AuroraMiddleware/qt5base-lts
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
AuQT66LTS
qt5base-lts/cmake/QtFlagHandlingHelpers.cmake

1136 lines
46 KiB
CMake
Raw Permalink Normal View History

Add license headers to cmake files CMakeLists.txt and .cmake files of significant size (more than 2 lines according to our check in tst_license.pl) now have the copyright and license header. Existing copyright statements remain intact Task-number: QTBUG-88621 Change-Id: I3b98cdc55ead806ec81ce09af9271f9b95af97fa Reviewed-by: Jörg Bornemann <joerg.bornemann@qt.io>
2022-07-05 11:26:52 +00:00
# Copyright (C) 2022 The Qt Company Ltd.
Change the license of all CMakeLists.txt and *.cmake files to BSD Task-number: QTBUG-105718 Change-Id: I5d3ef70a31235868b9be6cb479b7621bf2a8ba39 Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org> Reviewed-by: Jörg Bornemann <joerg.bornemann@qt.io>
2022-08-19 13:21:34 +00:00
# SPDX-License-Identifier: BSD-3-Clause
Add license headers to cmake files CMakeLists.txt and .cmake files of significant size (more than 2 lines according to our check in tst_license.pl) now have the copyright and license header. Existing copyright statements remain intact Task-number: QTBUG-88621 Change-Id: I3b98cdc55ead806ec81ce09af9271f9b95af97fa Reviewed-by: Jörg Bornemann <joerg.bornemann@qt.io>
2022-07-05 11:26:52 +00:00
Remove Qt_6_PRIVATE_API ELF version from a symbol used by QFuture::then() QFuture::then() uses QtPrivate::Continuation::create(), which in turn uses private API from an inline function: f->d.setContinuation(ContinuationWrapper(std::move(continuation)), fi.d); f->d is QFutureInterfaceBase (a public class), but its setContinuation() takes QFutureInterfaceBasePrivate by pointer. Our ELF versioning scripts mark everything that uses that class as private, resulting in: 4806: 0000000000287d70 365 FUNC GLOBAL PROTECTED 16 _ZN20QFutureInterfaceBase15setContinuationESt8functionIFvRKS_EEP27QFutureInterfaceBasePrivate@@Qt_6_PRIVATE_API This commit adds an exception for this symbol, causing it to go back to the regular "Qt_6" ELF version: 5629: 00000000003d6a16 366 FUNC GLOBAL PROTECTED 16 _ZN20QFutureInterfaceBase15setContinuationESt8functionIFvRKS_EEP27QFutureInterfaceBasePrivate@@Qt_6 This solution can probably be cleaned up a bit by moving the marker into the header files parsed by syncqt, so they follow code motion without having to remember to update the CMakeLists.txt. That requires some surgery with syncqt, so not suitable for cherry-picking. As a drive-by, fix the target_type check for the _qt_extra_linker_script_content genex property Fixes: QTBUG-117514 Pick-to: 6.6 Change-Id: I85599ea5ca7a4b79a8bbfffd178b92e73dbe11de Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2023-10-06 16:53:26 +00:00
# Sets '${var}' to a genex that extracts the target's property.
# Sets 'have_${var}' to a genex that checks that the property has a
# non-empty value.
macro(qt_internal_genex_get_property var target property)
set(${var} "$<TARGET_PROPERTY:${target},${property}>")
set(have_${var} "$<BOOL:${${var}}>")
endmacro()
# Sets '${var}' to a genex that will join the given property values
# using '${glue}' and will surround the entire output with '${prefix}'
# and '${suffix}'.
macro(qt_internal_genex_get_joined_property var target property prefix suffix glue)
qt_internal_genex_get_property("${var}" "${target}" "${property}")
set(${var}
"$<${have_${var}}:${prefix}$<JOIN:${${var}},${glue}>${suffix}>")
endmacro()
Add "content file" mode for the qt_internal_add_linker_version_script Add the support of pre-cooked content for the LD version script. The content can be generated without using the perl script at configure or build time. Change-Id: I1316e114a1d5550b2fdcf3482a51f336fb311a29 Reviewed-by: Axel Spoerl <axel.spoerl@qt.io> Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
2022-05-17 17:17:40 +00:00
# This function generates LD version script for the target and uses it in the target linker line.
# Function has two modes dependending on the specified arguments.
# Arguments:
# PRIVATE_CONTENT_FILE specifies the pre-cooked content of Qt_<version>_PRIVATE_API section.
# Requires the content file available at build time.
CMake: Split QtBuild.cmake into smaller files QtBuild.cmake is huge. Split it. Move module, plugin, tools, executables and test related functions out of QtBuild.cmake into separate files. Do the same for many other things too. An additional requirement is that all the new Helpers files only define functions and macros. No global variable definitions are allowed, nor execution of commands with side effects. Some notes: qt_install_qml_files is removed because it's dead code. Some functions still need to be figured out, because they are interspersed and depend on various global state assignments. Task-number: QTBUG-86035 Change-Id: I21d79ff02eef923c202eb1000422888727cb0e2c Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-08-13 15:37:47 +00:00
function(qt_internal_add_linker_version_script target)
wasm: add support for building Qt as shared libraries This implements the build system bits required to build Qt as as separate wasm modules a.k.a Emscripten side modules. Enable by configuring with the "-shared" flag. This is the first step towards shared library support and gets us as far as being able to load QtCore and instantiate a QCoreApplication. Task-number: QTBUG-63925 Change-Id: Ib8f07f80fb5b13c8dbba65c7db735dc557b70d0e Reviewed-by: Morten Johan Sørvig <morten.sorvig@qt.io>
2022-08-19 12:25:49 +00:00
if(WASM)
return()
endif()
Retire the qt_parse_all_arguments This replaces the qt_parse_all_arguments macro with the built-in `cmake_parse_arguments(PARSE_ARGV`. In addition, a new function, _qt_internal_validate_all_args_are_parsed, can be used to check whether any _UNPARSED_ARGUMENTS have been passed to the function. Fixes: QTBUG-99238 Change-Id: I8cee83dc92dc6acdaaf747ea6ff9084c11dc649b Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
2023-01-11 14:36:18 +00:00
cmake_parse_arguments(PARSE_ARGV 1 arg
Add "content file" mode for the qt_internal_add_linker_version_script Add the support of pre-cooked content for the LD version script. The content can be generated without using the perl script at configure or build time. Change-Id: I1316e114a1d5550b2fdcf3482a51f336fb311a29 Reviewed-by: Axel Spoerl <axel.spoerl@qt.io> Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
2022-05-17 17:17:40 +00:00
""
"PRIVATE_CONTENT_FILE"
"PRIVATE_HEADERS"
)
Retire the qt_parse_all_arguments This replaces the qt_parse_all_arguments macro with the built-in `cmake_parse_arguments(PARSE_ARGV`. In addition, a new function, _qt_internal_validate_all_args_are_parsed, can be used to check whether any _UNPARSED_ARGUMENTS have been passed to the function. Fixes: QTBUG-99238 Change-Id: I8cee83dc92dc6acdaaf747ea6ff9084c11dc649b Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
2023-01-11 14:36:18 +00:00
_qt_internal_validate_all_args_are_parsed(arg)
CMake: Split QtBuild.cmake into smaller files QtBuild.cmake is huge. Split it. Move module, plugin, tools, executables and test related functions out of QtBuild.cmake into separate files. Do the same for many other things too. An additional requirement is that all the new Helpers files only define functions and macros. No global variable definitions are allowed, nor execution of commands with side effects. Some notes: qt_install_qml_files is removed because it's dead code. Some functions still need to be figured out, because they are interspersed and depend on various global state assignments. Task-number: QTBUG-86035 Change-Id: I21d79ff02eef923c202eb1000422888727cb0e2c Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-08-13 15:37:47 +00:00
Add "content file" mode for the qt_internal_add_linker_version_script Add the support of pre-cooked content for the LD version script. The content can be generated without using the perl script at configure or build time. Change-Id: I1316e114a1d5550b2fdcf3482a51f336fb311a29 Reviewed-by: Axel Spoerl <axel.spoerl@qt.io> Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
2022-05-17 17:17:40 +00:00
if(arg_PRIVATE_CONTENT_FILE AND arg_PRIVATE_HEADERS)
message(FATAL_ERROR "Both PRIVATE_CONTENT_FILE and PRIVATE_HEADERS are specified.")
endif()
if(TEST_ld_version_script)
CMake/ELF: replace 'extern "C++"' with exhaustive mangled expansions Commit 946f15efb76fffda37b77f7d194d679b904305b1 added the line extern "C++" { std::* }; but that was too good to be true. The intention was to catch Standard Library inline symbols that got emitted in our own libraries, not use of Standard Library types in our symbols. Unfortunately, that glob expression matches the demangling literally and return types for templates are demangled in their usual position to the left of the function name. For example, std::random_device qFoo(); // mangles as _Z4qFoov → "qFoo()" but template <typename T> T qFoo(); template std::random_device qFoo<std::random_device>(); mangles as _Z4qFooISt13random_deviceET_v and that demangles to "std::random_device qFoo<std::random_device>()". Therefore, we replace that with a full expansion according to the mangling scheme. This includes a minor fix for the RTTI symbols, to match nested names too (those with "N" in the name). It can't match virtual override thunks ("Tv" and "Th"), because those have variable length names and the matching is done by fnmatch(), not regex. Pick-to: 6.6 Change-Id: I9d43e5b91eb142d6945cfffd178449cf68669cb6 Reviewed-by: Alexey Edelev <alexey.edelev@qt.io>
2023-09-12 23:15:14 +00:00
# Create a list of mangled symbol matches for all "std::" symbols. This
# list will catch most symbols, but will miss global-namespace symbols
# that only have std parameters.
# See https://itanium-cxx-abi.github.io/cxx-abi/abi.html#mangle.name for reference
set(contents "NonQt {\nlocal:")
# For types: vtable, VTT, typeinfo, typeinfo name
foreach(ptrqualifier "" "P" "PK") # T, T *, const T * (volatile ignored)
string(APPEND contents "\n _ZT[VTIS]${ptrqualifier}S*;"
"_ZT[VTIS]${ptrqualifier}NS*;")
endforeach()
# For functions and variables
foreach(special ""
"G[VR]" # guard variables, extended-lifetime references
"GTt") # transaction entry points
foreach(cvqualifier "" "[VK]" "VK") # plain, const|volatile, const volatile
string(APPEND contents "\n ")
foreach(refqualifier "" "[RO]") # plain, & or &&
# For names in the std:: namespace, compression applies
# (https://itanium-cxx-abi.github.io/cxx-abi/abi.html#mangling-compression)
string(APPEND contents
" _Z${special}${cvqualifier}${refqualifier}S*;" # plain
" _Z${special}N${cvqualifier}${refqualifier}S*;" # nested name
)
endforeach()
endforeach()
endforeach()
string(APPEND contents "\n};\nQt_${PROJECT_VERSION_MAJOR}_PRIVATE_API { qt_private_api_tag*;\n")
Add "content file" mode for the qt_internal_add_linker_version_script Add the support of pre-cooked content for the LD version script. The content can be generated without using the perl script at configure or build time. Change-Id: I1316e114a1d5550b2fdcf3482a51f336fb311a29 Reviewed-by: Axel Spoerl <axel.spoerl@qt.io> Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
2022-05-17 17:17:40 +00:00
if(arg_PRIVATE_HEADERS)
foreach(ph ${arg_PRIVATE_HEADERS})
string(APPEND contents " @FILE:${ph}@\n")
endforeach()
else()
string(APPEND contents "@PRIVATE_CONTENT@")
endif()
Remove INTERNAL argument from qt_internal_add_linker_version_script The code branch for the INTERNAL argument did nothing, and we never call qt_internal_add_linker_version_script with INTERNAL. Change-Id: Ie369b4dac29cd1a977433ebfd662c198a3e1d0f2 Reviewed-by: Alexey Edelev <alexey.edelev@qt.io> Reviewed-by: Cristian Adam <cristian.adam@qt.io>
2021-02-15 15:40:08 +00:00
string(APPEND contents "};\n")
set(current "Qt_${PROJECT_VERSION_MAJOR}")
Remove Qt_6_PRIVATE_API ELF version from a symbol used by QFuture::then() QFuture::then() uses QtPrivate::Continuation::create(), which in turn uses private API from an inline function: f->d.setContinuation(ContinuationWrapper(std::move(continuation)), fi.d); f->d is QFutureInterfaceBase (a public class), but its setContinuation() takes QFutureInterfaceBasePrivate by pointer. Our ELF versioning scripts mark everything that uses that class as private, resulting in: 4806: 0000000000287d70 365 FUNC GLOBAL PROTECTED 16 _ZN20QFutureInterfaceBase15setContinuationESt8functionIFvRKS_EEP27QFutureInterfaceBasePrivate@@Qt_6_PRIVATE_API This commit adds an exception for this symbol, causing it to go back to the regular "Qt_6" ELF version: 5629: 00000000003d6a16 366 FUNC GLOBAL PROTECTED 16 _ZN20QFutureInterfaceBase15setContinuationESt8functionIFvRKS_EEP27QFutureInterfaceBasePrivate@@Qt_6 This solution can probably be cleaned up a bit by moving the marker into the header files parsed by syncqt, so they follow code motion without having to remember to update the CMakeLists.txt. That requires some surgery with syncqt, so not suitable for cherry-picking. As a drive-by, fix the target_type check for the _qt_extra_linker_script_content genex property Fixes: QTBUG-117514 Pick-to: 6.6 Change-Id: I85599ea5ca7a4b79a8bbfffd178b92e73dbe11de Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2023-10-06 16:53:26 +00:00
string(APPEND contents "${current} {\n *;")
get_target_property(target_type ${target} TYPE)
if(NOT target_type STREQUAL "INTERFACE_LIBRARY")
set(genex_prefix "\n ")
set(genex_glue "$<SEMICOLON>\n ")
set(genex_suffix "$<SEMICOLON>")
qt_internal_genex_get_joined_property(
linker_exports "${target}" _qt_extra_linker_script_exports
"${genex_prefix}" "${genex_suffix}" "${genex_glue}"
)
string(APPEND contents "${linker_exports}")
endif()
string(APPEND contents "\n};\n")
Remove INTERNAL argument from qt_internal_add_linker_version_script The code branch for the INTERNAL argument did nothing, and we never call qt_internal_add_linker_version_script with INTERNAL. Change-Id: Ie369b4dac29cd1a977433ebfd662c198a3e1d0f2 Reviewed-by: Alexey Edelev <alexey.edelev@qt.io> Reviewed-by: Cristian Adam <cristian.adam@qt.io>
2021-02-15 15:40:08 +00:00
CMake: Fix linkage with lld 16.0 lld 16.0 is more picky about symbol versioning than previous versions (and other linkers such as ld.bfd, gold or mold). It now errors out if a symbol is versioned but not defined (see 8796677de8900dc154aef45f8620c3f987a40291). Outside of detecting support for symbol versioning (fixed by 462832), this causes linking Qt6 libraries other than Qt6Core to fail because their linker scripts try to add versioning to qt_version_tag, which is defined in Qt6Core rather than the library being linked. The obvious (and working) fix is to version qt_version_tag only where it is defined (Qt6Core), but this is not what the original intent seems to be. Task-number: QTBUG-111514 Change-Id: I963d417befb0f6b2260c57f059eeda1fe79200c9 Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
2023-02-25 18:10:06 +00:00
if(NOT target_type STREQUAL "INTERFACE_LIBRARY")
set(property_genex "$<TARGET_PROPERTY:${target},_qt_extra_linker_script_content>")
set(check_genex "$<BOOL:${property_genex}>")
string(APPEND contents
"$<${check_genex}:${property_genex}>")
endif()
CMake: Split QtBuild.cmake into smaller files QtBuild.cmake is huge. Split it. Move module, plugin, tools, executables and test related functions out of QtBuild.cmake into separate files. Do the same for many other things too. An additional requirement is that all the new Helpers files only define functions and macros. No global variable definitions are allowed, nor execution of commands with side effects. Some notes: qt_install_qml_files is removed because it's dead code. Some functions still need to be figured out, because they are interspersed and depend on various global state assignments. Task-number: QTBUG-86035 Change-Id: I21d79ff02eef923c202eb1000422888727cb0e2c Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-08-13 15:37:47 +00:00
Remove INTERNAL argument from qt_internal_add_linker_version_script The code branch for the INTERNAL argument did nothing, and we never call qt_internal_add_linker_version_script with INTERNAL. Change-Id: Ie369b4dac29cd1a977433ebfd662c198a3e1d0f2 Reviewed-by: Alexey Edelev <alexey.edelev@qt.io> Reviewed-by: Cristian Adam <cristian.adam@qt.io>
2021-02-15 15:40:08 +00:00
set(infile "${CMAKE_CURRENT_BINARY_DIR}/${target}.version.in")
set(outfile "${CMAKE_CURRENT_BINARY_DIR}/${target}.version")
CMake: Split QtBuild.cmake into smaller files QtBuild.cmake is huge. Split it. Move module, plugin, tools, executables and test related functions out of QtBuild.cmake into separate files. Do the same for many other things too. An additional requirement is that all the new Helpers files only define functions and macros. No global variable definitions are allowed, nor execution of commands with side effects. Some notes: qt_install_qml_files is removed because it's dead code. Some functions still need to be figured out, because they are interspersed and depend on various global state assignments. Task-number: QTBUG-86035 Change-Id: I21d79ff02eef923c202eb1000422888727cb0e2c Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-08-13 15:37:47 +00:00
Remove INTERNAL argument from qt_internal_add_linker_version_script The code branch for the INTERNAL argument did nothing, and we never call qt_internal_add_linker_version_script with INTERNAL. Change-Id: Ie369b4dac29cd1a977433ebfd662c198a3e1d0f2 Reviewed-by: Alexey Edelev <alexey.edelev@qt.io> Reviewed-by: Cristian Adam <cristian.adam@qt.io>
2021-02-15 15:40:08 +00:00
file(GENERATE OUTPUT "${infile}" CONTENT "${contents}")
CMake: Split QtBuild.cmake into smaller files QtBuild.cmake is huge. Split it. Move module, plugin, tools, executables and test related functions out of QtBuild.cmake into separate files. Do the same for many other things too. An additional requirement is that all the new Helpers files only define functions and macros. No global variable definitions are allowed, nor execution of commands with side effects. Some notes: qt_install_qml_files is removed because it's dead code. Some functions still need to be figured out, because they are interspersed and depend on various global state assignments. Task-number: QTBUG-86035 Change-Id: I21d79ff02eef923c202eb1000422888727cb0e2c Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-08-13 15:37:47 +00:00
Remove perl related functionality from CMake files Avoid using perl in CMake scripts. Remove the syncqt.pl specific code. Task-number: QTBUG-87480 Change-Id: I7fcd5cc83d173ec463c275b5b50b84f25044a118 Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
2022-10-05 14:32:31 +00:00
if(NOT arg_PRIVATE_CONTENT_FILE)
set(arg_PRIVATE_CONTENT_FILE "")
Add "content file" mode for the qt_internal_add_linker_version_script Add the support of pre-cooked content for the LD version script. The content can be generated without using the perl script at configure or build time. Change-Id: I1316e114a1d5550b2fdcf3482a51f336fb311a29 Reviewed-by: Axel Spoerl <axel.spoerl@qt.io> Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
2022-05-17 17:17:40 +00:00
endif()
Remove perl related functionality from CMake files Avoid using perl in CMake scripts. Remove the syncqt.pl specific code. Task-number: QTBUG-87480 Change-Id: I7fcd5cc83d173ec463c275b5b50b84f25044a118 Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
2022-10-05 14:32:31 +00:00
set(generator_command ${CMAKE_COMMAND}
"-DIN_FILE=${infile}"
"-DPRIVATE_CONTENT_FILE=${arg_PRIVATE_CONTENT_FILE}"
"-DOUT_FILE=${outfile}"
-P "${QT_CMAKE_DIR}/QtGenerateVersionScript.cmake"
)
set(generator_dependencies
"${arg_PRIVATE_CONTENT_FILE}"
"${QT_CMAKE_DIR}/QtGenerateVersionScript.cmake"
)
Avoid using add_custom_command with PRE_LINK for version script add_custom_command with PRE_LINK doesn't work correctly with Multi-Config builds. The better solution is to introduce a custom target that generates the final version script and link the target to the library target as the dependency. Change-Id: Ib7420af752a6a46f29f411f9f0dc8557410b4f22 Reviewed-by: Axel Spoerl <axel.spoerl@qt.io> Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
2022-05-18 16:14:12 +00:00
add_custom_command(
OUTPUT "${outfile}"
COMMAND ${generator_command}
DEPENDS ${generator_dependencies}
Remove INTERNAL argument from qt_internal_add_linker_version_script The code branch for the INTERNAL argument did nothing, and we never call qt_internal_add_linker_version_script with INTERNAL. Change-Id: Ie369b4dac29cd1a977433ebfd662c198a3e1d0f2 Reviewed-by: Alexey Edelev <alexey.edelev@qt.io> Reviewed-by: Cristian Adam <cristian.adam@qt.io>
2021-02-15 15:40:08 +00:00
WORKING_DIRECTORY "${CMAKE_CURRENT_SOURCE_DIR}"
Avoid using add_custom_command with PRE_LINK for version script add_custom_command with PRE_LINK doesn't work correctly with Multi-Config builds. The better solution is to introduce a custom target that generates the final version script and link the target to the library target as the dependency. Change-Id: Ib7420af752a6a46f29f411f9f0dc8557410b4f22 Reviewed-by: Axel Spoerl <axel.spoerl@qt.io> Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
2022-05-18 16:14:12 +00:00
COMMENT "Generating version linker script for target ${target}"
VERBATIM
)
add_custom_target(${target}_version_script DEPENDS ${outfile})
add_dependencies(${target} ${target}_version_script)
target_link_options(${target} PRIVATE "-Wl,--version-script,${outfile}")
CMake: Split QtBuild.cmake into smaller files QtBuild.cmake is huge. Split it. Move module, plugin, tools, executables and test related functions out of QtBuild.cmake into separate files. Do the same for many other things too. An additional requirement is that all the new Helpers files only define functions and macros. No global variable definitions are allowed, nor execution of commands with side effects. Some notes: qt_install_qml_files is removed because it's dead code. Some functions still need to be figured out, because they are interspersed and depend on various global state assignments. Task-number: QTBUG-86035 Change-Id: I21d79ff02eef923c202eb1000422888727cb0e2c Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-08-13 15:37:47 +00:00
endif()
endfunction()
function(qt_internal_add_link_flags_no_undefined target)
wasm: add support for building Qt as shared libraries This implements the build system bits required to build Qt as as separate wasm modules a.k.a Emscripten side modules. Enable by configuring with the "-shared" flag. This is the first step towards shared library support and gets us as far as being able to load QtCore and instantiate a QCoreApplication. Task-number: QTBUG-63925 Change-Id: Ib8f07f80fb5b13c8dbba65c7db735dc557b70d0e Reviewed-by: Morten Johan Sørvig <morten.sorvig@qt.io>
2022-08-19 12:25:49 +00:00
if (NOT QT_BUILD_SHARED_LIBS OR WASM)
CMake: Split QtBuild.cmake into smaller files QtBuild.cmake is huge. Split it. Move module, plugin, tools, executables and test related functions out of QtBuild.cmake into separate files. Do the same for many other things too. An additional requirement is that all the new Helpers files only define functions and macros. No global variable definitions are allowed, nor execution of commands with side effects. Some notes: qt_install_qml_files is removed because it's dead code. Some functions still need to be figured out, because they are interspersed and depend on various global state assignments. Task-number: QTBUG-86035 Change-Id: I21d79ff02eef923c202eb1000422888727cb0e2c Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-08-13 15:37:47 +00:00
return()
endif()
Remove --no-undefined for VxWorks Qt shared libs Build system adds this flag to any module that it uses when built as a shared library and not WASM. VxWorks is specific in regards to type of applications: - static application, which do not use any dynamic libraries - dynamic application, which uses shared libraries, and as such needs to link against libdl.a The latter is the only source of symbols related to dynamic loading and thread local storage. This is an architectural decision, as application is required to embed loading tools in itself (as opposed to Unix/Linux, which contains dynamic loader which provides these symbols separately). libdl.a is standard static library on VxWorks, compiled without PIC, and as such can't be linked into shared library. This results in compilation errors when building dynamic libraries which use `thread_local`-related symbols, since these symbols will be fulfilled at runtime (contained in executable which requires shared lib) and are not available during linking. Don't add `-Wl,--no-undefined` flag on VxWorks for Qt shared lib builds. Task-number: QTBUG-115777 Change-Id: I47e5d68afcba2c3b4203e4c0beded2e18ea2563b Reviewed-by: Pasi Petäjäjärvi <pasi.petajajarvi@qt.io>
2023-09-13 17:14:57 +00:00
if (VXWORKS)
# VxWorks requires thread_local-related symbols to be found at
# runtime, resulting in linker error when no-undefined flag is
# set and thread_local is used
return()
endif()
cmake: Don't pass -undefined,error to Apple linker Error is the default for this option in ld64, and with the new linker in Xcode 15 the option was deprecated, causing a warning, so let's just skip adding it and assume undefined symbols results in errors. Fixes: QTBUG-116732 Pick-to: 6.6 6.5 Change-Id: I736a891ae102fe1dfb2cdf869f42c665e231adba Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
2023-09-11 11:18:25 +00:00
if(CMAKE_CXX_COMPILER_ID STREQUAL "AppleClang")
# ld64 defaults to -undefined,error, and in Xcode 15
# passing this option is deprecated, causing a warning.
return()
endif()
Fix the win32-clang-msvc target Detect clang-cl mkspec target Pick-to: 6.2 Change-Id: If0ca31ae2da3b44a4e2bd116933007139cc02fdc Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2021-06-15 10:45:50 +00:00
if ((GCC OR CLANG) AND NOT MSVC)
CMake: Fix sanitizer build when using Clang on Linux Apparently the combination of the --no-undefined linker flag together with ASAN when building on Linux with Clang does not work. Disable --no-undefined flag in such a scenario. Note that linux-clang mkspec doesn't add that flag at all, which is why asan builds work there. Change-Id: I6167c757ce4be5d2263311bc84e5fb445b0f7c2d Fixes: QTBUG-86879 Reviewed-by: Cristian Adam <cristian.adam@qt.io> Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-09-24 15:00:25 +00:00
if(CLANG AND QT_FEATURE_sanitizer)
return()
endif()
CMake: Split QtBuild.cmake into smaller files QtBuild.cmake is huge. Split it. Move module, plugin, tools, executables and test related functions out of QtBuild.cmake into separate files. Do the same for many other things too. An additional requirement is that all the new Helpers files only define functions and macros. No global variable definitions are allowed, nor execution of commands with side effects. Some notes: qt_install_qml_files is removed because it's dead code. Some functions still need to be figured out, because they are interspersed and depend on various global state assignments. Task-number: QTBUG-86035 Change-Id: I21d79ff02eef923c202eb1000422888727cb0e2c Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-08-13 15:37:47 +00:00
set(previous_CMAKE_REQUIRED_LINK_OPTIONS ${CMAKE_REQUIRED_LINK_OPTIONS})
set(CMAKE_REQUIRED_LINK_OPTIONS "-Wl,-undefined,error")
check_cxx_source_compiles("int main() {}" HAVE_DASH_UNDEFINED_SYMBOLS)
if(HAVE_DASH_UNDEFINED_SYMBOLS)
set(no_undefined_flag "-Wl,-undefined,error")
endif()
set(CMAKE_REQUIRED_LINK_OPTIONS "-Wl,--no-undefined")
check_cxx_source_compiles("int main() {}" HAVE_DASH_DASH_NO_UNDEFINED)
if(HAVE_DASH_DASH_NO_UNDEFINED)
set(no_undefined_flag "-Wl,--no-undefined")
endif()
set(CMAKE_REQUIRED_LINK_OPTIONS ${previous_CMAKE_REQUIRED_LINK_OPTIONS})
if (NOT HAVE_DASH_UNDEFINED_SYMBOLS AND NOT HAVE_DASH_DASH_NO_UNDEFINED)
message(FATAL_ERROR "Platform linker doesn't support erroring upon encountering undefined symbols. Target:\"${target}\".")
endif()
target_link_options("${target}" PRIVATE "${no_undefined_flag}")
endif()
endfunction()
function(qt_internal_apply_gc_binaries_conditional target visibility)
# Should only be applied when the feature is enabled, aka for static builds.
if(NOT QT_FEATURE_gc_binaries)
return()
endif()
qt_internal_apply_gc_binaries("${target}" "${visibility}")
endfunction()
function(qt_internal_apply_gc_binaries target visibility)
set(possible_visibilities PRIVATE INTERFACE PUBLIC)
Replace list(FIND with simplified IN_LIST check Pick-to: 6.6 6.5 Change-Id: I2ed1803adabc36e646edae3b2bb07e5ff2194d9b Reviewed-by: Amir Masoud Abdol <amir.abdol@qt.io> Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
2023-06-08 10:30:07 +00:00
if(NOT visibility IN_LIST possible_visibilities)
CMake: Split QtBuild.cmake into smaller files QtBuild.cmake is huge. Split it. Move module, plugin, tools, executables and test related functions out of QtBuild.cmake into separate files. Do the same for many other things too. An additional requirement is that all the new Helpers files only define functions and macros. No global variable definitions are allowed, nor execution of commands with side effects. Some notes: qt_install_qml_files is removed because it's dead code. Some functions still need to be figured out, because they are interspersed and depend on various global state assignments. Task-number: QTBUG-86035 Change-Id: I21d79ff02eef923c202eb1000422888727cb0e2c Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-08-13 15:37:47 +00:00
message(FATAL_ERROR "Visibitily setting must be one of PRIVATE, INTERFACE or PUBLIC.")
endif()
Wrap the GNU/Clang compiler-dependent flags with genex conditions If flags use CMake scopes that propagate them to user libraries, it may lead to an issue, if user projects are built using different compiler. We need to guard these flags to make sure that they only will apply to respective compilers. Pick-to: 6.5 6.6 Change-Id: I0fd5847447bd8373e8e07f64dae11f27f48c915d Reviewed-by: Alexandru Croitor (OOO) <alexandru.croitor@qt.io> Reviewed-by: Amir Masoud Abdol <amir.abdol@qt.io>
2023-05-12 15:21:22 +00:00
string(JOIN "" clang_or_gcc_begin
"$<$<OR:"
"$<CXX_COMPILER_ID:GNU>,"
"$<CXX_COMPILER_ID:Clang>,"
"$<CXX_COMPILER_ID:AppleClang>,"
"$<CXX_COMPILER_ID:IntelLLVM>"
">:"
)
set(clang_or_gcc_end ">")
wasm: add cmake build support A few configure defines get changed: QMAKE_WASM_PTHREAD_POOL_SIZE is now QT_WASM_PTHREAD_POOL_SIZE QMAKE_WASM_TOTAL_MEMORY is now QT_WASM_INITIAL_MEMORY QMAKE_WASM_SOURCE_MAP_BASE is now QT_WASM_SOURCE_MAP_BASE device-option EMSCRIPTEN_ASYNCIFY=1 is QT_EMSCRIPTEN_ASYNCIFY=1 To create source maps for debugging. use device-option QT_WASM_SOURCE_MAP=1 Task-number: QTBUG-78647 Change-Id: If9f30cd7fb408c386d6d69b5f7b1beecf1ab44b5 Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
2020-09-03 08:29:41 +00:00
if ((GCC OR CLANG) AND NOT WASM AND NOT UIKIT AND NOT MSVC)
CMake: Split QtBuild.cmake into smaller files QtBuild.cmake is huge. Split it. Move module, plugin, tools, executables and test related functions out of QtBuild.cmake into separate files. Do the same for many other things too. An additional requirement is that all the new Helpers files only define functions and macros. No global variable definitions are allowed, nor execution of commands with side effects. Some notes: qt_install_qml_files is removed because it's dead code. Some functions still need to be figured out, because they are interspersed and depend on various global state assignments. Task-number: QTBUG-86035 Change-Id: I21d79ff02eef923c202eb1000422888727cb0e2c Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-08-13 15:37:47 +00:00
if(APPLE)
set(gc_sections_flag "-Wl,-dead_strip")
elseif(SOLARIS)
set(gc_sections_flag "-Wl,-z,ignore")
elseif(LINUX OR BSD OR WIN32 OR ANDROID)
set(gc_sections_flag "-Wl,--gc-sections")
endif()
CMake: Fix gc_sections genex evaluation in pkgconfig file creation The gc_sections linker flag was recently wrapped in a $<CXX_COMPILER_ID> genex to prevent adding it the command line when using an incompatible compiler. This causes an issue when generating .pc pkg-config files because $<CXX_COMPILER_ID> can't be used in the output of a file(GENERATE) call. Record the flag in a global property, both the genex-wrapped and bare forms, so that we can perform a string replacement when generating the pkg-config file to remove the genex wrapping. This is not perfect, in the sense that consumers of the .pc file might get the wrong flag if using an incompatible compiler, but it's better than outright failing the Qt build. Distros will be expected to patch the .pc files if necessary. Note the issue does not usually happen for regular Qt builds because gc_sections is only enabled automatically for static builds, but for static builds we don't currently generate .pc files. So the issue only happens in shared Qt builds where the gc_sections feature is enabled manually. Amends a2b6c2f3437bf1779da787e719bea08bc6f28622 Pick-to: 6.5 6.6 Fixes: QTBUG-115243 Change-Id: I3f6bdf86c24ee90b6da04994e458b438cc41fc7a Reviewed-by: Amir Masoud Abdol <amir.abdol@qt.io>
2023-07-14 13:24:19 +00:00
# Save the flag value with and without genex wrapping, so we can remove the wrapping
# when generating .pc pkgconfig files.
set_property(GLOBAL PROPERTY _qt_internal_gc_sections_without_genex "${gc_sections_flag}")
Wrap the GNU/Clang compiler-dependent flags with genex conditions If flags use CMake scopes that propagate them to user libraries, it may lead to an issue, if user projects are built using different compiler. We need to guard these flags to make sure that they only will apply to respective compilers. Pick-to: 6.5 6.6 Change-Id: I0fd5847447bd8373e8e07f64dae11f27f48c915d Reviewed-by: Alexandru Croitor (OOO) <alexandru.croitor@qt.io> Reviewed-by: Amir Masoud Abdol <amir.abdol@qt.io>
2023-05-12 15:21:22 +00:00
set(gc_sections_flag
"${clang_or_gcc_begin}${gc_sections_flag}${clang_or_gcc_end}")
CMake: Fix gc_sections genex evaluation in pkgconfig file creation The gc_sections linker flag was recently wrapped in a $<CXX_COMPILER_ID> genex to prevent adding it the command line when using an incompatible compiler. This causes an issue when generating .pc pkg-config files because $<CXX_COMPILER_ID> can't be used in the output of a file(GENERATE) call. Record the flag in a global property, both the genex-wrapped and bare forms, so that we can perform a string replacement when generating the pkg-config file to remove the genex wrapping. This is not perfect, in the sense that consumers of the .pc file might get the wrong flag if using an incompatible compiler, but it's better than outright failing the Qt build. Distros will be expected to patch the .pc files if necessary. Note the issue does not usually happen for regular Qt builds because gc_sections is only enabled automatically for static builds, but for static builds we don't currently generate .pc files. So the issue only happens in shared Qt builds where the gc_sections feature is enabled manually. Amends a2b6c2f3437bf1779da787e719bea08bc6f28622 Pick-to: 6.5 6.6 Fixes: QTBUG-115243 Change-Id: I3f6bdf86c24ee90b6da04994e458b438cc41fc7a Reviewed-by: Amir Masoud Abdol <amir.abdol@qt.io>
2023-07-14 13:24:19 +00:00
set_property(GLOBAL PROPERTY _qt_internal_gc_sections_with_genex "${gc_sections_flag}")
CMake: Split QtBuild.cmake into smaller files QtBuild.cmake is huge. Split it. Move module, plugin, tools, executables and test related functions out of QtBuild.cmake into separate files. Do the same for many other things too. An additional requirement is that all the new Helpers files only define functions and macros. No global variable definitions are allowed, nor execution of commands with side effects. Some notes: qt_install_qml_files is removed because it's dead code. Some functions still need to be figured out, because they are interspersed and depend on various global state assignments. Task-number: QTBUG-86035 Change-Id: I21d79ff02eef923c202eb1000422888727cb0e2c Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-08-13 15:37:47 +00:00
endif()
if(gc_sections_flag)
target_link_options("${target}" ${visibility} "${gc_sections_flag}")
endif()
CMake: remove support for building Qt with the old Intel compiler This hasn't worked for some time. It's not in our CI and I don't think it was working at all. When I tried to build it, I ran into several problems with C++17 and an Internal Compiler Error I did not have any interest in working around. After discussing with the Intel compiler team, it was decided that fixing those issues in the old compiler is not going to happen. Instead, their recommendation is to adopt the new LLVM-based compiler, which the last commit added support for. This commit does not remove qmake support for the old ICC. It's possible someone is using qmake with a non-Qt6 project and ICC. Change-Id: Icb2516126f674e7b8bb3fffd16ad6350ddbd49e5 Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
2021-10-12 20:39:37 +00:00
if((GCC OR CLANG) AND NOT WASM AND NOT UIKIT AND NOT MSVC)
Wrap the GNU/Clang compiler-dependent flags with genex conditions If flags use CMake scopes that propagate them to user libraries, it may lead to an issue, if user projects are built using different compiler. We need to guard these flags to make sure that they only will apply to respective compilers. Pick-to: 6.5 6.6 Change-Id: I0fd5847447bd8373e8e07f64dae11f27f48c915d Reviewed-by: Alexandru Croitor (OOO) <alexandru.croitor@qt.io> Reviewed-by: Amir Masoud Abdol <amir.abdol@qt.io>
2023-05-12 15:21:22 +00:00
set(split_sections_flags
"${clang_or_gcc_begin}-ffunction-sections;-fdata-sections${clang_or_gcc_end}")
CMake: Split QtBuild.cmake into smaller files QtBuild.cmake is huge. Split it. Move module, plugin, tools, executables and test related functions out of QtBuild.cmake into separate files. Do the same for many other things too. An additional requirement is that all the new Helpers files only define functions and macros. No global variable definitions are allowed, nor execution of commands with side effects. Some notes: qt_install_qml_files is removed because it's dead code. Some functions still need to be figured out, because they are interspersed and depend on various global state assignments. Task-number: QTBUG-86035 Change-Id: I21d79ff02eef923c202eb1000422888727cb0e2c Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-08-13 15:37:47 +00:00
endif()
if(split_sections_flags)
Wrap the GNU/Clang compiler-dependent flags with genex conditions If flags use CMake scopes that propagate them to user libraries, it may lead to an issue, if user projects are built using different compiler. We need to guard these flags to make sure that they only will apply to respective compilers. Pick-to: 6.5 6.6 Change-Id: I0fd5847447bd8373e8e07f64dae11f27f48c915d Reviewed-by: Alexandru Croitor (OOO) <alexandru.croitor@qt.io> Reviewed-by: Amir Masoud Abdol <amir.abdol@qt.io>
2023-05-12 15:21:22 +00:00
target_compile_options("${target}" ${visibility} "${split_sections_flags}")
CMake: Split QtBuild.cmake into smaller files QtBuild.cmake is huge. Split it. Move module, plugin, tools, executables and test related functions out of QtBuild.cmake into separate files. Do the same for many other things too. An additional requirement is that all the new Helpers files only define functions and macros. No global variable definitions are allowed, nor execution of commands with side effects. Some notes: qt_install_qml_files is removed because it's dead code. Some functions still need to be figured out, because they are interspersed and depend on various global state assignments. Task-number: QTBUG-86035 Change-Id: I21d79ff02eef923c202eb1000422888727cb0e2c Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-08-13 15:37:47 +00:00
endif()
endfunction()
function(qt_internal_apply_intel_cet target visibility)
if(NOT QT_FEATURE_intelcet)
return()
endif()
set(possible_visibilities PRIVATE INTERFACE PUBLIC)
Replace list(FIND with simplified IN_LIST check Pick-to: 6.6 6.5 Change-Id: I2ed1803adabc36e646edae3b2bb07e5ff2194d9b Reviewed-by: Amir Masoud Abdol <amir.abdol@qt.io> Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
2023-06-08 10:30:07 +00:00
if(NOT visibility IN_LIST possible_visibilities)
CMake: Split QtBuild.cmake into smaller files QtBuild.cmake is huge. Split it. Move module, plugin, tools, executables and test related functions out of QtBuild.cmake into separate files. Do the same for many other things too. An additional requirement is that all the new Helpers files only define functions and macros. No global variable definitions are allowed, nor execution of commands with side effects. Some notes: qt_install_qml_files is removed because it's dead code. Some functions still need to be figured out, because they are interspersed and depend on various global state assignments. Task-number: QTBUG-86035 Change-Id: I21d79ff02eef923c202eb1000422888727cb0e2c Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-08-13 15:37:47 +00:00
message(FATAL_ERROR "Visibitily setting must be one of PRIVATE, INTERFACE or PUBLIC.")
endif()
if(GCC)
Wrap the GNU/Clang compiler-dependent flags with genex conditions If flags use CMake scopes that propagate them to user libraries, it may lead to an issue, if user projects are built using different compiler. We need to guard these flags to make sure that they only will apply to respective compilers. Pick-to: 6.5 6.6 Change-Id: I0fd5847447bd8373e8e07f64dae11f27f48c915d Reviewed-by: Alexandru Croitor (OOO) <alexandru.croitor@qt.io> Reviewed-by: Amir Masoud Abdol <amir.abdol@qt.io>
2023-05-12 15:21:22 +00:00
string(JOIN "" flags
"$<$<OR:"
"$<CXX_COMPILER_ID:GNU>,"
"$<CXX_COMPILER_ID:Clang>,"
"$<CXX_COMPILER_ID:AppleClang>"
">:-mshstk>")
CMake: Split QtBuild.cmake into smaller files QtBuild.cmake is huge. Split it. Move module, plugin, tools, executables and test related functions out of QtBuild.cmake into separate files. Do the same for many other things too. An additional requirement is that all the new Helpers files only define functions and macros. No global variable definitions are allowed, nor execution of commands with side effects. Some notes: qt_install_qml_files is removed because it's dead code. Some functions still need to be figured out, because they are interspersed and depend on various global state assignments. Task-number: QTBUG-86035 Change-Id: I21d79ff02eef923c202eb1000422888727cb0e2c Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-08-13 15:37:47 +00:00
endif()
if(flags)
target_compile_options("${target}" ${visibility} "${flags}")
endif()
endfunction()
function(qt_internal_library_deprecation_level result)
Rename QT_DISABLE_DEPRECATED_BEFORE -> QT_DISABLE_DEPRECATED_UP_TO The new name describes the behavior in a better way. [ChangeLog][Build System] The QT_DISABLE_DEPRECATED_BEFORE macro is renamed to QT_DISABLE_DEPRECATED_UP_TO. The old name is deprecated, but is still recognized if it is defined during configuration and the new name is not defined. Task-number: QTBUG-104944 Change-Id: Ifc34323e0bbd9e3dc2f86c3e80d4d0940ebccbb8 Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io> Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
2022-08-16 10:26:25 +00:00
# QT_DISABLE_DEPRECATED_UP_TO controls which version we use as a cut-off
Bump QT_DEPRECATED_WARNINGS_SINCE When building Qt we should have warnings of any internal use of deprecated API, so set it to Qt7. Also added comments to clarify what the macros do. Change-Id: Ib47278fed9ab1ec4411ed9f69a9a9c0f811db02d Reviewed-by: Marc Mutz <marc.mutz@qt.io>
2022-03-04 09:21:06 +00:00
# compiling in to the library. E.g. if it is set to QT_VERSION then no
# code which was deprecated before QT_VERSION will be compiled in.
Add -disable-deprecated-up-to parameter to configure script And also teach CMake to treat it properly instead of hardcoding the version number. [ChangeLog][Build System] The configure script now accepts a new parameter -disable-deprecated-up-to which is used to remove all deprecated code from API and ABI while building the libraries. The version number must be specified in a hex format. For example, it can be used like this: /path/to/qt/configure -disable-deprecated-up-to 0x060500 to remove all code deprecated in Qt 6.5.0 or earlier releases. Task-number: QTBUG-101510 Change-Id: I557cf83e29b867fa1052bb097985e144b5eaf34d Reviewed-by: Edward Welbourne <edward.welbourne@qt.io> Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
2022-08-16 11:02:27 +00:00
if (NOT DEFINED QT_DISABLE_DEPRECATED_UP_TO)
if(WIN32)
# On Windows, due to the way DLLs work, we need to export all functions,
# including the inlines
list(APPEND deprecations "QT_DISABLE_DEPRECATED_UP_TO=0x040800")
else()
# On other platforms, Qt's own compilation does need to compile the Qt 5.0 API
list(APPEND deprecations "QT_DISABLE_DEPRECATED_UP_TO=0x050000")
endif()
CMake: Split QtBuild.cmake into smaller files QtBuild.cmake is huge. Split it. Move module, plugin, tools, executables and test related functions out of QtBuild.cmake into separate files. Do the same for many other things too. An additional requirement is that all the new Helpers files only define functions and macros. No global variable definitions are allowed, nor execution of commands with side effects. Some notes: qt_install_qml_files is removed because it's dead code. Some functions still need to be figured out, because they are interspersed and depend on various global state assignments. Task-number: QTBUG-86035 Change-Id: I21d79ff02eef923c202eb1000422888727cb0e2c Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-08-13 15:37:47 +00:00
else()
Add -disable-deprecated-up-to parameter to configure script And also teach CMake to treat it properly instead of hardcoding the version number. [ChangeLog][Build System] The configure script now accepts a new parameter -disable-deprecated-up-to which is used to remove all deprecated code from API and ABI while building the libraries. The version number must be specified in a hex format. For example, it can be used like this: /path/to/qt/configure -disable-deprecated-up-to 0x060500 to remove all code deprecated in Qt 6.5.0 or earlier releases. Task-number: QTBUG-101510 Change-Id: I557cf83e29b867fa1052bb097985e144b5eaf34d Reviewed-by: Edward Welbourne <edward.welbourne@qt.io> Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
2022-08-16 11:02:27 +00:00
list(APPEND deprecations "QT_DISABLE_DEPRECATED_UP_TO=${QT_DISABLE_DEPRECATED_UP_TO}")
CMake: Split QtBuild.cmake into smaller files QtBuild.cmake is huge. Split it. Move module, plugin, tools, executables and test related functions out of QtBuild.cmake into separate files. Do the same for many other things too. An additional requirement is that all the new Helpers files only define functions and macros. No global variable definitions are allowed, nor execution of commands with side effects. Some notes: qt_install_qml_files is removed because it's dead code. Some functions still need to be figured out, because they are interspersed and depend on various global state assignments. Task-number: QTBUG-86035 Change-Id: I21d79ff02eef923c202eb1000422888727cb0e2c Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-08-13 15:37:47 +00:00
endif()
Rename QT_DEPRECATED_WARNINGS_SINCE -> QT_WARN_DEPRECATED_UP_TO The new name describes the behavior in a better way. [ChangeLog][Build System] The QT_DEPRECATED_WARNINGS_SINCE macro is renamed to QT_WARN_DEPRECATED_UP_TO. The old name is deprecated, but is still recognized if it is defined during configuration and the new name is not defined. Fixes: QTBUG-104944 Change-Id: I320c033010dfab120db6922598454f95169657f7 Reviewed-by: Edward Welbourne <edward.welbourne@qt.io> Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
2022-08-16 10:52:42 +00:00
# QT_WARN_DEPRECATED_UP_TO controls the upper-bound of deprecation
# warnings that are emitted. E.g. if it is set to 0x060500 then all use of
# things deprecated in or before 6.5.0 will be warned against.
list(APPEND deprecations "QT_WARN_DEPRECATED_UP_TO=0x070000")
CMake: Fix deprecation defines not being set properly Instead of propagating the deprecation wranings, we compiled all code with -Ddeprecations. Change-Id: I0233ddc85bdbdcb93d366073b2cea5d47bdbe52a Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org> Reviewed-by: Cristian Adam <cristian.adam@qt.io>
2020-10-19 17:34:17 +00:00
set("${result}" "${deprecations}" PARENT_SCOPE)
CMake: Split QtBuild.cmake into smaller files QtBuild.cmake is huge. Split it. Move module, plugin, tools, executables and test related functions out of QtBuild.cmake into separate files. Do the same for many other things too. An additional requirement is that all the new Helpers files only define functions and macros. No global variable definitions are allowed, nor execution of commands with side effects. Some notes: qt_install_qml_files is removed because it's dead code. Some functions still need to be figured out, because they are interspersed and depend on various global state assignments. Task-number: QTBUG-86035 Change-Id: I21d79ff02eef923c202eb1000422888727cb0e2c Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-08-13 15:37:47 +00:00
endfunction()
Clear non-relevant CMake compiler flags Add internal function to cleanup compiler flags out of the CMAKE_xxx_FLAGS_xxx variables. Use introduced interface to clear the '/EHsc' flag for the MSVC compiler family. This adjusts the CMake behavior to the qmake one. Change the 'EXCEPTIONS' option handling in helper functions. Add ability to add enabling and disabling exception flags. Previously it was only possible to add disabling exception flags. Fixes: QTBUG-89952 Change-Id: I60d47660a97ae9b5a1d1f4107d352c9e97890144 Reviewed-by: Craig Scott <craig.scott@qt.io> Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2021-01-21 10:17:58 +00:00
# Sets the exceptions flags for the given target according to exceptions_on
function(qt_internal_set_exceptions_flags target exceptions_on)
set(_defs "")
set(_flag "")
if(exceptions_on)
if(MSVC)
set(_flag "/EHsc")
Fix build with clang-cl 1. Clang-CL can't recognize "/d2FH4" and it's causing lots of warnings when compiling. So don't apply it when building with clang-cl. 2. The definition of "FS_INFORMATION_CLASS" need to be visible to clang as well. Don't know why it was excluded originally. Pick-to: 6.3 Change-Id: I7b6e14999eea0ba1f0d73962ff03a35548f88a5a Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
2022-01-21 07:25:46 +00:00
if((MSVC_VERSION GREATER_EQUAL 1929) AND NOT CLANG)
cmake: add some note of MSVC undocumented flag I added this flag but only explained what it is in the original commit message, it may confuse future code readers without some inline comments. So add some comments to avoid such issue. Change-Id: I6a3c0f53c2bc58646cc70e45cacb1d0e40656c0a Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
2023-10-07 06:03:03 +00:00
# Use the undocumented compiler flag to make our binary smaller on x64.
# https://devblogs.microsoft.com/cppblog/making-cpp-exception-handling-smaller-x64/
# NOTE: It seems we'll use this new exception handling model unconditionally without
# this hack since some unknown MSVC version.
MSVC: Use dubbed FH4 to make C++ exception handling smaller Visual Studio 2019 introduced the dubbed FH4 feature which can make C++ exception handling smaller on x64. According to the article [1], it's enabled by default for UWP applications, and Microsoft also use it in their own widely-known commercial products such as Office to reduce the binary size. So make use of this feature for Qt when possible, to get smaller binary. As a drive-by, add "/EHs-c-" explicitly to the flags when we want to disable C++ exception handling. [1] Official article that introduces dubbed FH4: https://devblogs.microsoft.com/cppblog/making-cpp-exception-handling-smaller-x64/ Change-Id: I2e3330de477f78372cf7903d0ef7a732b09552a9 Reviewed-by: Thiago Macieira <thiago.macieira@intel.com> Reviewed-by: André de la Rocha <andre.rocha@qt.io> Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org> Reviewed-by: Oliver Wolff <oliver.wolff@qt.io>
2021-12-02 09:14:00 +00:00
set(_flag ${_flag} "/d2FH4")
endif()
cmake: simplify exceptions handling code A little code simplification. And we can enable exception handling explicitly for GCC/Clang by adding "-fexceptions", it was missing in the old code, add it as well. [ChangeLog] [Build System] Qt explicitly pass -fexceptions now on non-MSVC toolchains, if exception handling is not disabled in CMake configure. Change-Id: Id9d61d3ee8b7d490f4a743e34e8be321af080be0 Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io> Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2023-10-07 06:17:30 +00:00
else()
set(_flag "-fexceptions")
Clear non-relevant CMake compiler flags Add internal function to cleanup compiler flags out of the CMAKE_xxx_FLAGS_xxx variables. Use introduced interface to clear the '/EHsc' flag for the MSVC compiler family. This adjusts the CMake behavior to the qmake one. Change the 'EXCEPTIONS' option handling in helper functions. Add ability to add enabling and disabling exception flags. Previously it was only possible to add disabling exception flags. Fixes: QTBUG-89952 Change-Id: I60d47660a97ae9b5a1d1f4107d352c9e97890144 Reviewed-by: Craig Scott <craig.scott@qt.io> Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2021-01-21 10:17:58 +00:00
endif()
else()
set(_defs "QT_NO_EXCEPTIONS")
cmake: simplify exceptions handling code A little code simplification. And we can enable exception handling explicitly for GCC/Clang by adding "-fexceptions", it was missing in the old code, add it as well. [ChangeLog] [Build System] Qt explicitly pass -fexceptions now on non-MSVC toolchains, if exception handling is not disabled in CMake configure. Change-Id: Id9d61d3ee8b7d490f4a743e34e8be321af080be0 Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io> Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2023-10-07 06:17:30 +00:00
if(MSVC)
MSVC: Use dubbed FH4 to make C++ exception handling smaller Visual Studio 2019 introduced the dubbed FH4 feature which can make C++ exception handling smaller on x64. According to the article [1], it's enabled by default for UWP applications, and Microsoft also use it in their own widely-known commercial products such as Office to reduce the binary size. So make use of this feature for Qt when possible, to get smaller binary. As a drive-by, add "/EHs-c-" explicitly to the flags when we want to disable C++ exception handling. [1] Official article that introduces dubbed FH4: https://devblogs.microsoft.com/cppblog/making-cpp-exception-handling-smaller-x64/ Change-Id: I2e3330de477f78372cf7903d0ef7a732b09552a9 Reviewed-by: Thiago Macieira <thiago.macieira@intel.com> Reviewed-by: André de la Rocha <andre.rocha@qt.io> Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org> Reviewed-by: Oliver Wolff <oliver.wolff@qt.io>
2021-12-02 09:14:00 +00:00
set(_flag "/EHs-c-" "/wd4530" "/wd4577")
cmake: simplify exceptions handling code A little code simplification. And we can enable exception handling explicitly for GCC/Clang by adding "-fexceptions", it was missing in the old code, add it as well. [ChangeLog] [Build System] Qt explicitly pass -fexceptions now on non-MSVC toolchains, if exception handling is not disabled in CMake configure. Change-Id: Id9d61d3ee8b7d490f4a743e34e8be321af080be0 Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io> Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2023-10-07 06:17:30 +00:00
else()
Clear non-relevant CMake compiler flags Add internal function to cleanup compiler flags out of the CMAKE_xxx_FLAGS_xxx variables. Use introduced interface to clear the '/EHsc' flag for the MSVC compiler family. This adjusts the CMake behavior to the qmake one. Change the 'EXCEPTIONS' option handling in helper functions. Add ability to add enabling and disabling exception flags. Previously it was only possible to add disabling exception flags. Fixes: QTBUG-89952 Change-Id: I60d47660a97ae9b5a1d1f4107d352c9e97890144 Reviewed-by: Craig Scott <craig.scott@qt.io> Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2021-01-21 10:17:58 +00:00
set(_flag "-fno-exceptions")
Improve clang-cl support for Qt6 1. clang-cl doesn't support "-fno-exceptions", it uses msvc's parameter. 2. some parameters supported by msvc are not supported by clang-cl and they are causing huge warning message flood, don't add them. 3. use correct optimize parameter for clang-cl. Change-Id: Idbadf139127143c5fa6c49068588cb26f47da7a2 Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
2020-11-02 04:36:14 +00:00
endif()
CMake: Split QtBuild.cmake into smaller files QtBuild.cmake is huge. Split it. Move module, plugin, tools, executables and test related functions out of QtBuild.cmake into separate files. Do the same for many other things too. An additional requirement is that all the new Helpers files only define functions and macros. No global variable definitions are allowed, nor execution of commands with side effects. Some notes: qt_install_qml_files is removed because it's dead code. Some functions still need to be figured out, because they are interspersed and depend on various global state assignments. Task-number: QTBUG-86035 Change-Id: I21d79ff02eef923c202eb1000422888727cb0e2c Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-08-13 15:37:47 +00:00
endif()
Clear non-relevant CMake compiler flags Add internal function to cleanup compiler flags out of the CMAKE_xxx_FLAGS_xxx variables. Use introduced interface to clear the '/EHsc' flag for the MSVC compiler family. This adjusts the CMake behavior to the qmake one. Change the 'EXCEPTIONS' option handling in helper functions. Add ability to add enabling and disabling exception flags. Previously it was only possible to add disabling exception flags. Fixes: QTBUG-89952 Change-Id: I60d47660a97ae9b5a1d1f4107d352c9e97890144 Reviewed-by: Craig Scott <craig.scott@qt.io> Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2021-01-21 10:17:58 +00:00
target_compile_definitions("${target}" PRIVATE ${_defs})
target_compile_options("${target}" PRIVATE ${_flag})
CMake: Split QtBuild.cmake into smaller files QtBuild.cmake is huge. Split it. Move module, plugin, tools, executables and test related functions out of QtBuild.cmake into separate files. Do the same for many other things too. An additional requirement is that all the new Helpers files only define functions and macros. No global variable definitions are allowed, nor execution of commands with side effects. Some notes: qt_install_qml_files is removed because it's dead code. Some functions still need to be figured out, because they are interspersed and depend on various global state assignments. Task-number: QTBUG-86035 Change-Id: I21d79ff02eef923c202eb1000422888727cb0e2c Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-08-13 15:37:47 +00:00
endfunction()
function(qt_skip_warnings_are_errors target)
get_target_property(target_type "${target}" TYPE)
if(target_type STREQUAL "INTERFACE_LIBRARY")
return()
endif()
set_target_properties("${target}" PROPERTIES QT_SKIP_WARNINGS_ARE_ERRORS ON)
endfunction()
function(qt_skip_warnings_are_errors_when_repo_unclean target)
if(QT_REPO_NOT_WARNINGS_CLEAN)
qt_skip_warnings_are_errors("${target}")
endif()
endfunction()
function(qt_disable_warnings target)
get_target_property(target_type "${target}" TYPE)
if(target_type STREQUAL "INTERFACE_LIBRARY")
return()
endif()
set_target_properties("${target}" PROPERTIES QT_COMPILE_OPTIONS_DISABLE_WARNINGS ON)
endfunction()
function(qt_set_symbol_visibility_preset target value)
get_target_property(target_type "${target}" TYPE)
if(target_type STREQUAL "INTERFACE_LIBRARY")
return()
endif()
set_target_properties("${target}" PROPERTIES C_VISIBILITY_PRESET "${value}")
set_target_properties("${target}" PROPERTIES CXX_VISIBILITY_PRESET "${value}")
set_target_properties("${target}" PROPERTIES OBJC_VISIBILITY_PRESET "${value}")
set_target_properties("${target}" PROPERTIES OBJCXX_VISIBILITY_PRESET "${value}")
endfunction()
function(qt_set_symbol_visibility_hidden target)
qt_set_symbol_visibility_preset("${target}" "hidden")
endfunction()
function(qt_set_language_standards)
## Use the latest standard the compiler supports (same as qt_common.prf)
C++23/c++2b support Change-Id: I33b2a48312ae94e3d5ebb4097e50c4953e14d533 Reviewed-by: Jörg Bornemann <joerg.bornemann@qt.io>
2022-02-04 10:52:55 +00:00
if (QT_FEATURE_cxx2b)
set(CMAKE_CXX_STANDARD 23 PARENT_SCOPE)
elseif (QT_FEATURE_cxx20)
CMake: Split QtBuild.cmake into smaller files QtBuild.cmake is huge. Split it. Move module, plugin, tools, executables and test related functions out of QtBuild.cmake into separate files. Do the same for many other things too. An additional requirement is that all the new Helpers files only define functions and macros. No global variable definitions are allowed, nor execution of commands with side effects. Some notes: qt_install_qml_files is removed because it's dead code. Some functions still need to be figured out, because they are interspersed and depend on various global state assignments. Task-number: QTBUG-86035 Change-Id: I21d79ff02eef923c202eb1000422888727cb0e2c Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-08-13 15:37:47 +00:00
set(CMAKE_CXX_STANDARD 20 PARENT_SCOPE)
else()
set(CMAKE_CXX_STANDARD 17 PARENT_SCOPE)
endif()
CMake: Mark chosen C++ standard as required for building Qt Change-Id: I7d3187e74251059c7536f3e34d8c87f54c673146 Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org> Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
2022-06-03 05:58:59 +00:00
set(CMAKE_CXX_STANDARD_REQUIRED ON PARENT_SCOPE)
CMake: Split QtBuild.cmake into smaller files QtBuild.cmake is huge. Split it. Move module, plugin, tools, executables and test related functions out of QtBuild.cmake into separate files. Do the same for many other things too. An additional requirement is that all the new Helpers files only define functions and macros. No global variable definitions are allowed, nor execution of commands with side effects. Some notes: qt_install_qml_files is removed because it's dead code. Some functions still need to be figured out, because they are interspersed and depend on various global state assignments. Task-number: QTBUG-86035 Change-Id: I21d79ff02eef923c202eb1000422888727cb0e2c Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-08-13 15:37:47 +00:00
CMake: up the minimum C version to C11 We're in 2022. A 11-year-old C standard probably suffices, especially since we require C++17 anyway. Pick-to: 6.4 Change-Id: Ibcde9b9795ad42ac9978fffd16f3555327097ded Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
2022-05-28 17:35:03 +00:00
set(CMAKE_C_STANDARD 11 PARENT_SCOPE)
set(CMAKE_C_STANDARD_REQUIRED ON PARENT_SCOPE)
CMake: Split QtBuild.cmake into smaller files QtBuild.cmake is huge. Split it. Move module, plugin, tools, executables and test related functions out of QtBuild.cmake into separate files. Do the same for many other things too. An additional requirement is that all the new Helpers files only define functions and macros. No global variable definitions are allowed, nor execution of commands with side effects. Some notes: qt_install_qml_files is removed because it's dead code. Some functions still need to be figured out, because they are interspersed and depend on various global state assignments. Task-number: QTBUG-86035 Change-Id: I21d79ff02eef923c202eb1000422888727cb0e2c Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-08-13 15:37:47 +00:00
endfunction()
function(qt_set_language_standards_interface_compile_features target)
# Regardless of which C++ standard is used to build Qt itself, require C++17 when building
# Qt applications using CMake (because the Qt header files use C++17 features).
set(cpp_feature "cxx_std_17")
target_compile_features("${target}" INTERFACE ${cpp_feature})
endfunction()
Build Qt (and client apps using it) with /permissive- *Not* using /permissive- exposes Qt and client apps to interesting bugs and/or build failures, (e.g. QTBUG-87225, or 19b5520abfb5f66d4b83c7a18cc72d68673d098a). We demand strict conformance by any other compiler, it's time to demand it from MSVC too. The Windows headers themselves are clean starting from the Windows Fall Creators SDK (10.0.16299.0), and moreover Qt 6 will drop WinRT; therefore, the comment in the mkspecs does not apply any more. Since /permissive- implies /Zc:referenceBinding, drop that option. The other implied options are set on MSVC < 2017, but I leave them in to avoid tinkering with the fragile lists of C/C++ flags. Rename the CMake internal helper function to better describe what it does. Fixes: QTBUG-85633 Fixes: QTBUG-85637 Fixes: QTBUG-85635 Fixes: QTBUG-88244 Change-Id: Ie03fddb61aa066fdc14b7231c22e7108b4a02fbb Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io> Reviewed-by: Kai Koehne <kai.koehne@qt.io> Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
2020-10-08 20:29:04 +00:00
function(qt_set_msvc_cplusplus_options target visibility)
CMake: Split QtBuild.cmake into smaller files QtBuild.cmake is huge. Split it. Move module, plugin, tools, executables and test related functions out of QtBuild.cmake into separate files. Do the same for many other things too. An additional requirement is that all the new Helpers files only define functions and macros. No global variable definitions are allowed, nor execution of commands with side effects. Some notes: qt_install_qml_files is removed because it's dead code. Some functions still need to be figured out, because they are interspersed and depend on various global state assignments. Task-number: QTBUG-86035 Change-Id: I21d79ff02eef923c202eb1000422888727cb0e2c Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-08-13 15:37:47 +00:00
# For MSVC we need to explicitly pass -Zc:__cplusplus to get correct __cplusplus.
# Check qt_config_compile_test for more info.
CMake: Add support for building with clang-cl qmake had support for building with clang-cl as the win32-clang-msvc mkspec. Task-number: QTBUG-89642 Task-number: QTBUG-88081 Change-Id: I0709c289f90fedb121620d1e67ef841602219816 Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
2021-06-03 10:42:07 +00:00
if(MSVC AND MSVC_VERSION GREATER_EQUAL 1913)
CMake: Add C/C++ compiler options to C/C++ sources only When compiling CUDA sources in a user project, the Qt6::Platform target would pull in C/C++ related compiler flags, leading to compiler errors. Make sure that we only add those flags to C/C++ source files. Pick-to: 6.2 6.3 Fixes: QTBUG-99548 Change-Id: Idbccd65fe8f66abd1da3fce95e563065d1ed3cc6 Reviewed-by: Kai Koehne <kai.koehne@qt.io>
2022-01-04 09:30:04 +00:00
set(flags "-Zc:__cplusplus" "-permissive-")
Make sure that -Zc:__cplusplus, and -permissive- only apply to MSVC As described in the bug report, in cases where host Qt is built using MSVC, these flags may leak to user projects if they are set to be configured by a different compiler, e.g., Clang. Pick-to: 6.5 Fixes: QTBUG-112737 Change-Id: Iad922e24cc7e7f835e08ed37271dfbedc6e38dbe Reviewed-by: Alexey Edelev <alexey.edelev@qt.io> Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
2023-05-08 14:19:35 +00:00
target_compile_options("${target}" ${visibility}
"$<$<AND:$<CXX_COMPILER_ID:MSVC>,$<COMPILE_LANGUAGE:CXX>>:${flags}>")
CMake: Split QtBuild.cmake into smaller files QtBuild.cmake is huge. Split it. Move module, plugin, tools, executables and test related functions out of QtBuild.cmake into separate files. Do the same for many other things too. An additional requirement is that all the new Helpers files only define functions and macros. No global variable definitions are allowed, nor execution of commands with side effects. Some notes: qt_install_qml_files is removed because it's dead code. Some functions still need to be figured out, because they are interspersed and depend on various global state assignments. Task-number: QTBUG-86035 Change-Id: I21d79ff02eef923c202eb1000422888727cb0e2c Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-08-13 15:37:47 +00:00
endif()
endfunction()
function(qt_enable_utf8_sources target)
set(utf8_flags "")
if(MSVC)
Make sure that -utf-8 flag only apply to MSVC INTERFACE scope propagates the '-utf-8' flag to the target dependencies. So if Qt is built using MSVC, but the depending targets use different compiler this flag will break the compilation. Guard the flag using genex. Amends e3cc2487ce63cae42d8054b38fdb54abe8545007 Pick-to: 6.5 Task-number: QTBUG-112737 Change-Id: Ie0576667108820dd61035debfc1fcc030ef3536a Reviewed-by: Lars Schmertmann <lars.schmertmann@governikus.de> Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2023-05-12 14:23:15 +00:00
list(APPEND utf8_flags "$<$<CXX_COMPILER_ID:MSVC>:-utf-8>")
CMake: Split QtBuild.cmake into smaller files QtBuild.cmake is huge. Split it. Move module, plugin, tools, executables and test related functions out of QtBuild.cmake into separate files. Do the same for many other things too. An additional requirement is that all the new Helpers files only define functions and macros. No global variable definitions are allowed, nor execution of commands with side effects. Some notes: qt_install_qml_files is removed because it's dead code. Some functions still need to be figured out, because they are interspersed and depend on various global state assignments. Task-number: QTBUG-86035 Change-Id: I21d79ff02eef923c202eb1000422888727cb0e2c Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-08-13 15:37:47 +00:00
endif()
if(utf8_flags)
# Allow opting out by specifying the QT_NO_UTF8_SOURCE target property.
CMake: Add C/C++ compiler options to C/C++ sources only When compiling CUDA sources in a user project, the Qt6::Platform target would pull in C/C++ related compiler flags, leading to compiler errors. Make sure that we only add those flags to C/C++ source files. Pick-to: 6.2 6.3 Fixes: QTBUG-99548 Change-Id: Idbccd65fe8f66abd1da3fce95e563065d1ed3cc6 Reviewed-by: Kai Koehne <kai.koehne@qt.io>
2022-01-04 09:30:04 +00:00
set(opt_out_condition "$<NOT:$<BOOL:$<TARGET_PROPERTY:QT_NO_UTF8_SOURCE>>>")
# Only set the compiler option for C and C++.
set(language_condition "$<COMPILE_LANGUAGE:C,CXX>")
# Compose the full condition.
set(genex_condition "$<AND:${opt_out_condition},${language_condition}>")
CMake: Split QtBuild.cmake into smaller files QtBuild.cmake is huge. Split it. Move module, plugin, tools, executables and test related functions out of QtBuild.cmake into separate files. Do the same for many other things too. An additional requirement is that all the new Helpers files only define functions and macros. No global variable definitions are allowed, nor execution of commands with side effects. Some notes: qt_install_qml_files is removed because it's dead code. Some functions still need to be figured out, because they are interspersed and depend on various global state assignments. Task-number: QTBUG-86035 Change-Id: I21d79ff02eef923c202eb1000422888727cb0e2c Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-08-13 15:37:47 +00:00
set(utf8_flags "$<${genex_condition}:${utf8_flags}>")
target_compile_options("${target}" INTERFACE "${utf8_flags}")
endif()
endfunction()
CMake: Adjust compiler flag optimizations to qmake mkspec ones There are inconsistencies in the default optimization flags added by CMake across configurations like Release and RelWithDebInfo. In particular Release uses -O3, whereas RelWithDebInfo uses -O2, as well as usage of /INCREMENTAL in release configs with MSVC, etc. To make sure that the Qt 6 binaries built with CMake are consistent across configs, as well as consistent with the flags we used when building Qt 5 with qmake, add a horrible search and replace mechanism to replaces the CMake flags with what our mkspecs indicate to use. Ideally this would be done by providing custom CMake toolchain files for each platform we support, and we might revisit that later if the need really arises. To implement the replacing, we first need the flags that should be added. Port the QMAKE_CFLAGS_OPTIMIZE variables to CMake, which is done in QtCompilerOptimization.cmake. Then a new function called qt_internal_set_up_config_optimizations_like_in_qmake will look for any kind of optimization flags set in the CMAKE_<LANG>_FLAGS_<CONFIG> style variables, remove them, and add the appropriate flags that qmake mkspecs provide. On some platforms (like Windows MSVC) the function also alters the linker CMAKE_${TYPE}_LINKER_FLAGS_<CONFIG> style variables. The mechanism allows opting out of this replacing by setting the QT_USE_DEFAULT_CMAKE_OPTIMIZATION_FLAGS value. It also allows opting into removal of flags for custom configs by providing QT_ADDITIONAL_OPTIMIZATION_FLAG_CONFIGS. It's only removal, because we wouldn't know what kind of config it is, and thus what flags to add. The currently modified configs are: Release, RelWithDebInfo, MinSizeRel, Debug aka the usual default CMake provided ones. The mechanism is only applied to C-like languages. ASM is not handled to be on the safe side due to not knowing what kind of compiler flags the platform assembler might take. It's also important to skip RC on MSVC platforms. Task-number: QTBUG-85992 Change-Id: I3712d5cd5a34fceab54f56a6fa46b4e678952362 Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org> Reviewed-by: Cristian Adam <cristian.adam@qt.io>
2020-08-17 15:32:28 +00:00
Enable UNICODE for all Qt targets and Qt consumers by default After discussion we decided to opt-out the UNICODE definintion behavior. To disable UNICODE in user projects the qt6_disable_unicode_defines function could be used. Amends 5b64e5950cf984abb7d47e0802bcb4b5a21e06fa [ChangeLog][CMake] Enables the UNICODE and _UNICODE definitions on WIN32 platforms by default for all cmake projects to reflect the qmake behavior. Use qt6_disable_unicode_defines function to disable the default unicode definitions. Pick-to: 6.1 Fixes: QTBUG-93895 Change-Id: Id70ff7dcf8c74f660ec851f8b950e1e3b94d9fb4 Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
2021-05-21 08:22:04 +00:00
function(qt_internal_enable_unicode_defines)
if(WIN32)
set(no_unicode_condition
"$<NOT:$<BOOL:$<TARGET_PROPERTY:QT_NO_UNICODE_DEFINES>>>")
target_compile_definitions(Platform
INTERFACE "$<${no_unicode_condition}:UNICODE;_UNICODE>")
endif()
endfunction()
CMake: Refactor optimization flag handling and add optimize_full Introduce a bunch of helper functions to manipulate compiler flags and linker flags for the CMAKE_<LANG>_FLAGS_<CONFIG> and CMAKE_<LINK_TYPE>_LINKER_FLAGS_<CONFIG> CMake variables. These variables can be assigned and modified either in the cache or for a specific subdirectory scope, which will apply the flags only to targets in that scope. Add qt_internal_add_optimize_full_flags() function which mimics qmake's CONFIG += optimize_full behavior. Calling it will force usage of the '-O3' optimization flag on supported platforms (falling back '-O2' where not supported). Use the function for the Core and Gui subdirectories, to enable full optimization for the respective Qt modules as it is done in the qmake projects. To ensure that the global qmake-like compiler flags are assigned eveywhere, qt_internal_set_up_config_optimizations_like_in_qmake() needs to be called after Qt global features like optimize_size and optimize_full are available. This means that qtbase and its standalone tests need some special handling in regards to when to call that function. Task-number: QTBUG-86866 Change-Id: Ic7ac23de0265561cb06a0ba55089b6c0d3347441 Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-10-02 13:38:47 +00:00
# Saves the list of known optimization flags for the current compiler in out_var.
#
# Mostly used for removing them before adding new ones.
CMake: Adjust compiler flag optimizations to qmake mkspec ones There are inconsistencies in the default optimization flags added by CMake across configurations like Release and RelWithDebInfo. In particular Release uses -O3, whereas RelWithDebInfo uses -O2, as well as usage of /INCREMENTAL in release configs with MSVC, etc. To make sure that the Qt 6 binaries built with CMake are consistent across configs, as well as consistent with the flags we used when building Qt 5 with qmake, add a horrible search and replace mechanism to replaces the CMake flags with what our mkspecs indicate to use. Ideally this would be done by providing custom CMake toolchain files for each platform we support, and we might revisit that later if the need really arises. To implement the replacing, we first need the flags that should be added. Port the QMAKE_CFLAGS_OPTIMIZE variables to CMake, which is done in QtCompilerOptimization.cmake. Then a new function called qt_internal_set_up_config_optimizations_like_in_qmake will look for any kind of optimization flags set in the CMAKE_<LANG>_FLAGS_<CONFIG> style variables, remove them, and add the appropriate flags that qmake mkspecs provide. On some platforms (like Windows MSVC) the function also alters the linker CMAKE_${TYPE}_LINKER_FLAGS_<CONFIG> style variables. The mechanism allows opting out of this replacing by setting the QT_USE_DEFAULT_CMAKE_OPTIMIZATION_FLAGS value. It also allows opting into removal of flags for custom configs by providing QT_ADDITIONAL_OPTIMIZATION_FLAG_CONFIGS. It's only removal, because we wouldn't know what kind of config it is, and thus what flags to add. The currently modified configs are: Release, RelWithDebInfo, MinSizeRel, Debug aka the usual default CMake provided ones. The mechanism is only applied to C-like languages. ASM is not handled to be on the safe side due to not knowing what kind of compiler flags the platform assembler might take. It's also important to skip RC on MSVC platforms. Task-number: QTBUG-85992 Change-Id: I3712d5cd5a34fceab54f56a6fa46b4e678952362 Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org> Reviewed-by: Cristian Adam <cristian.adam@qt.io>
2020-08-17 15:32:28 +00:00
function(qt_internal_get_all_possible_optimization_flag_values out_var)
set(flag_values "")
set(vars QT_CFLAGS_OPTIMIZE QT_CFLAGS_OPTIMIZE_FULL
QT_CFLAGS_OPTIMIZE_DEBUG QT_CFLAGS_OPTIMIZE_SIZE)
foreach(optimize_var ${vars})
set(value "${${optimize_var}}")
if(value)
list(APPEND flag_values "${value}")
endif()
endforeach()
# Additional flag values which might not be used in qmake mkspecs, but might be set by CMake,
# aka flags that are recognized by the compile which we might want to remove.
if(QT_CFLAGS_OPTIMIZE_VALID_VALUES)
list(APPEND flag_values ${QT_CFLAGS_OPTIMIZE_VALID_VALUES})
endif()
set("${out_var}" "${flag_values}" PARENT_SCOPE)
endfunction()
CMake: Refactor optimization flag handling and add optimize_full Introduce a bunch of helper functions to manipulate compiler flags and linker flags for the CMAKE_<LANG>_FLAGS_<CONFIG> and CMAKE_<LINK_TYPE>_LINKER_FLAGS_<CONFIG> CMake variables. These variables can be assigned and modified either in the cache or for a specific subdirectory scope, which will apply the flags only to targets in that scope. Add qt_internal_add_optimize_full_flags() function which mimics qmake's CONFIG += optimize_full behavior. Calling it will force usage of the '-O3' optimization flag on supported platforms (falling back '-O2' where not supported). Use the function for the Core and Gui subdirectories, to enable full optimization for the respective Qt modules as it is done in the qmake projects. To ensure that the global qmake-like compiler flags are assigned eveywhere, qt_internal_set_up_config_optimizations_like_in_qmake() needs to be called after Qt global features like optimize_size and optimize_full are available. This means that qtbase and its standalone tests need some special handling in regards to when to call that function. Task-number: QTBUG-86866 Change-Id: Ic7ac23de0265561cb06a0ba55089b6c0d3347441 Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-10-02 13:38:47 +00:00
# Return's the current compiler's optimize_full flags if available.
# Otherwise returns the regular optimization level flag.
function(qt_internal_get_optimize_full_flags out_var)
set(optimize_full_flags "${QT_CFLAGS_OPTIMIZE_FULL}")
if(NOT optimize_full_flags)
set(optimize_full_flags "${QT_CFLAGS_OPTIMIZE}")
endif()
set(${out_var} "${optimize_full_flags}" PARENT_SCOPE)
endfunction()
# Prints the compiler and linker flags for each configuration, language and target type.
#
# Usually it would print the cache variables, but one may also override the variables
# in a specific directory scope, so this is useful for debugging.
#
# Basically dumps either scoped or cached
# CMAKE_<LANG>_FLAGS_CONFIG> and CMAKE_<TYPE>_LINKER_FLAGS_<CONFIG> variables.
function(qt_internal_print_optimization_flags_values)
qt_internal_get_enabled_languages_for_flag_manipulation(languages)
qt_internal_get_configs_for_flag_manipulation(configs)
qt_internal_get_target_link_types_for_flag_manipulation(target_link_types)
qt_internal_print_optimization_flags_values_helper(
"${languages}" "${configs}" "${target_link_types}")
endfunction()
# Helper function for printing the optimization flags.
function(qt_internal_print_optimization_flags_values_helper languages configs target_link_types)
CMake: Adjust compiler flag optimizations to qmake mkspec ones There are inconsistencies in the default optimization flags added by CMake across configurations like Release and RelWithDebInfo. In particular Release uses -O3, whereas RelWithDebInfo uses -O2, as well as usage of /INCREMENTAL in release configs with MSVC, etc. To make sure that the Qt 6 binaries built with CMake are consistent across configs, as well as consistent with the flags we used when building Qt 5 with qmake, add a horrible search and replace mechanism to replaces the CMake flags with what our mkspecs indicate to use. Ideally this would be done by providing custom CMake toolchain files for each platform we support, and we might revisit that later if the need really arises. To implement the replacing, we first need the flags that should be added. Port the QMAKE_CFLAGS_OPTIMIZE variables to CMake, which is done in QtCompilerOptimization.cmake. Then a new function called qt_internal_set_up_config_optimizations_like_in_qmake will look for any kind of optimization flags set in the CMAKE_<LANG>_FLAGS_<CONFIG> style variables, remove them, and add the appropriate flags that qmake mkspecs provide. On some platforms (like Windows MSVC) the function also alters the linker CMAKE_${TYPE}_LINKER_FLAGS_<CONFIG> style variables. The mechanism allows opting out of this replacing by setting the QT_USE_DEFAULT_CMAKE_OPTIMIZATION_FLAGS value. It also allows opting into removal of flags for custom configs by providing QT_ADDITIONAL_OPTIMIZATION_FLAG_CONFIGS. It's only removal, because we wouldn't know what kind of config it is, and thus what flags to add. The currently modified configs are: Release, RelWithDebInfo, MinSizeRel, Debug aka the usual default CMake provided ones. The mechanism is only applied to C-like languages. ASM is not handled to be on the safe side due to not knowing what kind of compiler flags the platform assembler might take. It's also important to skip RC on MSVC platforms. Task-number: QTBUG-85992 Change-Id: I3712d5cd5a34fceab54f56a6fa46b4e678952362 Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org> Reviewed-by: Cristian Adam <cristian.adam@qt.io>
2020-08-17 15:32:28 +00:00
foreach(lang ${languages})
set(flag_var_name "CMAKE_${lang}_FLAGS")
message(STATUS "${flag_var_name}: ${${flag_var_name}}")
foreach(config ${configs})
set(flag_var_name "CMAKE_${lang}_FLAGS_${config}")
message(STATUS "${flag_var_name}: ${${flag_var_name}}")
endforeach()
endforeach()
foreach(t ${target_link_types})
set(flag_var_name "CMAKE_${t}_LINKER_FLAGS")
message(STATUS "${flag_var_name}: ${${flag_var_name}}")
foreach(config ${configs})
set(flag_var_name "CMAKE_${t}_LINKER_FLAGS_${config}")
message(STATUS "${flag_var_name}: ${${flag_var_name}}")
endforeach()
endforeach()
endfunction()
CMake: Refactor optimization flag handling and add optimize_full Introduce a bunch of helper functions to manipulate compiler flags and linker flags for the CMAKE_<LANG>_FLAGS_<CONFIG> and CMAKE_<LINK_TYPE>_LINKER_FLAGS_<CONFIG> CMake variables. These variables can be assigned and modified either in the cache or for a specific subdirectory scope, which will apply the flags only to targets in that scope. Add qt_internal_add_optimize_full_flags() function which mimics qmake's CONFIG += optimize_full behavior. Calling it will force usage of the '-O3' optimization flag on supported platforms (falling back '-O2' where not supported). Use the function for the Core and Gui subdirectories, to enable full optimization for the respective Qt modules as it is done in the qmake projects. To ensure that the global qmake-like compiler flags are assigned eveywhere, qt_internal_set_up_config_optimizations_like_in_qmake() needs to be called after Qt global features like optimize_size and optimize_full are available. This means that qtbase and its standalone tests need some special handling in regards to when to call that function. Task-number: QTBUG-86866 Change-Id: Ic7ac23de0265561cb06a0ba55089b6c0d3347441 Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-10-02 13:38:47 +00:00
# Saves the list of configs for which flag manipulation will occur.
function(qt_internal_get_configs_for_flag_manipulation out_var)
set(configs RELEASE RELWITHDEBINFO MINSIZEREL DEBUG)
# Opt into additional non-standard configs for flag removal only.
if(QT_ADDITIONAL_OPTIMIZATION_FLAG_CONFIGS)
list(APPEND configs ${QT_ADDITIONAL_OPTIMIZATION_FLAG_CONFIGS})
CMake: Adjust compiler flag optimizations to qmake mkspec ones There are inconsistencies in the default optimization flags added by CMake across configurations like Release and RelWithDebInfo. In particular Release uses -O3, whereas RelWithDebInfo uses -O2, as well as usage of /INCREMENTAL in release configs with MSVC, etc. To make sure that the Qt 6 binaries built with CMake are consistent across configs, as well as consistent with the flags we used when building Qt 5 with qmake, add a horrible search and replace mechanism to replaces the CMake flags with what our mkspecs indicate to use. Ideally this would be done by providing custom CMake toolchain files for each platform we support, and we might revisit that later if the need really arises. To implement the replacing, we first need the flags that should be added. Port the QMAKE_CFLAGS_OPTIMIZE variables to CMake, which is done in QtCompilerOptimization.cmake. Then a new function called qt_internal_set_up_config_optimizations_like_in_qmake will look for any kind of optimization flags set in the CMAKE_<LANG>_FLAGS_<CONFIG> style variables, remove them, and add the appropriate flags that qmake mkspecs provide. On some platforms (like Windows MSVC) the function also alters the linker CMAKE_${TYPE}_LINKER_FLAGS_<CONFIG> style variables. The mechanism allows opting out of this replacing by setting the QT_USE_DEFAULT_CMAKE_OPTIMIZATION_FLAGS value. It also allows opting into removal of flags for custom configs by providing QT_ADDITIONAL_OPTIMIZATION_FLAG_CONFIGS. It's only removal, because we wouldn't know what kind of config it is, and thus what flags to add. The currently modified configs are: Release, RelWithDebInfo, MinSizeRel, Debug aka the usual default CMake provided ones. The mechanism is only applied to C-like languages. ASM is not handled to be on the safe side due to not knowing what kind of compiler flags the platform assembler might take. It's also important to skip RC on MSVC platforms. Task-number: QTBUG-85992 Change-Id: I3712d5cd5a34fceab54f56a6fa46b4e678952362 Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org> Reviewed-by: Cristian Adam <cristian.adam@qt.io>
2020-08-17 15:32:28 +00:00
endif()
CMake: Refactor optimization flag handling and add optimize_full Introduce a bunch of helper functions to manipulate compiler flags and linker flags for the CMAKE_<LANG>_FLAGS_<CONFIG> and CMAKE_<LINK_TYPE>_LINKER_FLAGS_<CONFIG> CMake variables. These variables can be assigned and modified either in the cache or for a specific subdirectory scope, which will apply the flags only to targets in that scope. Add qt_internal_add_optimize_full_flags() function which mimics qmake's CONFIG += optimize_full behavior. Calling it will force usage of the '-O3' optimization flag on supported platforms (falling back '-O2' where not supported). Use the function for the Core and Gui subdirectories, to enable full optimization for the respective Qt modules as it is done in the qmake projects. To ensure that the global qmake-like compiler flags are assigned eveywhere, qt_internal_set_up_config_optimizations_like_in_qmake() needs to be called after Qt global features like optimize_size and optimize_full are available. This means that qtbase and its standalone tests need some special handling in regards to when to call that function. Task-number: QTBUG-86866 Change-Id: Ic7ac23de0265561cb06a0ba55089b6c0d3347441 Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-10-02 13:38:47 +00:00
set(${out_var} "${configs}" PARENT_SCOPE)
endfunction()
# Saves the list of target link types for which flag manipulation will occur.
function(qt_internal_get_target_link_types_for_flag_manipulation out_var)
set(target_link_types EXE SHARED MODULE STATIC)
set(${out_var} "${target_link_types}" PARENT_SCOPE)
endfunction()
# Saves list of enabled languages for which it is safe to manipulate compilation flags.
function(qt_internal_get_enabled_languages_for_flag_manipulation out_var)
CMake: Adjust compiler flag optimizations to qmake mkspec ones There are inconsistencies in the default optimization flags added by CMake across configurations like Release and RelWithDebInfo. In particular Release uses -O3, whereas RelWithDebInfo uses -O2, as well as usage of /INCREMENTAL in release configs with MSVC, etc. To make sure that the Qt 6 binaries built with CMake are consistent across configs, as well as consistent with the flags we used when building Qt 5 with qmake, add a horrible search and replace mechanism to replaces the CMake flags with what our mkspecs indicate to use. Ideally this would be done by providing custom CMake toolchain files for each platform we support, and we might revisit that later if the need really arises. To implement the replacing, we first need the flags that should be added. Port the QMAKE_CFLAGS_OPTIMIZE variables to CMake, which is done in QtCompilerOptimization.cmake. Then a new function called qt_internal_set_up_config_optimizations_like_in_qmake will look for any kind of optimization flags set in the CMAKE_<LANG>_FLAGS_<CONFIG> style variables, remove them, and add the appropriate flags that qmake mkspecs provide. On some platforms (like Windows MSVC) the function also alters the linker CMAKE_${TYPE}_LINKER_FLAGS_<CONFIG> style variables. The mechanism allows opting out of this replacing by setting the QT_USE_DEFAULT_CMAKE_OPTIMIZATION_FLAGS value. It also allows opting into removal of flags for custom configs by providing QT_ADDITIONAL_OPTIMIZATION_FLAG_CONFIGS. It's only removal, because we wouldn't know what kind of config it is, and thus what flags to add. The currently modified configs are: Release, RelWithDebInfo, MinSizeRel, Debug aka the usual default CMake provided ones. The mechanism is only applied to C-like languages. ASM is not handled to be on the safe side due to not knowing what kind of compiler flags the platform assembler might take. It's also important to skip RC on MSVC platforms. Task-number: QTBUG-85992 Change-Id: I3712d5cd5a34fceab54f56a6fa46b4e678952362 Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org> Reviewed-by: Cristian Adam <cristian.adam@qt.io>
2020-08-17 15:32:28 +00:00
# Limit flag modification to c-like code. We don't want to accidentally add incompatible
# flags to MSVC's RC or Swift.
Ensure consistent flags for ASM language and MODULE targets By not including ASM in the languages for which compiler flags are adjusted, we can end up with inconsistent optimization and debug settings between different languages. Make sure they are treated the same. Similarly, linker flags for MODULE target types should also have their debugging and incremental linking flags adjusted, just like EXE and SHARED targets. When building with MSVC, MODULE targets were having the relevant flags stripped, but not then replaced with the desired ones, leaving them empty in typical scenarios. This would primarily affect plugins, which are built as MODULE rather than SHARED libraries. Fixes: QTBUG-90237 Change-Id: I648ea74be1654d24cbecc592ce0ca4b59b2ae839 Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io> Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
2021-01-14 05:47:19 +00:00
set(languages_to_process ASM C CXX OBJC OBJCXX)
CMake: Adjust compiler flag optimizations to qmake mkspec ones There are inconsistencies in the default optimization flags added by CMake across configurations like Release and RelWithDebInfo. In particular Release uses -O3, whereas RelWithDebInfo uses -O2, as well as usage of /INCREMENTAL in release configs with MSVC, etc. To make sure that the Qt 6 binaries built with CMake are consistent across configs, as well as consistent with the flags we used when building Qt 5 with qmake, add a horrible search and replace mechanism to replaces the CMake flags with what our mkspecs indicate to use. Ideally this would be done by providing custom CMake toolchain files for each platform we support, and we might revisit that later if the need really arises. To implement the replacing, we first need the flags that should be added. Port the QMAKE_CFLAGS_OPTIMIZE variables to CMake, which is done in QtCompilerOptimization.cmake. Then a new function called qt_internal_set_up_config_optimizations_like_in_qmake will look for any kind of optimization flags set in the CMAKE_<LANG>_FLAGS_<CONFIG> style variables, remove them, and add the appropriate flags that qmake mkspecs provide. On some platforms (like Windows MSVC) the function also alters the linker CMAKE_${TYPE}_LINKER_FLAGS_<CONFIG> style variables. The mechanism allows opting out of this replacing by setting the QT_USE_DEFAULT_CMAKE_OPTIMIZATION_FLAGS value. It also allows opting into removal of flags for custom configs by providing QT_ADDITIONAL_OPTIMIZATION_FLAG_CONFIGS. It's only removal, because we wouldn't know what kind of config it is, and thus what flags to add. The currently modified configs are: Release, RelWithDebInfo, MinSizeRel, Debug aka the usual default CMake provided ones. The mechanism is only applied to C-like languages. ASM is not handled to be on the safe side due to not knowing what kind of compiler flags the platform assembler might take. It's also important to skip RC on MSVC platforms. Task-number: QTBUG-85992 Change-Id: I3712d5cd5a34fceab54f56a6fa46b4e678952362 Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org> Reviewed-by: Cristian Adam <cristian.adam@qt.io>
2020-08-17 15:32:28 +00:00
get_property(globally_enabled_languages GLOBAL PROPERTY ENABLED_LANGUAGES)
set(enabled_languages "")
foreach(lang ${languages_to_process})
if(lang IN_LIST globally_enabled_languages)
list(APPEND enabled_languages "${lang}")
endif()
endforeach()
CMake: Refactor optimization flag handling and add optimize_full Introduce a bunch of helper functions to manipulate compiler flags and linker flags for the CMAKE_<LANG>_FLAGS_<CONFIG> and CMAKE_<LINK_TYPE>_LINKER_FLAGS_<CONFIG> CMake variables. These variables can be assigned and modified either in the cache or for a specific subdirectory scope, which will apply the flags only to targets in that scope. Add qt_internal_add_optimize_full_flags() function which mimics qmake's CONFIG += optimize_full behavior. Calling it will force usage of the '-O3' optimization flag on supported platforms (falling back '-O2' where not supported). Use the function for the Core and Gui subdirectories, to enable full optimization for the respective Qt modules as it is done in the qmake projects. To ensure that the global qmake-like compiler flags are assigned eveywhere, qt_internal_set_up_config_optimizations_like_in_qmake() needs to be called after Qt global features like optimize_size and optimize_full are available. This means that qtbase and its standalone tests need some special handling in regards to when to call that function. Task-number: QTBUG-86866 Change-Id: Ic7ac23de0265561cb06a0ba55089b6c0d3347441 Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-10-02 13:38:47 +00:00
set(${out_var} "${enabled_languages}" PARENT_SCOPE)
endfunction()
CMake: Adjust compiler flag optimizations to qmake mkspec ones There are inconsistencies in the default optimization flags added by CMake across configurations like Release and RelWithDebInfo. In particular Release uses -O3, whereas RelWithDebInfo uses -O2, as well as usage of /INCREMENTAL in release configs with MSVC, etc. To make sure that the Qt 6 binaries built with CMake are consistent across configs, as well as consistent with the flags we used when building Qt 5 with qmake, add a horrible search and replace mechanism to replaces the CMake flags with what our mkspecs indicate to use. Ideally this would be done by providing custom CMake toolchain files for each platform we support, and we might revisit that later if the need really arises. To implement the replacing, we first need the flags that should be added. Port the QMAKE_CFLAGS_OPTIMIZE variables to CMake, which is done in QtCompilerOptimization.cmake. Then a new function called qt_internal_set_up_config_optimizations_like_in_qmake will look for any kind of optimization flags set in the CMAKE_<LANG>_FLAGS_<CONFIG> style variables, remove them, and add the appropriate flags that qmake mkspecs provide. On some platforms (like Windows MSVC) the function also alters the linker CMAKE_${TYPE}_LINKER_FLAGS_<CONFIG> style variables. The mechanism allows opting out of this replacing by setting the QT_USE_DEFAULT_CMAKE_OPTIMIZATION_FLAGS value. It also allows opting into removal of flags for custom configs by providing QT_ADDITIONAL_OPTIMIZATION_FLAG_CONFIGS. It's only removal, because we wouldn't know what kind of config it is, and thus what flags to add. The currently modified configs are: Release, RelWithDebInfo, MinSizeRel, Debug aka the usual default CMake provided ones. The mechanism is only applied to C-like languages. ASM is not handled to be on the safe side due to not knowing what kind of compiler flags the platform assembler might take. It's also important to skip RC on MSVC platforms. Task-number: QTBUG-85992 Change-Id: I3712d5cd5a34fceab54f56a6fa46b4e678952362 Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org> Reviewed-by: Cristian Adam <cristian.adam@qt.io>
2020-08-17 15:32:28 +00:00
Prevent compiler and linker flags from growing every cmake run Functions in QtFlagHandlingHelpers.cmake try to update the CMake cache variables for compiler and linker flags. These were using the current value of those variables and writing the modified ones back to the cache every time CMake ran. If a toolchain file sets or modifies any of these variables, that updated value is used and written back into the cache instead of the original cache variable's value. The next time CMake executes, the toolchain file re-applies the same change and the variable grows longer each time with flags repeated. With Ninja, this causes a complete rebuild every time CMake is re-run. The Android NDK toolchain file is one example where this behavior is triggered (the fault is shared, one could argue that the NDK should only be setting ..._INIT variables, but that's out of our control). Another related bug in the previous implementation was that the flags used to build after the first CMake execution could be different to those used for all builds after the second and later CMake runs. This is because the CMake cache was being updated, but not always the calling scope of the functions that modified them. If a toolchain file set any of the compiler or linker flag variables as non-cache variables, then updating the cache variable would have no effect on the calling scope. The non-cache variable would continue to take precedence for that scope for that run. The next time CMake executes though, the updated cache variable would now have been used by the toolchain file and the change *will* be part of the non-cache variable's value. The above are examples of why you should try to avoid updating these cache variables from project code. We could leave the cache alone and always update only non-cache variables, but then a developer looking at the cache may wonder why the values they see there don't match the values being used in builds. Or worse, they think the cache values are being used and don't realize the builds are using something different. Ultimately, we have to choose which downside we are happy to live with. The changes here preserve the previous intent of updating the cache, but it's still a bit fragile. Fixes: QTBUG-89821 Task-number: QTBUG-85992 Task-number: QTBUG-86866 Change-Id: I8a16753e159bde338e6d2e5dd2ce91fc8ac5c39d Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2021-01-07 05:53:43 +00:00
# Helper function used to update compiler and linker flags further below
function(qt_internal_replace_flags_impl flag_var_name match_string replace_string IN_CACHE)
# This must come before cache variable modification because setting the
# cache variable with FORCE will overwrite the non-cache variable, but
# we need to use the original value on entry to this function.
# Handle an empty input string and an empty match string as a set().
if(match_string STREQUAL "" AND "${${flag_var_name}}" STREQUAL "")
set(${flag_var_name} "${replace_string}" PARENT_SCOPE)
else()
string(REPLACE
"${match_string}" "${replace_string}"
${flag_var_name} "${${flag_var_name}}")
string(STRIP "${${flag_var_name}}" ${flag_var_name})
set(${flag_var_name} "${${flag_var_name}}" PARENT_SCOPE)
endif()
if(IN_CACHE)
# We must not use the non-cache variable's value because toolchain files
# might be appending things to the cache variable's value and storing it
# in a non-cache variable (e.g. Android NDK toolchain file does this).
# Work exclusively on cache variable value only.
get_property(help_text CACHE "${flag_var_name}" PROPERTY HELPSTRING)
# Handle an empty input string and an empty match string as a set().
if(match_string STREQUAL "" AND "$CACHE{${flag_var_name}}" STREQUAL "")
set(${flag_var_name} "${replace_string}" CACHE STRING "${help_text}" FORCE)
else()
set(mod_flags "$CACHE{${flag_var_name}}")
string(REPLACE
"${match_string}" "${replace_string}"
mod_flags "${mod_flags}")
string(STRIP "${mod_flags}" mod_flags)
set(${flag_var_name} "${mod_flags}" CACHE STRING "${help_text}" FORCE)
endif()
endif()
endfunction()
# Helper function used to update compiler and linker flags further below
function(qt_internal_remove_flags_impl flag_var_name flag_values IN_CACHE)
Clear non-relevant CMake compiler flags Add internal function to cleanup compiler flags out of the CMAKE_xxx_FLAGS_xxx variables. Use introduced interface to clear the '/EHsc' flag for the MSVC compiler family. This adjusts the CMake behavior to the qmake one. Change the 'EXCEPTIONS' option handling in helper functions. Add ability to add enabling and disabling exception flags. Previously it was only possible to add disabling exception flags. Fixes: QTBUG-89952 Change-Id: I60d47660a97ae9b5a1d1f4107d352c9e97890144 Reviewed-by: Craig Scott <craig.scott@qt.io> Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2021-01-21 10:17:58 +00:00
cmake_parse_arguments(arg "REGEX" "" "" ${ARGN})
set(replace_type REPLACE)
if(arg_REGEX)
list(PREPEND replace_type REGEX)
endif()
Prevent compiler and linker flags from growing every cmake run Functions in QtFlagHandlingHelpers.cmake try to update the CMake cache variables for compiler and linker flags. These were using the current value of those variables and writing the modified ones back to the cache every time CMake ran. If a toolchain file sets or modifies any of these variables, that updated value is used and written back into the cache instead of the original cache variable's value. The next time CMake executes, the toolchain file re-applies the same change and the variable grows longer each time with flags repeated. With Ninja, this causes a complete rebuild every time CMake is re-run. The Android NDK toolchain file is one example where this behavior is triggered (the fault is shared, one could argue that the NDK should only be setting ..._INIT variables, but that's out of our control). Another related bug in the previous implementation was that the flags used to build after the first CMake execution could be different to those used for all builds after the second and later CMake runs. This is because the CMake cache was being updated, but not always the calling scope of the functions that modified them. If a toolchain file set any of the compiler or linker flag variables as non-cache variables, then updating the cache variable would have no effect on the calling scope. The non-cache variable would continue to take precedence for that scope for that run. The next time CMake executes though, the updated cache variable would now have been used by the toolchain file and the change *will* be part of the non-cache variable's value. The above are examples of why you should try to avoid updating these cache variables from project code. We could leave the cache alone and always update only non-cache variables, but then a developer looking at the cache may wonder why the values they see there don't match the values being used in builds. Or worse, they think the cache values are being used and don't realize the builds are using something different. Ultimately, we have to choose which downside we are happy to live with. The changes here preserve the previous intent of updating the cache, but it's still a bit fragile. Fixes: QTBUG-89821 Task-number: QTBUG-85992 Task-number: QTBUG-86866 Change-Id: I8a16753e159bde338e6d2e5dd2ce91fc8ac5c39d Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2021-01-07 05:53:43 +00:00
# This must come before cache variable modification because setting the
# cache variable with FORCE will overwrite the non-cache variable in this
# function scope, but we need to use the original value before that change.
foreach(flag_value IN LISTS flag_values)
Replace flags with spaces when removing them Removing flags based on REGEX may include spaces, so flags might be glued. Replace flags with spaces to keep at least one space for the described case. Pick-to: 6.2 6.1 Fixes: QTBUG-94400 Change-Id: Ice268da36174ef5cf4398d2aee8fcd4731336316 Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2021-06-11 11:12:50 +00:00
string(${replace_type} "${flag_value}" " " ${flag_var_name} "${${flag_var_name}}")
Prevent compiler and linker flags from growing every cmake run Functions in QtFlagHandlingHelpers.cmake try to update the CMake cache variables for compiler and linker flags. These were using the current value of those variables and writing the modified ones back to the cache every time CMake ran. If a toolchain file sets or modifies any of these variables, that updated value is used and written back into the cache instead of the original cache variable's value. The next time CMake executes, the toolchain file re-applies the same change and the variable grows longer each time with flags repeated. With Ninja, this causes a complete rebuild every time CMake is re-run. The Android NDK toolchain file is one example where this behavior is triggered (the fault is shared, one could argue that the NDK should only be setting ..._INIT variables, but that's out of our control). Another related bug in the previous implementation was that the flags used to build after the first CMake execution could be different to those used for all builds after the second and later CMake runs. This is because the CMake cache was being updated, but not always the calling scope of the functions that modified them. If a toolchain file set any of the compiler or linker flag variables as non-cache variables, then updating the cache variable would have no effect on the calling scope. The non-cache variable would continue to take precedence for that scope for that run. The next time CMake executes though, the updated cache variable would now have been used by the toolchain file and the change *will* be part of the non-cache variable's value. The above are examples of why you should try to avoid updating these cache variables from project code. We could leave the cache alone and always update only non-cache variables, but then a developer looking at the cache may wonder why the values they see there don't match the values being used in builds. Or worse, they think the cache values are being used and don't realize the builds are using something different. Ultimately, we have to choose which downside we are happy to live with. The changes here preserve the previous intent of updating the cache, but it's still a bit fragile. Fixes: QTBUG-89821 Task-number: QTBUG-85992 Task-number: QTBUG-86866 Change-Id: I8a16753e159bde338e6d2e5dd2ce91fc8ac5c39d Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2021-01-07 05:53:43 +00:00
endforeach()
string(STRIP "${${flag_var_name}}" ${flag_var_name})
set(${flag_var_name} "${${flag_var_name}}" PARENT_SCOPE)
if(IN_CACHE)
# We must not use the non-cache variable's value because toolchain files
# might be appending things to the cache variable's value and storing it
# in a non-cache variable (e.g. Android NDK toolchain file does this).
# Work exclusively on cache variable value only.
set(mod_flags $CACHE{${flag_var_name}})
foreach(flag_value IN LISTS flag_values)
Replace flags with spaces when removing them Removing flags based on REGEX may include spaces, so flags might be glued. Replace flags with spaces to keep at least one space for the described case. Pick-to: 6.2 6.1 Fixes: QTBUG-94400 Change-Id: Ice268da36174ef5cf4398d2aee8fcd4731336316 Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2021-06-11 11:12:50 +00:00
string(${replace_type} "${flag_value}" " " mod_flags "${mod_flags}")
Prevent compiler and linker flags from growing every cmake run Functions in QtFlagHandlingHelpers.cmake try to update the CMake cache variables for compiler and linker flags. These were using the current value of those variables and writing the modified ones back to the cache every time CMake ran. If a toolchain file sets or modifies any of these variables, that updated value is used and written back into the cache instead of the original cache variable's value. The next time CMake executes, the toolchain file re-applies the same change and the variable grows longer each time with flags repeated. With Ninja, this causes a complete rebuild every time CMake is re-run. The Android NDK toolchain file is one example where this behavior is triggered (the fault is shared, one could argue that the NDK should only be setting ..._INIT variables, but that's out of our control). Another related bug in the previous implementation was that the flags used to build after the first CMake execution could be different to those used for all builds after the second and later CMake runs. This is because the CMake cache was being updated, but not always the calling scope of the functions that modified them. If a toolchain file set any of the compiler or linker flag variables as non-cache variables, then updating the cache variable would have no effect on the calling scope. The non-cache variable would continue to take precedence for that scope for that run. The next time CMake executes though, the updated cache variable would now have been used by the toolchain file and the change *will* be part of the non-cache variable's value. The above are examples of why you should try to avoid updating these cache variables from project code. We could leave the cache alone and always update only non-cache variables, but then a developer looking at the cache may wonder why the values they see there don't match the values being used in builds. Or worse, they think the cache values are being used and don't realize the builds are using something different. Ultimately, we have to choose which downside we are happy to live with. The changes here preserve the previous intent of updating the cache, but it's still a bit fragile. Fixes: QTBUG-89821 Task-number: QTBUG-85992 Task-number: QTBUG-86866 Change-Id: I8a16753e159bde338e6d2e5dd2ce91fc8ac5c39d Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2021-01-07 05:53:43 +00:00
endforeach()
string(STRIP "${mod_flags}" mod_flags)
get_property(help_text CACHE ${flag_var_name} PROPERTY HELPSTRING)
set(${flag_var_name} "${mod_flags}" CACHE STRING "${help_text}" FORCE)
endif()
endfunction()
# Helper function used to update compiler and linker flags further below
function(qt_internal_add_flags_impl flag_var_name flags IN_CACHE)
# This must come before cache variable modification because setting the
# cache variable with FORCE will overwrite the non-cache variable, but
# we need to use the original value on entry to this function.
set(${flag_var_name} "${${flag_var_name}} ${flags}")
string(STRIP "${${flag_var_name}}" ${flag_var_name})
set(${flag_var_name} "${${flag_var_name}}" PARENT_SCOPE)
if(IN_CACHE)
# We must not use the non-cache variable's value because toolchain files
# might be appending things to the cache variable's value and storing it
# in a non-cache variable (e.g. Android NDK toolchain file does this).
# Work exclusively on cache variable value only.
set(mod_flags "$CACHE{${flag_var_name}} ${flags}")
string(STRIP "${mod_flags}" mod_flags)
get_property(help_text CACHE ${flag_var_name} PROPERTY HELPSTRING)
set(${flag_var_name} "${mod_flags}" CACHE STRING "${help_text}" FORCE)
endif()
endfunction()
CMake: Refactor optimization flag handling and add optimize_full Introduce a bunch of helper functions to manipulate compiler flags and linker flags for the CMAKE_<LANG>_FLAGS_<CONFIG> and CMAKE_<LINK_TYPE>_LINKER_FLAGS_<CONFIG> CMake variables. These variables can be assigned and modified either in the cache or for a specific subdirectory scope, which will apply the flags only to targets in that scope. Add qt_internal_add_optimize_full_flags() function which mimics qmake's CONFIG += optimize_full behavior. Calling it will force usage of the '-O3' optimization flag on supported platforms (falling back '-O2' where not supported). Use the function for the Core and Gui subdirectories, to enable full optimization for the respective Qt modules as it is done in the qmake projects. To ensure that the global qmake-like compiler flags are assigned eveywhere, qt_internal_set_up_config_optimizations_like_in_qmake() needs to be called after Qt global features like optimize_size and optimize_full are available. This means that qtbase and its standalone tests need some special handling in regards to when to call that function. Task-number: QTBUG-86866 Change-Id: Ic7ac23de0265561cb06a0ba55089b6c0d3347441 Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-10-02 13:38:47 +00:00
# Removes all known compiler optimization flags for the given CONFIGS, for all enabled 'safe'
Prevent compiler and linker flags from growing every cmake run Functions in QtFlagHandlingHelpers.cmake try to update the CMake cache variables for compiler and linker flags. These were using the current value of those variables and writing the modified ones back to the cache every time CMake ran. If a toolchain file sets or modifies any of these variables, that updated value is used and written back into the cache instead of the original cache variable's value. The next time CMake executes, the toolchain file re-applies the same change and the variable grows longer each time with flags repeated. With Ninja, this causes a complete rebuild every time CMake is re-run. The Android NDK toolchain file is one example where this behavior is triggered (the fault is shared, one could argue that the NDK should only be setting ..._INIT variables, but that's out of our control). Another related bug in the previous implementation was that the flags used to build after the first CMake execution could be different to those used for all builds after the second and later CMake runs. This is because the CMake cache was being updated, but not always the calling scope of the functions that modified them. If a toolchain file set any of the compiler or linker flag variables as non-cache variables, then updating the cache variable would have no effect on the calling scope. The non-cache variable would continue to take precedence for that scope for that run. The next time CMake executes though, the updated cache variable would now have been used by the toolchain file and the change *will* be part of the non-cache variable's value. The above are examples of why you should try to avoid updating these cache variables from project code. We could leave the cache alone and always update only non-cache variables, but then a developer looking at the cache may wonder why the values they see there don't match the values being used in builds. Or worse, they think the cache values are being used and don't realize the builds are using something different. Ultimately, we have to choose which downside we are happy to live with. The changes here preserve the previous intent of updating the cache, but it's still a bit fragile. Fixes: QTBUG-89821 Task-number: QTBUG-85992 Task-number: QTBUG-86866 Change-Id: I8a16753e159bde338e6d2e5dd2ce91fc8ac5c39d Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2021-01-07 05:53:43 +00:00
# languages. The flag variables are always updated in the calling scope, even if they did not
# exist beforehand.
CMake: Refactor optimization flag handling and add optimize_full Introduce a bunch of helper functions to manipulate compiler flags and linker flags for the CMAKE_<LANG>_FLAGS_<CONFIG> and CMAKE_<LINK_TYPE>_LINKER_FLAGS_<CONFIG> CMake variables. These variables can be assigned and modified either in the cache or for a specific subdirectory scope, which will apply the flags only to targets in that scope. Add qt_internal_add_optimize_full_flags() function which mimics qmake's CONFIG += optimize_full behavior. Calling it will force usage of the '-O3' optimization flag on supported platforms (falling back '-O2' where not supported). Use the function for the Core and Gui subdirectories, to enable full optimization for the respective Qt modules as it is done in the qmake projects. To ensure that the global qmake-like compiler flags are assigned eveywhere, qt_internal_set_up_config_optimizations_like_in_qmake() needs to be called after Qt global features like optimize_size and optimize_full are available. This means that qtbase and its standalone tests need some special handling in regards to when to call that function. Task-number: QTBUG-86866 Change-Id: Ic7ac23de0265561cb06a0ba55089b6c0d3347441 Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-10-02 13:38:47 +00:00
#
Prevent compiler and linker flags from growing every cmake run Functions in QtFlagHandlingHelpers.cmake try to update the CMake cache variables for compiler and linker flags. These were using the current value of those variables and writing the modified ones back to the cache every time CMake ran. If a toolchain file sets or modifies any of these variables, that updated value is used and written back into the cache instead of the original cache variable's value. The next time CMake executes, the toolchain file re-applies the same change and the variable grows longer each time with flags repeated. With Ninja, this causes a complete rebuild every time CMake is re-run. The Android NDK toolchain file is one example where this behavior is triggered (the fault is shared, one could argue that the NDK should only be setting ..._INIT variables, but that's out of our control). Another related bug in the previous implementation was that the flags used to build after the first CMake execution could be different to those used for all builds after the second and later CMake runs. This is because the CMake cache was being updated, but not always the calling scope of the functions that modified them. If a toolchain file set any of the compiler or linker flag variables as non-cache variables, then updating the cache variable would have no effect on the calling scope. The non-cache variable would continue to take precedence for that scope for that run. The next time CMake executes though, the updated cache variable would now have been used by the toolchain file and the change *will* be part of the non-cache variable's value. The above are examples of why you should try to avoid updating these cache variables from project code. We could leave the cache alone and always update only non-cache variables, but then a developer looking at the cache may wonder why the values they see there don't match the values being used in builds. Or worse, they think the cache values are being used and don't realize the builds are using something different. Ultimately, we have to choose which downside we are happy to live with. The changes here preserve the previous intent of updating the cache, but it's still a bit fragile. Fixes: QTBUG-89821 Task-number: QTBUG-85992 Task-number: QTBUG-86866 Change-Id: I8a16753e159bde338e6d2e5dd2ce91fc8ac5c39d Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2021-01-07 05:53:43 +00:00
# IN_CACHE - remove them from the corresponding cache variable too. Note that the cache
# variable may have a different value to the non-cache variable.
# CONFIGS - should be a list of upper case configs like DEBUG, RELEASE, RELWITHDEBINFO.
# LANGUAGES - optional list of languages like 'C', 'CXX', for which to remove the flags
# if not provided, defaults to the list of enabled C-like languages
CMake: Refactor optimization flag handling and add optimize_full Introduce a bunch of helper functions to manipulate compiler flags and linker flags for the CMAKE_<LANG>_FLAGS_<CONFIG> and CMAKE_<LINK_TYPE>_LINKER_FLAGS_<CONFIG> CMake variables. These variables can be assigned and modified either in the cache or for a specific subdirectory scope, which will apply the flags only to targets in that scope. Add qt_internal_add_optimize_full_flags() function which mimics qmake's CONFIG += optimize_full behavior. Calling it will force usage of the '-O3' optimization flag on supported platforms (falling back '-O2' where not supported). Use the function for the Core and Gui subdirectories, to enable full optimization for the respective Qt modules as it is done in the qmake projects. To ensure that the global qmake-like compiler flags are assigned eveywhere, qt_internal_set_up_config_optimizations_like_in_qmake() needs to be called after Qt global features like optimize_size and optimize_full are available. This means that qtbase and its standalone tests need some special handling in regards to when to call that function. Task-number: QTBUG-86866 Change-Id: Ic7ac23de0265561cb06a0ba55089b6c0d3347441 Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-10-02 13:38:47 +00:00
function(qt_internal_remove_known_optimization_flags)
Retire the qt_parse_all_arguments This replaces the qt_parse_all_arguments macro with the built-in `cmake_parse_arguments(PARSE_ARGV`. In addition, a new function, _qt_internal_validate_all_args_are_parsed, can be used to check whether any _UNPARSED_ARGUMENTS have been passed to the function. Fixes: QTBUG-99238 Change-Id: I8cee83dc92dc6acdaaf747ea6ff9084c11dc649b Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
2023-01-11 14:36:18 +00:00
cmake_parse_arguments(PARSE_ARGV 0 arg
Prevent compiler and linker flags from growing every cmake run Functions in QtFlagHandlingHelpers.cmake try to update the CMake cache variables for compiler and linker flags. These were using the current value of those variables and writing the modified ones back to the cache every time CMake ran. If a toolchain file sets or modifies any of these variables, that updated value is used and written back into the cache instead of the original cache variable's value. The next time CMake executes, the toolchain file re-applies the same change and the variable grows longer each time with flags repeated. With Ninja, this causes a complete rebuild every time CMake is re-run. The Android NDK toolchain file is one example where this behavior is triggered (the fault is shared, one could argue that the NDK should only be setting ..._INIT variables, but that's out of our control). Another related bug in the previous implementation was that the flags used to build after the first CMake execution could be different to those used for all builds after the second and later CMake runs. This is because the CMake cache was being updated, but not always the calling scope of the functions that modified them. If a toolchain file set any of the compiler or linker flag variables as non-cache variables, then updating the cache variable would have no effect on the calling scope. The non-cache variable would continue to take precedence for that scope for that run. The next time CMake executes though, the updated cache variable would now have been used by the toolchain file and the change *will* be part of the non-cache variable's value. The above are examples of why you should try to avoid updating these cache variables from project code. We could leave the cache alone and always update only non-cache variables, but then a developer looking at the cache may wonder why the values they see there don't match the values being used in builds. Or worse, they think the cache values are being used and don't realize the builds are using something different. Ultimately, we have to choose which downside we are happy to live with. The changes here preserve the previous intent of updating the cache, but it's still a bit fragile. Fixes: QTBUG-89821 Task-number: QTBUG-85992 Task-number: QTBUG-86866 Change-Id: I8a16753e159bde338e6d2e5dd2ce91fc8ac5c39d Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2021-01-07 05:53:43 +00:00
"IN_CACHE"
CMake: Refactor optimization flag handling and add optimize_full Introduce a bunch of helper functions to manipulate compiler flags and linker flags for the CMAKE_<LANG>_FLAGS_<CONFIG> and CMAKE_<LINK_TYPE>_LINKER_FLAGS_<CONFIG> CMake variables. These variables can be assigned and modified either in the cache or for a specific subdirectory scope, which will apply the flags only to targets in that scope. Add qt_internal_add_optimize_full_flags() function which mimics qmake's CONFIG += optimize_full behavior. Calling it will force usage of the '-O3' optimization flag on supported platforms (falling back '-O2' where not supported). Use the function for the Core and Gui subdirectories, to enable full optimization for the respective Qt modules as it is done in the qmake projects. To ensure that the global qmake-like compiler flags are assigned eveywhere, qt_internal_set_up_config_optimizations_like_in_qmake() needs to be called after Qt global features like optimize_size and optimize_full are available. This means that qtbase and its standalone tests need some special handling in regards to when to call that function. Task-number: QTBUG-86866 Change-Id: Ic7ac23de0265561cb06a0ba55089b6c0d3347441 Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-10-02 13:38:47 +00:00
""
Retire the qt_parse_all_arguments This replaces the qt_parse_all_arguments macro with the built-in `cmake_parse_arguments(PARSE_ARGV`. In addition, a new function, _qt_internal_validate_all_args_are_parsed, can be used to check whether any _UNPARSED_ARGUMENTS have been passed to the function. Fixes: QTBUG-99238 Change-Id: I8cee83dc92dc6acdaaf747ea6ff9084c11dc649b Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
2023-01-11 14:36:18 +00:00
"CONFIGS;LANGUAGES")
_qt_internal_validate_all_args_are_parsed(arg)
CMake: Refactor optimization flag handling and add optimize_full Introduce a bunch of helper functions to manipulate compiler flags and linker flags for the CMAKE_<LANG>_FLAGS_<CONFIG> and CMAKE_<LINK_TYPE>_LINKER_FLAGS_<CONFIG> CMake variables. These variables can be assigned and modified either in the cache or for a specific subdirectory scope, which will apply the flags only to targets in that scope. Add qt_internal_add_optimize_full_flags() function which mimics qmake's CONFIG += optimize_full behavior. Calling it will force usage of the '-O3' optimization flag on supported platforms (falling back '-O2' where not supported). Use the function for the Core and Gui subdirectories, to enable full optimization for the respective Qt modules as it is done in the qmake projects. To ensure that the global qmake-like compiler flags are assigned eveywhere, qt_internal_set_up_config_optimizations_like_in_qmake() needs to be called after Qt global features like optimize_size and optimize_full are available. This means that qtbase and its standalone tests need some special handling in regards to when to call that function. Task-number: QTBUG-86866 Change-Id: Ic7ac23de0265561cb06a0ba55089b6c0d3347441 Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-10-02 13:38:47 +00:00
if(NOT arg_CONFIGS)
message(FATAL_ERROR
Prevent compiler and linker flags from growing every cmake run Functions in QtFlagHandlingHelpers.cmake try to update the CMake cache variables for compiler and linker flags. These were using the current value of those variables and writing the modified ones back to the cache every time CMake ran. If a toolchain file sets or modifies any of these variables, that updated value is used and written back into the cache instead of the original cache variable's value. The next time CMake executes, the toolchain file re-applies the same change and the variable grows longer each time with flags repeated. With Ninja, this causes a complete rebuild every time CMake is re-run. The Android NDK toolchain file is one example where this behavior is triggered (the fault is shared, one could argue that the NDK should only be setting ..._INIT variables, but that's out of our control). Another related bug in the previous implementation was that the flags used to build after the first CMake execution could be different to those used for all builds after the second and later CMake runs. This is because the CMake cache was being updated, but not always the calling scope of the functions that modified them. If a toolchain file set any of the compiler or linker flag variables as non-cache variables, then updating the cache variable would have no effect on the calling scope. The non-cache variable would continue to take precedence for that scope for that run. The next time CMake executes though, the updated cache variable would now have been used by the toolchain file and the change *will* be part of the non-cache variable's value. The above are examples of why you should try to avoid updating these cache variables from project code. We could leave the cache alone and always update only non-cache variables, but then a developer looking at the cache may wonder why the values they see there don't match the values being used in builds. Or worse, they think the cache values are being used and don't realize the builds are using something different. Ultimately, we have to choose which downside we are happy to live with. The changes here preserve the previous intent of updating the cache, but it's still a bit fragile. Fixes: QTBUG-89821 Task-number: QTBUG-85992 Task-number: QTBUG-86866 Change-Id: I8a16753e159bde338e6d2e5dd2ce91fc8ac5c39d Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2021-01-07 05:53:43 +00:00
"You must specify at least one configuration for which to remove the flags.")
CMake: Adjust compiler flag optimizations to qmake mkspec ones There are inconsistencies in the default optimization flags added by CMake across configurations like Release and RelWithDebInfo. In particular Release uses -O3, whereas RelWithDebInfo uses -O2, as well as usage of /INCREMENTAL in release configs with MSVC, etc. To make sure that the Qt 6 binaries built with CMake are consistent across configs, as well as consistent with the flags we used when building Qt 5 with qmake, add a horrible search and replace mechanism to replaces the CMake flags with what our mkspecs indicate to use. Ideally this would be done by providing custom CMake toolchain files for each platform we support, and we might revisit that later if the need really arises. To implement the replacing, we first need the flags that should be added. Port the QMAKE_CFLAGS_OPTIMIZE variables to CMake, which is done in QtCompilerOptimization.cmake. Then a new function called qt_internal_set_up_config_optimizations_like_in_qmake will look for any kind of optimization flags set in the CMAKE_<LANG>_FLAGS_<CONFIG> style variables, remove them, and add the appropriate flags that qmake mkspecs provide. On some platforms (like Windows MSVC) the function also alters the linker CMAKE_${TYPE}_LINKER_FLAGS_<CONFIG> style variables. The mechanism allows opting out of this replacing by setting the QT_USE_DEFAULT_CMAKE_OPTIMIZATION_FLAGS value. It also allows opting into removal of flags for custom configs by providing QT_ADDITIONAL_OPTIMIZATION_FLAG_CONFIGS. It's only removal, because we wouldn't know what kind of config it is, and thus what flags to add. The currently modified configs are: Release, RelWithDebInfo, MinSizeRel, Debug aka the usual default CMake provided ones. The mechanism is only applied to C-like languages. ASM is not handled to be on the safe side due to not knowing what kind of compiler flags the platform assembler might take. It's also important to skip RC on MSVC platforms. Task-number: QTBUG-85992 Change-Id: I3712d5cd5a34fceab54f56a6fa46b4e678952362 Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org> Reviewed-by: Cristian Adam <cristian.adam@qt.io>
2020-08-17 15:32:28 +00:00
endif()
CMake: Refactor optimization flag handling and add optimize_full Introduce a bunch of helper functions to manipulate compiler flags and linker flags for the CMAKE_<LANG>_FLAGS_<CONFIG> and CMAKE_<LINK_TYPE>_LINKER_FLAGS_<CONFIG> CMake variables. These variables can be assigned and modified either in the cache or for a specific subdirectory scope, which will apply the flags only to targets in that scope. Add qt_internal_add_optimize_full_flags() function which mimics qmake's CONFIG += optimize_full behavior. Calling it will force usage of the '-O3' optimization flag on supported platforms (falling back '-O2' where not supported). Use the function for the Core and Gui subdirectories, to enable full optimization for the respective Qt modules as it is done in the qmake projects. To ensure that the global qmake-like compiler flags are assigned eveywhere, qt_internal_set_up_config_optimizations_like_in_qmake() needs to be called after Qt global features like optimize_size and optimize_full are available. This means that qtbase and its standalone tests need some special handling in regards to when to call that function. Task-number: QTBUG-86866 Change-Id: Ic7ac23de0265561cb06a0ba55089b6c0d3347441 Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-10-02 13:38:47 +00:00
if(arg_LANGUAGES)
set(enabled_languages "${arg_LANGUAGES}")
else()
qt_internal_get_enabled_languages_for_flag_manipulation(enabled_languages)
CMake: Adjust compiler flag optimizations to qmake mkspec ones There are inconsistencies in the default optimization flags added by CMake across configurations like Release and RelWithDebInfo. In particular Release uses -O3, whereas RelWithDebInfo uses -O2, as well as usage of /INCREMENTAL in release configs with MSVC, etc. To make sure that the Qt 6 binaries built with CMake are consistent across configs, as well as consistent with the flags we used when building Qt 5 with qmake, add a horrible search and replace mechanism to replaces the CMake flags with what our mkspecs indicate to use. Ideally this would be done by providing custom CMake toolchain files for each platform we support, and we might revisit that later if the need really arises. To implement the replacing, we first need the flags that should be added. Port the QMAKE_CFLAGS_OPTIMIZE variables to CMake, which is done in QtCompilerOptimization.cmake. Then a new function called qt_internal_set_up_config_optimizations_like_in_qmake will look for any kind of optimization flags set in the CMAKE_<LANG>_FLAGS_<CONFIG> style variables, remove them, and add the appropriate flags that qmake mkspecs provide. On some platforms (like Windows MSVC) the function also alters the linker CMAKE_${TYPE}_LINKER_FLAGS_<CONFIG> style variables. The mechanism allows opting out of this replacing by setting the QT_USE_DEFAULT_CMAKE_OPTIMIZATION_FLAGS value. It also allows opting into removal of flags for custom configs by providing QT_ADDITIONAL_OPTIMIZATION_FLAG_CONFIGS. It's only removal, because we wouldn't know what kind of config it is, and thus what flags to add. The currently modified configs are: Release, RelWithDebInfo, MinSizeRel, Debug aka the usual default CMake provided ones. The mechanism is only applied to C-like languages. ASM is not handled to be on the safe side due to not knowing what kind of compiler flags the platform assembler might take. It's also important to skip RC on MSVC platforms. Task-number: QTBUG-85992 Change-Id: I3712d5cd5a34fceab54f56a6fa46b4e678952362 Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org> Reviewed-by: Cristian Adam <cristian.adam@qt.io>
2020-08-17 15:32:28 +00:00
endif()
qt_internal_get_all_possible_optimization_flag_values(flag_values)
CMake: Refactor optimization flag handling and add optimize_full Introduce a bunch of helper functions to manipulate compiler flags and linker flags for the CMAKE_<LANG>_FLAGS_<CONFIG> and CMAKE_<LINK_TYPE>_LINKER_FLAGS_<CONFIG> CMake variables. These variables can be assigned and modified either in the cache or for a specific subdirectory scope, which will apply the flags only to targets in that scope. Add qt_internal_add_optimize_full_flags() function which mimics qmake's CONFIG += optimize_full behavior. Calling it will force usage of the '-O3' optimization flag on supported platforms (falling back '-O2' where not supported). Use the function for the Core and Gui subdirectories, to enable full optimization for the respective Qt modules as it is done in the qmake projects. To ensure that the global qmake-like compiler flags are assigned eveywhere, qt_internal_set_up_config_optimizations_like_in_qmake() needs to be called after Qt global features like optimize_size and optimize_full are available. This means that qtbase and its standalone tests need some special handling in regards to when to call that function. Task-number: QTBUG-86866 Change-Id: Ic7ac23de0265561cb06a0ba55089b6c0d3347441 Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-10-02 13:38:47 +00:00
set(configs ${arg_CONFIGS})
CMake: Adjust compiler flag optimizations to qmake mkspec ones There are inconsistencies in the default optimization flags added by CMake across configurations like Release and RelWithDebInfo. In particular Release uses -O3, whereas RelWithDebInfo uses -O2, as well as usage of /INCREMENTAL in release configs with MSVC, etc. To make sure that the Qt 6 binaries built with CMake are consistent across configs, as well as consistent with the flags we used when building Qt 5 with qmake, add a horrible search and replace mechanism to replaces the CMake flags with what our mkspecs indicate to use. Ideally this would be done by providing custom CMake toolchain files for each platform we support, and we might revisit that later if the need really arises. To implement the replacing, we first need the flags that should be added. Port the QMAKE_CFLAGS_OPTIMIZE variables to CMake, which is done in QtCompilerOptimization.cmake. Then a new function called qt_internal_set_up_config_optimizations_like_in_qmake will look for any kind of optimization flags set in the CMAKE_<LANG>_FLAGS_<CONFIG> style variables, remove them, and add the appropriate flags that qmake mkspecs provide. On some platforms (like Windows MSVC) the function also alters the linker CMAKE_${TYPE}_LINKER_FLAGS_<CONFIG> style variables. The mechanism allows opting out of this replacing by setting the QT_USE_DEFAULT_CMAKE_OPTIMIZATION_FLAGS value. It also allows opting into removal of flags for custom configs by providing QT_ADDITIONAL_OPTIMIZATION_FLAG_CONFIGS. It's only removal, because we wouldn't know what kind of config it is, and thus what flags to add. The currently modified configs are: Release, RelWithDebInfo, MinSizeRel, Debug aka the usual default CMake provided ones. The mechanism is only applied to C-like languages. ASM is not handled to be on the safe side due to not knowing what kind of compiler flags the platform assembler might take. It's also important to skip RC on MSVC platforms. Task-number: QTBUG-85992 Change-Id: I3712d5cd5a34fceab54f56a6fa46b4e678952362 Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org> Reviewed-by: Cristian Adam <cristian.adam@qt.io>
2020-08-17 15:32:28 +00:00
foreach(lang ${enabled_languages})
foreach(config ${configs})
set(flag_var_name "CMAKE_${lang}_FLAGS_${config}")
Prevent compiler and linker flags from growing every cmake run Functions in QtFlagHandlingHelpers.cmake try to update the CMake cache variables for compiler and linker flags. These were using the current value of those variables and writing the modified ones back to the cache every time CMake ran. If a toolchain file sets or modifies any of these variables, that updated value is used and written back into the cache instead of the original cache variable's value. The next time CMake executes, the toolchain file re-applies the same change and the variable grows longer each time with flags repeated. With Ninja, this causes a complete rebuild every time CMake is re-run. The Android NDK toolchain file is one example where this behavior is triggered (the fault is shared, one could argue that the NDK should only be setting ..._INIT variables, but that's out of our control). Another related bug in the previous implementation was that the flags used to build after the first CMake execution could be different to those used for all builds after the second and later CMake runs. This is because the CMake cache was being updated, but not always the calling scope of the functions that modified them. If a toolchain file set any of the compiler or linker flag variables as non-cache variables, then updating the cache variable would have no effect on the calling scope. The non-cache variable would continue to take precedence for that scope for that run. The next time CMake executes though, the updated cache variable would now have been used by the toolchain file and the change *will* be part of the non-cache variable's value. The above are examples of why you should try to avoid updating these cache variables from project code. We could leave the cache alone and always update only non-cache variables, but then a developer looking at the cache may wonder why the values they see there don't match the values being used in builds. Or worse, they think the cache values are being used and don't realize the builds are using something different. Ultimately, we have to choose which downside we are happy to live with. The changes here preserve the previous intent of updating the cache, but it's still a bit fragile. Fixes: QTBUG-89821 Task-number: QTBUG-85992 Task-number: QTBUG-86866 Change-Id: I8a16753e159bde338e6d2e5dd2ce91fc8ac5c39d Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2021-01-07 05:53:43 +00:00
qt_internal_remove_flags_impl(${flag_var_name} "${flag_values}" "${arg_IN_CACHE}")
set(${flag_var_name} "${${flag_var_name}}" PARENT_SCOPE)
CMake: Adjust compiler flag optimizations to qmake mkspec ones There are inconsistencies in the default optimization flags added by CMake across configurations like Release and RelWithDebInfo. In particular Release uses -O3, whereas RelWithDebInfo uses -O2, as well as usage of /INCREMENTAL in release configs with MSVC, etc. To make sure that the Qt 6 binaries built with CMake are consistent across configs, as well as consistent with the flags we used when building Qt 5 with qmake, add a horrible search and replace mechanism to replaces the CMake flags with what our mkspecs indicate to use. Ideally this would be done by providing custom CMake toolchain files for each platform we support, and we might revisit that later if the need really arises. To implement the replacing, we first need the flags that should be added. Port the QMAKE_CFLAGS_OPTIMIZE variables to CMake, which is done in QtCompilerOptimization.cmake. Then a new function called qt_internal_set_up_config_optimizations_like_in_qmake will look for any kind of optimization flags set in the CMAKE_<LANG>_FLAGS_<CONFIG> style variables, remove them, and add the appropriate flags that qmake mkspecs provide. On some platforms (like Windows MSVC) the function also alters the linker CMAKE_${TYPE}_LINKER_FLAGS_<CONFIG> style variables. The mechanism allows opting out of this replacing by setting the QT_USE_DEFAULT_CMAKE_OPTIMIZATION_FLAGS value. It also allows opting into removal of flags for custom configs by providing QT_ADDITIONAL_OPTIMIZATION_FLAG_CONFIGS. It's only removal, because we wouldn't know what kind of config it is, and thus what flags to add. The currently modified configs are: Release, RelWithDebInfo, MinSizeRel, Debug aka the usual default CMake provided ones. The mechanism is only applied to C-like languages. ASM is not handled to be on the safe side due to not knowing what kind of compiler flags the platform assembler might take. It's also important to skip RC on MSVC platforms. Task-number: QTBUG-85992 Change-Id: I3712d5cd5a34fceab54f56a6fa46b4e678952362 Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org> Reviewed-by: Cristian Adam <cristian.adam@qt.io>
2020-08-17 15:32:28 +00:00
endforeach()
endforeach()
CMake: Refactor optimization flag handling and add optimize_full Introduce a bunch of helper functions to manipulate compiler flags and linker flags for the CMAKE_<LANG>_FLAGS_<CONFIG> and CMAKE_<LINK_TYPE>_LINKER_FLAGS_<CONFIG> CMake variables. These variables can be assigned and modified either in the cache or for a specific subdirectory scope, which will apply the flags only to targets in that scope. Add qt_internal_add_optimize_full_flags() function which mimics qmake's CONFIG += optimize_full behavior. Calling it will force usage of the '-O3' optimization flag on supported platforms (falling back '-O2' where not supported). Use the function for the Core and Gui subdirectories, to enable full optimization for the respective Qt modules as it is done in the qmake projects. To ensure that the global qmake-like compiler flags are assigned eveywhere, qt_internal_set_up_config_optimizations_like_in_qmake() needs to be called after Qt global features like optimize_size and optimize_full are available. This means that qtbase and its standalone tests need some special handling in regards to when to call that function. Task-number: QTBUG-86866 Change-Id: Ic7ac23de0265561cb06a0ba55089b6c0d3347441 Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-10-02 13:38:47 +00:00
endfunction()
Clear non-relevant CMake compiler flags Add internal function to cleanup compiler flags out of the CMAKE_xxx_FLAGS_xxx variables. Use introduced interface to clear the '/EHsc' flag for the MSVC compiler family. This adjusts the CMake behavior to the qmake one. Change the 'EXCEPTIONS' option handling in helper functions. Add ability to add enabling and disabling exception flags. Previously it was only possible to add disabling exception flags. Fixes: QTBUG-89952 Change-Id: I60d47660a97ae9b5a1d1f4107d352c9e97890144 Reviewed-by: Craig Scott <craig.scott@qt.io> Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2021-01-21 10:17:58 +00:00
# Removes specified flags from CMAKE_<LANGUAGES>_FLAGS[_CONFIGS] variables
#
Make CONFIGS arguments of qt_internal_remove_compiler_flags optional There was a discrepancy between what the comment was saying, and what the function was doing before, and with this patch, we opt for what the comment was saying, which makes more sense. In addition, I cleaned up its documentation a bit. Pick-to: 6.5 Change-Id: I07c20f93aa5c8e9bc942f5e69f0cf8299f4a813e Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2023-05-05 14:07:50 +00:00
# Option Arguments:
# IN_CACHE
# Enables flags removal from CACHE
# REGEX
# Enables the flag processing as a regular expression.
#
# Multi-value Arguments:
# CONFIGS
# List of configurations that need to clear flags. Clears all configs by default if not
# specified.
#
# LANGUAGES
# List of LANGUAGES that need clear flags. Clears all languages by default if not
# specified.
Clear non-relevant CMake compiler flags Add internal function to cleanup compiler flags out of the CMAKE_xxx_FLAGS_xxx variables. Use introduced interface to clear the '/EHsc' flag for the MSVC compiler family. This adjusts the CMake behavior to the qmake one. Change the 'EXCEPTIONS' option handling in helper functions. Add ability to add enabling and disabling exception flags. Previously it was only possible to add disabling exception flags. Fixes: QTBUG-89952 Change-Id: I60d47660a97ae9b5a1d1f4107d352c9e97890144 Reviewed-by: Craig Scott <craig.scott@qt.io> Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2021-01-21 10:17:58 +00:00
function(qt_internal_remove_compiler_flags flags)
Retire the qt_parse_all_arguments This replaces the qt_parse_all_arguments macro with the built-in `cmake_parse_arguments(PARSE_ARGV`. In addition, a new function, _qt_internal_validate_all_args_are_parsed, can be used to check whether any _UNPARSED_ARGUMENTS have been passed to the function. Fixes: QTBUG-99238 Change-Id: I8cee83dc92dc6acdaaf747ea6ff9084c11dc649b Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
2023-01-11 14:36:18 +00:00
cmake_parse_arguments(PARSE_ARGV 1 arg
Clear non-relevant CMake compiler flags Add internal function to cleanup compiler flags out of the CMAKE_xxx_FLAGS_xxx variables. Use introduced interface to clear the '/EHsc' flag for the MSVC compiler family. This adjusts the CMake behavior to the qmake one. Change the 'EXCEPTIONS' option handling in helper functions. Add ability to add enabling and disabling exception flags. Previously it was only possible to add disabling exception flags. Fixes: QTBUG-89952 Change-Id: I60d47660a97ae9b5a1d1f4107d352c9e97890144 Reviewed-by: Craig Scott <craig.scott@qt.io> Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2021-01-21 10:17:58 +00:00
"IN_CACHE;REGEX"
""
"CONFIGS;LANGUAGES"
)
Retire the qt_parse_all_arguments This replaces the qt_parse_all_arguments macro with the built-in `cmake_parse_arguments(PARSE_ARGV`. In addition, a new function, _qt_internal_validate_all_args_are_parsed, can be used to check whether any _UNPARSED_ARGUMENTS have been passed to the function. Fixes: QTBUG-99238 Change-Id: I8cee83dc92dc6acdaaf747ea6ff9084c11dc649b Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
2023-01-11 14:36:18 +00:00
_qt_internal_validate_all_args_are_parsed(arg)
Clear non-relevant CMake compiler flags Add internal function to cleanup compiler flags out of the CMAKE_xxx_FLAGS_xxx variables. Use introduced interface to clear the '/EHsc' flag for the MSVC compiler family. This adjusts the CMake behavior to the qmake one. Change the 'EXCEPTIONS' option handling in helper functions. Add ability to add enabling and disabling exception flags. Previously it was only possible to add disabling exception flags. Fixes: QTBUG-89952 Change-Id: I60d47660a97ae9b5a1d1f4107d352c9e97890144 Reviewed-by: Craig Scott <craig.scott@qt.io> Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2021-01-21 10:17:58 +00:00
if("${flags}" STREQUAL "")
message(WARNING "qt_internal_remove_compiler_flags was called without any flags specified.")
return()
endif()
if(arg_LANGUAGES)
set(languages "${arg_LANGUAGES}")
else()
qt_internal_get_enabled_languages_for_flag_manipulation(languages)
endif()
if(arg_CONFIGS)
set(configs "${arg_CONFIGS}")
else()
Make CONFIGS arguments of qt_internal_remove_compiler_flags optional There was a discrepancy between what the comment was saying, and what the function was doing before, and with this patch, we opt for what the comment was saying, which makes more sense. In addition, I cleaned up its documentation a bit. Pick-to: 6.5 Change-Id: I07c20f93aa5c8e9bc942f5e69f0cf8299f4a813e Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2023-05-05 14:07:50 +00:00
qt_internal_get_configs_for_flag_manipulation(configs)
Clear non-relevant CMake compiler flags Add internal function to cleanup compiler flags out of the CMAKE_xxx_FLAGS_xxx variables. Use introduced interface to clear the '/EHsc' flag for the MSVC compiler family. This adjusts the CMake behavior to the qmake one. Change the 'EXCEPTIONS' option handling in helper functions. Add ability to add enabling and disabling exception flags. Previously it was only possible to add disabling exception flags. Fixes: QTBUG-89952 Change-Id: I60d47660a97ae9b5a1d1f4107d352c9e97890144 Reviewed-by: Craig Scott <craig.scott@qt.io> Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2021-01-21 10:17:58 +00:00
endif()
if(arg_REGEX)
list(APPEND extra_options "REGEX")
endif()
foreach(lang ${languages})
set(flag_var_name "CMAKE_${lang}_FLAGS")
qt_internal_remove_flags_impl(${flag_var_name}
"${flags}"
"${arg_IN_CACHE}"
${extra_options}
)
set(${flag_var_name} "${${flag_var_name}}" PARENT_SCOPE)
foreach(config ${configs})
set(flag_var_name "CMAKE_${lang}_FLAGS_${config}")
qt_internal_remove_flags_impl(${flag_var_name}
"${flags}"
"${arg_IN_CACHE}"
${extra_options}
)
set(${flag_var_name} "${${flag_var_name}}" PARENT_SCOPE)
endforeach()
endforeach()
endfunction()
Prevent compiler and linker flags from growing every cmake run Functions in QtFlagHandlingHelpers.cmake try to update the CMake cache variables for compiler and linker flags. These were using the current value of those variables and writing the modified ones back to the cache every time CMake ran. If a toolchain file sets or modifies any of these variables, that updated value is used and written back into the cache instead of the original cache variable's value. The next time CMake executes, the toolchain file re-applies the same change and the variable grows longer each time with flags repeated. With Ninja, this causes a complete rebuild every time CMake is re-run. The Android NDK toolchain file is one example where this behavior is triggered (the fault is shared, one could argue that the NDK should only be setting ..._INIT variables, but that's out of our control). Another related bug in the previous implementation was that the flags used to build after the first CMake execution could be different to those used for all builds after the second and later CMake runs. This is because the CMake cache was being updated, but not always the calling scope of the functions that modified them. If a toolchain file set any of the compiler or linker flag variables as non-cache variables, then updating the cache variable would have no effect on the calling scope. The non-cache variable would continue to take precedence for that scope for that run. The next time CMake executes though, the updated cache variable would now have been used by the toolchain file and the change *will* be part of the non-cache variable's value. The above are examples of why you should try to avoid updating these cache variables from project code. We could leave the cache alone and always update only non-cache variables, but then a developer looking at the cache may wonder why the values they see there don't match the values being used in builds. Or worse, they think the cache values are being used and don't realize the builds are using something different. Ultimately, we have to choose which downside we are happy to live with. The changes here preserve the previous intent of updating the cache, but it's still a bit fragile. Fixes: QTBUG-89821 Task-number: QTBUG-85992 Task-number: QTBUG-86866 Change-Id: I8a16753e159bde338e6d2e5dd2ce91fc8ac5c39d Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2021-01-07 05:53:43 +00:00
# Adds compiler flags for the given CONFIGS in the calling scope. Can also update the cache
# if asked to do so. The flag variables are always updated in the calling scope, even if they
# did not exist beforehand.
CMake: Refactor optimization flag handling and add optimize_full Introduce a bunch of helper functions to manipulate compiler flags and linker flags for the CMAKE_<LANG>_FLAGS_<CONFIG> and CMAKE_<LINK_TYPE>_LINKER_FLAGS_<CONFIG> CMake variables. These variables can be assigned and modified either in the cache or for a specific subdirectory scope, which will apply the flags only to targets in that scope. Add qt_internal_add_optimize_full_flags() function which mimics qmake's CONFIG += optimize_full behavior. Calling it will force usage of the '-O3' optimization flag on supported platforms (falling back '-O2' where not supported). Use the function for the Core and Gui subdirectories, to enable full optimization for the respective Qt modules as it is done in the qmake projects. To ensure that the global qmake-like compiler flags are assigned eveywhere, qt_internal_set_up_config_optimizations_like_in_qmake() needs to be called after Qt global features like optimize_size and optimize_full are available. This means that qtbase and its standalone tests need some special handling in regards to when to call that function. Task-number: QTBUG-86866 Change-Id: Ic7ac23de0265561cb06a0ba55089b6c0d3347441 Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-10-02 13:38:47 +00:00
#
Prevent compiler and linker flags from growing every cmake run Functions in QtFlagHandlingHelpers.cmake try to update the CMake cache variables for compiler and linker flags. These were using the current value of those variables and writing the modified ones back to the cache every time CMake ran. If a toolchain file sets or modifies any of these variables, that updated value is used and written back into the cache instead of the original cache variable's value. The next time CMake executes, the toolchain file re-applies the same change and the variable grows longer each time with flags repeated. With Ninja, this causes a complete rebuild every time CMake is re-run. The Android NDK toolchain file is one example where this behavior is triggered (the fault is shared, one could argue that the NDK should only be setting ..._INIT variables, but that's out of our control). Another related bug in the previous implementation was that the flags used to build after the first CMake execution could be different to those used for all builds after the second and later CMake runs. This is because the CMake cache was being updated, but not always the calling scope of the functions that modified them. If a toolchain file set any of the compiler or linker flag variables as non-cache variables, then updating the cache variable would have no effect on the calling scope. The non-cache variable would continue to take precedence for that scope for that run. The next time CMake executes though, the updated cache variable would now have been used by the toolchain file and the change *will* be part of the non-cache variable's value. The above are examples of why you should try to avoid updating these cache variables from project code. We could leave the cache alone and always update only non-cache variables, but then a developer looking at the cache may wonder why the values they see there don't match the values being used in builds. Or worse, they think the cache values are being used and don't realize the builds are using something different. Ultimately, we have to choose which downside we are happy to live with. The changes here preserve the previous intent of updating the cache, but it's still a bit fragile. Fixes: QTBUG-89821 Task-number: QTBUG-85992 Task-number: QTBUG-86866 Change-Id: I8a16753e159bde338e6d2e5dd2ce91fc8ac5c39d Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2021-01-07 05:53:43 +00:00
# FLAGS - should be a single string of flags separated by spaces.
# IN_CACHE - add them to the corresponding cache variable too. Note that the cache
# variable may have a different value to the non-cache variable.
# CONFIGS - should be a list of upper case configs like DEBUG, RELEASE, RELWITHDEBINFO.
# LANGUAGES - optional list of languages like 'C', 'CXX', for which to add the flags
# if not provided, defaults to the list of enabled C-like languages
CMake: Refactor optimization flag handling and add optimize_full Introduce a bunch of helper functions to manipulate compiler flags and linker flags for the CMAKE_<LANG>_FLAGS_<CONFIG> and CMAKE_<LINK_TYPE>_LINKER_FLAGS_<CONFIG> CMake variables. These variables can be assigned and modified either in the cache or for a specific subdirectory scope, which will apply the flags only to targets in that scope. Add qt_internal_add_optimize_full_flags() function which mimics qmake's CONFIG += optimize_full behavior. Calling it will force usage of the '-O3' optimization flag on supported platforms (falling back '-O2' where not supported). Use the function for the Core and Gui subdirectories, to enable full optimization for the respective Qt modules as it is done in the qmake projects. To ensure that the global qmake-like compiler flags are assigned eveywhere, qt_internal_set_up_config_optimizations_like_in_qmake() needs to be called after Qt global features like optimize_size and optimize_full are available. This means that qtbase and its standalone tests need some special handling in regards to when to call that function. Task-number: QTBUG-86866 Change-Id: Ic7ac23de0265561cb06a0ba55089b6c0d3347441 Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-10-02 13:38:47 +00:00
function(qt_internal_add_compiler_flags)
Retire the qt_parse_all_arguments This replaces the qt_parse_all_arguments macro with the built-in `cmake_parse_arguments(PARSE_ARGV`. In addition, a new function, _qt_internal_validate_all_args_are_parsed, can be used to check whether any _UNPARSED_ARGUMENTS have been passed to the function. Fixes: QTBUG-99238 Change-Id: I8cee83dc92dc6acdaaf747ea6ff9084c11dc649b Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
2023-01-11 14:36:18 +00:00
cmake_parse_arguments(PARSE_ARGV 0 arg
Prevent compiler and linker flags from growing every cmake run Functions in QtFlagHandlingHelpers.cmake try to update the CMake cache variables for compiler and linker flags. These were using the current value of those variables and writing the modified ones back to the cache every time CMake ran. If a toolchain file sets or modifies any of these variables, that updated value is used and written back into the cache instead of the original cache variable's value. The next time CMake executes, the toolchain file re-applies the same change and the variable grows longer each time with flags repeated. With Ninja, this causes a complete rebuild every time CMake is re-run. The Android NDK toolchain file is one example where this behavior is triggered (the fault is shared, one could argue that the NDK should only be setting ..._INIT variables, but that's out of our control). Another related bug in the previous implementation was that the flags used to build after the first CMake execution could be different to those used for all builds after the second and later CMake runs. This is because the CMake cache was being updated, but not always the calling scope of the functions that modified them. If a toolchain file set any of the compiler or linker flag variables as non-cache variables, then updating the cache variable would have no effect on the calling scope. The non-cache variable would continue to take precedence for that scope for that run. The next time CMake executes though, the updated cache variable would now have been used by the toolchain file and the change *will* be part of the non-cache variable's value. The above are examples of why you should try to avoid updating these cache variables from project code. We could leave the cache alone and always update only non-cache variables, but then a developer looking at the cache may wonder why the values they see there don't match the values being used in builds. Or worse, they think the cache values are being used and don't realize the builds are using something different. Ultimately, we have to choose which downside we are happy to live with. The changes here preserve the previous intent of updating the cache, but it's still a bit fragile. Fixes: QTBUG-89821 Task-number: QTBUG-85992 Task-number: QTBUG-86866 Change-Id: I8a16753e159bde338e6d2e5dd2ce91fc8ac5c39d Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2021-01-07 05:53:43 +00:00
"IN_CACHE"
CMake: Refactor optimization flag handling and add optimize_full Introduce a bunch of helper functions to manipulate compiler flags and linker flags for the CMAKE_<LANG>_FLAGS_<CONFIG> and CMAKE_<LINK_TYPE>_LINKER_FLAGS_<CONFIG> CMake variables. These variables can be assigned and modified either in the cache or for a specific subdirectory scope, which will apply the flags only to targets in that scope. Add qt_internal_add_optimize_full_flags() function which mimics qmake's CONFIG += optimize_full behavior. Calling it will force usage of the '-O3' optimization flag on supported platforms (falling back '-O2' where not supported). Use the function for the Core and Gui subdirectories, to enable full optimization for the respective Qt modules as it is done in the qmake projects. To ensure that the global qmake-like compiler flags are assigned eveywhere, qt_internal_set_up_config_optimizations_like_in_qmake() needs to be called after Qt global features like optimize_size and optimize_full are available. This means that qtbase and its standalone tests need some special handling in regards to when to call that function. Task-number: QTBUG-86866 Change-Id: Ic7ac23de0265561cb06a0ba55089b6c0d3347441 Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-10-02 13:38:47 +00:00
"FLAGS"
Retire the qt_parse_all_arguments This replaces the qt_parse_all_arguments macro with the built-in `cmake_parse_arguments(PARSE_ARGV`. In addition, a new function, _qt_internal_validate_all_args_are_parsed, can be used to check whether any _UNPARSED_ARGUMENTS have been passed to the function. Fixes: QTBUG-99238 Change-Id: I8cee83dc92dc6acdaaf747ea6ff9084c11dc649b Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
2023-01-11 14:36:18 +00:00
"CONFIGS;LANGUAGES")
_qt_internal_validate_all_args_are_parsed(arg)
CMake: Refactor optimization flag handling and add optimize_full Introduce a bunch of helper functions to manipulate compiler flags and linker flags for the CMAKE_<LANG>_FLAGS_<CONFIG> and CMAKE_<LINK_TYPE>_LINKER_FLAGS_<CONFIG> CMake variables. These variables can be assigned and modified either in the cache or for a specific subdirectory scope, which will apply the flags only to targets in that scope. Add qt_internal_add_optimize_full_flags() function which mimics qmake's CONFIG += optimize_full behavior. Calling it will force usage of the '-O3' optimization flag on supported platforms (falling back '-O2' where not supported). Use the function for the Core and Gui subdirectories, to enable full optimization for the respective Qt modules as it is done in the qmake projects. To ensure that the global qmake-like compiler flags are assigned eveywhere, qt_internal_set_up_config_optimizations_like_in_qmake() needs to be called after Qt global features like optimize_size and optimize_full are available. This means that qtbase and its standalone tests need some special handling in regards to when to call that function. Task-number: QTBUG-86866 Change-Id: Ic7ac23de0265561cb06a0ba55089b6c0d3347441 Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-10-02 13:38:47 +00:00
if(NOT arg_CONFIGS)
message(FATAL_ERROR
"You must specify at least one configuration for which to add the flags.")
endif()
if(NOT arg_FLAGS)
message(FATAL_ERROR "You must specify at least one flag to add.")
endif()
if(arg_LANGUAGES)
set(enabled_languages "${arg_LANGUAGES}")
else()
qt_internal_get_enabled_languages_for_flag_manipulation(enabled_languages)
endif()
set(configs ${arg_CONFIGS})
foreach(lang ${enabled_languages})
foreach(config ${configs})
set(flag_var_name "CMAKE_${lang}_FLAGS_${config}")
Prevent compiler and linker flags from growing every cmake run Functions in QtFlagHandlingHelpers.cmake try to update the CMake cache variables for compiler and linker flags. These were using the current value of those variables and writing the modified ones back to the cache every time CMake ran. If a toolchain file sets or modifies any of these variables, that updated value is used and written back into the cache instead of the original cache variable's value. The next time CMake executes, the toolchain file re-applies the same change and the variable grows longer each time with flags repeated. With Ninja, this causes a complete rebuild every time CMake is re-run. The Android NDK toolchain file is one example where this behavior is triggered (the fault is shared, one could argue that the NDK should only be setting ..._INIT variables, but that's out of our control). Another related bug in the previous implementation was that the flags used to build after the first CMake execution could be different to those used for all builds after the second and later CMake runs. This is because the CMake cache was being updated, but not always the calling scope of the functions that modified them. If a toolchain file set any of the compiler or linker flag variables as non-cache variables, then updating the cache variable would have no effect on the calling scope. The non-cache variable would continue to take precedence for that scope for that run. The next time CMake executes though, the updated cache variable would now have been used by the toolchain file and the change *will* be part of the non-cache variable's value. The above are examples of why you should try to avoid updating these cache variables from project code. We could leave the cache alone and always update only non-cache variables, but then a developer looking at the cache may wonder why the values they see there don't match the values being used in builds. Or worse, they think the cache values are being used and don't realize the builds are using something different. Ultimately, we have to choose which downside we are happy to live with. The changes here preserve the previous intent of updating the cache, but it's still a bit fragile. Fixes: QTBUG-89821 Task-number: QTBUG-85992 Task-number: QTBUG-86866 Change-Id: I8a16753e159bde338e6d2e5dd2ce91fc8ac5c39d Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2021-01-07 05:53:43 +00:00
qt_internal_add_flags_impl(${flag_var_name} "${arg_FLAGS}" "${arg_IN_CACHE}")
set(${flag_var_name} "${${flag_var_name}}" PARENT_SCOPE)
CMake: Refactor optimization flag handling and add optimize_full Introduce a bunch of helper functions to manipulate compiler flags and linker flags for the CMAKE_<LANG>_FLAGS_<CONFIG> and CMAKE_<LINK_TYPE>_LINKER_FLAGS_<CONFIG> CMake variables. These variables can be assigned and modified either in the cache or for a specific subdirectory scope, which will apply the flags only to targets in that scope. Add qt_internal_add_optimize_full_flags() function which mimics qmake's CONFIG += optimize_full behavior. Calling it will force usage of the '-O3' optimization flag on supported platforms (falling back '-O2' where not supported). Use the function for the Core and Gui subdirectories, to enable full optimization for the respective Qt modules as it is done in the qmake projects. To ensure that the global qmake-like compiler flags are assigned eveywhere, qt_internal_set_up_config_optimizations_like_in_qmake() needs to be called after Qt global features like optimize_size and optimize_full are available. This means that qtbase and its standalone tests need some special handling in regards to when to call that function. Task-number: QTBUG-86866 Change-Id: Ic7ac23de0265561cb06a0ba55089b6c0d3347441 Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-10-02 13:38:47 +00:00
endforeach()
endforeach()
endfunction()
# Convenience function that adds compiler flags for all release configurations.
Prevent compiler and linker flags from growing every cmake run Functions in QtFlagHandlingHelpers.cmake try to update the CMake cache variables for compiler and linker flags. These were using the current value of those variables and writing the modified ones back to the cache every time CMake ran. If a toolchain file sets or modifies any of these variables, that updated value is used and written back into the cache instead of the original cache variable's value. The next time CMake executes, the toolchain file re-applies the same change and the variable grows longer each time with flags repeated. With Ninja, this causes a complete rebuild every time CMake is re-run. The Android NDK toolchain file is one example where this behavior is triggered (the fault is shared, one could argue that the NDK should only be setting ..._INIT variables, but that's out of our control). Another related bug in the previous implementation was that the flags used to build after the first CMake execution could be different to those used for all builds after the second and later CMake runs. This is because the CMake cache was being updated, but not always the calling scope of the functions that modified them. If a toolchain file set any of the compiler or linker flag variables as non-cache variables, then updating the cache variable would have no effect on the calling scope. The non-cache variable would continue to take precedence for that scope for that run. The next time CMake executes though, the updated cache variable would now have been used by the toolchain file and the change *will* be part of the non-cache variable's value. The above are examples of why you should try to avoid updating these cache variables from project code. We could leave the cache alone and always update only non-cache variables, but then a developer looking at the cache may wonder why the values they see there don't match the values being used in builds. Or worse, they think the cache values are being used and don't realize the builds are using something different. Ultimately, we have to choose which downside we are happy to live with. The changes here preserve the previous intent of updating the cache, but it's still a bit fragile. Fixes: QTBUG-89821 Task-number: QTBUG-85992 Task-number: QTBUG-86866 Change-Id: I8a16753e159bde338e6d2e5dd2ce91fc8ac5c39d Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2021-01-07 05:53:43 +00:00
# The flag variables are always updated in the calling scope, even if they did not
# exist beforehand.
CMake: Refactor optimization flag handling and add optimize_full Introduce a bunch of helper functions to manipulate compiler flags and linker flags for the CMAKE_<LANG>_FLAGS_<CONFIG> and CMAKE_<LINK_TYPE>_LINKER_FLAGS_<CONFIG> CMake variables. These variables can be assigned and modified either in the cache or for a specific subdirectory scope, which will apply the flags only to targets in that scope. Add qt_internal_add_optimize_full_flags() function which mimics qmake's CONFIG += optimize_full behavior. Calling it will force usage of the '-O3' optimization flag on supported platforms (falling back '-O2' where not supported). Use the function for the Core and Gui subdirectories, to enable full optimization for the respective Qt modules as it is done in the qmake projects. To ensure that the global qmake-like compiler flags are assigned eveywhere, qt_internal_set_up_config_optimizations_like_in_qmake() needs to be called after Qt global features like optimize_size and optimize_full are available. This means that qtbase and its standalone tests need some special handling in regards to when to call that function. Task-number: QTBUG-86866 Change-Id: Ic7ac23de0265561cb06a0ba55089b6c0d3347441 Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-10-02 13:38:47 +00:00
#
Prevent compiler and linker flags from growing every cmake run Functions in QtFlagHandlingHelpers.cmake try to update the CMake cache variables for compiler and linker flags. These were using the current value of those variables and writing the modified ones back to the cache every time CMake ran. If a toolchain file sets or modifies any of these variables, that updated value is used and written back into the cache instead of the original cache variable's value. The next time CMake executes, the toolchain file re-applies the same change and the variable grows longer each time with flags repeated. With Ninja, this causes a complete rebuild every time CMake is re-run. The Android NDK toolchain file is one example where this behavior is triggered (the fault is shared, one could argue that the NDK should only be setting ..._INIT variables, but that's out of our control). Another related bug in the previous implementation was that the flags used to build after the first CMake execution could be different to those used for all builds after the second and later CMake runs. This is because the CMake cache was being updated, but not always the calling scope of the functions that modified them. If a toolchain file set any of the compiler or linker flag variables as non-cache variables, then updating the cache variable would have no effect on the calling scope. The non-cache variable would continue to take precedence for that scope for that run. The next time CMake executes though, the updated cache variable would now have been used by the toolchain file and the change *will* be part of the non-cache variable's value. The above are examples of why you should try to avoid updating these cache variables from project code. We could leave the cache alone and always update only non-cache variables, but then a developer looking at the cache may wonder why the values they see there don't match the values being used in builds. Or worse, they think the cache values are being used and don't realize the builds are using something different. Ultimately, we have to choose which downside we are happy to live with. The changes here preserve the previous intent of updating the cache, but it's still a bit fragile. Fixes: QTBUG-89821 Task-number: QTBUG-85992 Task-number: QTBUG-86866 Change-Id: I8a16753e159bde338e6d2e5dd2ce91fc8ac5c39d Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2021-01-07 05:53:43 +00:00
# FLAGS - should be a single string of flags separated by spaces.
# IN_CACHE - add them to the corresponding cache variable too. Note that the cache
# variable may have a different value to the non-cache variable.
# LANGUAGES - optional list of languages like 'C', 'CXX', for which to add the flags
# if not provided, defaults to the list of enabled C-like languages
CMake: Refactor optimization flag handling and add optimize_full Introduce a bunch of helper functions to manipulate compiler flags and linker flags for the CMAKE_<LANG>_FLAGS_<CONFIG> and CMAKE_<LINK_TYPE>_LINKER_FLAGS_<CONFIG> CMake variables. These variables can be assigned and modified either in the cache or for a specific subdirectory scope, which will apply the flags only to targets in that scope. Add qt_internal_add_optimize_full_flags() function which mimics qmake's CONFIG += optimize_full behavior. Calling it will force usage of the '-O3' optimization flag on supported platforms (falling back '-O2' where not supported). Use the function for the Core and Gui subdirectories, to enable full optimization for the respective Qt modules as it is done in the qmake projects. To ensure that the global qmake-like compiler flags are assigned eveywhere, qt_internal_set_up_config_optimizations_like_in_qmake() needs to be called after Qt global features like optimize_size and optimize_full are available. This means that qtbase and its standalone tests need some special handling in regards to when to call that function. Task-number: QTBUG-86866 Change-Id: Ic7ac23de0265561cb06a0ba55089b6c0d3347441 Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-10-02 13:38:47 +00:00
function(qt_internal_add_compiler_flags_for_release_configs)
Retire the qt_parse_all_arguments This replaces the qt_parse_all_arguments macro with the built-in `cmake_parse_arguments(PARSE_ARGV`. In addition, a new function, _qt_internal_validate_all_args_are_parsed, can be used to check whether any _UNPARSED_ARGUMENTS have been passed to the function. Fixes: QTBUG-99238 Change-Id: I8cee83dc92dc6acdaaf747ea6ff9084c11dc649b Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
2023-01-11 14:36:18 +00:00
cmake_parse_arguments(PARSE_ARGV 0 arg
Prevent compiler and linker flags from growing every cmake run Functions in QtFlagHandlingHelpers.cmake try to update the CMake cache variables for compiler and linker flags. These were using the current value of those variables and writing the modified ones back to the cache every time CMake ran. If a toolchain file sets or modifies any of these variables, that updated value is used and written back into the cache instead of the original cache variable's value. The next time CMake executes, the toolchain file re-applies the same change and the variable grows longer each time with flags repeated. With Ninja, this causes a complete rebuild every time CMake is re-run. The Android NDK toolchain file is one example where this behavior is triggered (the fault is shared, one could argue that the NDK should only be setting ..._INIT variables, but that's out of our control). Another related bug in the previous implementation was that the flags used to build after the first CMake execution could be different to those used for all builds after the second and later CMake runs. This is because the CMake cache was being updated, but not always the calling scope of the functions that modified them. If a toolchain file set any of the compiler or linker flag variables as non-cache variables, then updating the cache variable would have no effect on the calling scope. The non-cache variable would continue to take precedence for that scope for that run. The next time CMake executes though, the updated cache variable would now have been used by the toolchain file and the change *will* be part of the non-cache variable's value. The above are examples of why you should try to avoid updating these cache variables from project code. We could leave the cache alone and always update only non-cache variables, but then a developer looking at the cache may wonder why the values they see there don't match the values being used in builds. Or worse, they think the cache values are being used and don't realize the builds are using something different. Ultimately, we have to choose which downside we are happy to live with. The changes here preserve the previous intent of updating the cache, but it's still a bit fragile. Fixes: QTBUG-89821 Task-number: QTBUG-85992 Task-number: QTBUG-86866 Change-Id: I8a16753e159bde338e6d2e5dd2ce91fc8ac5c39d Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2021-01-07 05:53:43 +00:00
"IN_CACHE"
CMake: Refactor optimization flag handling and add optimize_full Introduce a bunch of helper functions to manipulate compiler flags and linker flags for the CMAKE_<LANG>_FLAGS_<CONFIG> and CMAKE_<LINK_TYPE>_LINKER_FLAGS_<CONFIG> CMake variables. These variables can be assigned and modified either in the cache or for a specific subdirectory scope, which will apply the flags only to targets in that scope. Add qt_internal_add_optimize_full_flags() function which mimics qmake's CONFIG += optimize_full behavior. Calling it will force usage of the '-O3' optimization flag on supported platforms (falling back '-O2' where not supported). Use the function for the Core and Gui subdirectories, to enable full optimization for the respective Qt modules as it is done in the qmake projects. To ensure that the global qmake-like compiler flags are assigned eveywhere, qt_internal_set_up_config_optimizations_like_in_qmake() needs to be called after Qt global features like optimize_size and optimize_full are available. This means that qtbase and its standalone tests need some special handling in regards to when to call that function. Task-number: QTBUG-86866 Change-Id: Ic7ac23de0265561cb06a0ba55089b6c0d3347441 Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-10-02 13:38:47 +00:00
"FLAGS"
Retire the qt_parse_all_arguments This replaces the qt_parse_all_arguments macro with the built-in `cmake_parse_arguments(PARSE_ARGV`. In addition, a new function, _qt_internal_validate_all_args_are_parsed, can be used to check whether any _UNPARSED_ARGUMENTS have been passed to the function. Fixes: QTBUG-99238 Change-Id: I8cee83dc92dc6acdaaf747ea6ff9084c11dc649b Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
2023-01-11 14:36:18 +00:00
"LANGUAGES")
_qt_internal_validate_all_args_are_parsed(arg)
CMake: Refactor optimization flag handling and add optimize_full Introduce a bunch of helper functions to manipulate compiler flags and linker flags for the CMAKE_<LANG>_FLAGS_<CONFIG> and CMAKE_<LINK_TYPE>_LINKER_FLAGS_<CONFIG> CMake variables. These variables can be assigned and modified either in the cache or for a specific subdirectory scope, which will apply the flags only to targets in that scope. Add qt_internal_add_optimize_full_flags() function which mimics qmake's CONFIG += optimize_full behavior. Calling it will force usage of the '-O3' optimization flag on supported platforms (falling back '-O2' where not supported). Use the function for the Core and Gui subdirectories, to enable full optimization for the respective Qt modules as it is done in the qmake projects. To ensure that the global qmake-like compiler flags are assigned eveywhere, qt_internal_set_up_config_optimizations_like_in_qmake() needs to be called after Qt global features like optimize_size and optimize_full are available. This means that qtbase and its standalone tests need some special handling in regards to when to call that function. Task-number: QTBUG-86866 Change-Id: Ic7ac23de0265561cb06a0ba55089b6c0d3347441 Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-10-02 13:38:47 +00:00
set(args "")
if(arg_LANGUAGES)
set(enabled_languages "${arg_LANGUAGES}")
else()
qt_internal_get_enabled_languages_for_flag_manipulation(enabled_languages)
endif()
set(configs RELEASE RELWITHDEBINFO MINSIZEREL)
list(APPEND args CONFIGS ${configs})
if(arg_FLAGS)
list(APPEND args FLAGS "${arg_FLAGS}")
endif()
if(arg_IN_CACHE)
list(APPEND args IN_CACHE)
endif()
list(APPEND args LANGUAGES ${enabled_languages})
qt_internal_add_compiler_flags(${args})
Prevent compiler and linker flags from growing every cmake run Functions in QtFlagHandlingHelpers.cmake try to update the CMake cache variables for compiler and linker flags. These were using the current value of those variables and writing the modified ones back to the cache every time CMake ran. If a toolchain file sets or modifies any of these variables, that updated value is used and written back into the cache instead of the original cache variable's value. The next time CMake executes, the toolchain file re-applies the same change and the variable grows longer each time with flags repeated. With Ninja, this causes a complete rebuild every time CMake is re-run. The Android NDK toolchain file is one example where this behavior is triggered (the fault is shared, one could argue that the NDK should only be setting ..._INIT variables, but that's out of our control). Another related bug in the previous implementation was that the flags used to build after the first CMake execution could be different to those used for all builds after the second and later CMake runs. This is because the CMake cache was being updated, but not always the calling scope of the functions that modified them. If a toolchain file set any of the compiler or linker flag variables as non-cache variables, then updating the cache variable would have no effect on the calling scope. The non-cache variable would continue to take precedence for that scope for that run. The next time CMake executes though, the updated cache variable would now have been used by the toolchain file and the change *will* be part of the non-cache variable's value. The above are examples of why you should try to avoid updating these cache variables from project code. We could leave the cache alone and always update only non-cache variables, but then a developer looking at the cache may wonder why the values they see there don't match the values being used in builds. Or worse, they think the cache values are being used and don't realize the builds are using something different. Ultimately, we have to choose which downside we are happy to live with. The changes here preserve the previous intent of updating the cache, but it's still a bit fragile. Fixes: QTBUG-89821 Task-number: QTBUG-85992 Task-number: QTBUG-86866 Change-Id: I8a16753e159bde338e6d2e5dd2ce91fc8ac5c39d Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2021-01-07 05:53:43 +00:00
foreach(lang ${enabled_languages})
foreach(config ${configs})
set(flag_var_name "CMAKE_${lang}_FLAGS_${config}")
set("${flag_var_name}" "${${flag_var_name}}" PARENT_SCOPE)
CMake: Refactor optimization flag handling and add optimize_full Introduce a bunch of helper functions to manipulate compiler flags and linker flags for the CMAKE_<LANG>_FLAGS_<CONFIG> and CMAKE_<LINK_TYPE>_LINKER_FLAGS_<CONFIG> CMake variables. These variables can be assigned and modified either in the cache or for a specific subdirectory scope, which will apply the flags only to targets in that scope. Add qt_internal_add_optimize_full_flags() function which mimics qmake's CONFIG += optimize_full behavior. Calling it will force usage of the '-O3' optimization flag on supported platforms (falling back '-O2' where not supported). Use the function for the Core and Gui subdirectories, to enable full optimization for the respective Qt modules as it is done in the qmake projects. To ensure that the global qmake-like compiler flags are assigned eveywhere, qt_internal_set_up_config_optimizations_like_in_qmake() needs to be called after Qt global features like optimize_size and optimize_full are available. This means that qtbase and its standalone tests need some special handling in regards to when to call that function. Task-number: QTBUG-86866 Change-Id: Ic7ac23de0265561cb06a0ba55089b6c0d3347441 Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-10-02 13:38:47 +00:00
endforeach()
Prevent compiler and linker flags from growing every cmake run Functions in QtFlagHandlingHelpers.cmake try to update the CMake cache variables for compiler and linker flags. These were using the current value of those variables and writing the modified ones back to the cache every time CMake ran. If a toolchain file sets or modifies any of these variables, that updated value is used and written back into the cache instead of the original cache variable's value. The next time CMake executes, the toolchain file re-applies the same change and the variable grows longer each time with flags repeated. With Ninja, this causes a complete rebuild every time CMake is re-run. The Android NDK toolchain file is one example where this behavior is triggered (the fault is shared, one could argue that the NDK should only be setting ..._INIT variables, but that's out of our control). Another related bug in the previous implementation was that the flags used to build after the first CMake execution could be different to those used for all builds after the second and later CMake runs. This is because the CMake cache was being updated, but not always the calling scope of the functions that modified them. If a toolchain file set any of the compiler or linker flag variables as non-cache variables, then updating the cache variable would have no effect on the calling scope. The non-cache variable would continue to take precedence for that scope for that run. The next time CMake executes though, the updated cache variable would now have been used by the toolchain file and the change *will* be part of the non-cache variable's value. The above are examples of why you should try to avoid updating these cache variables from project code. We could leave the cache alone and always update only non-cache variables, but then a developer looking at the cache may wonder why the values they see there don't match the values being used in builds. Or worse, they think the cache values are being used and don't realize the builds are using something different. Ultimately, we have to choose which downside we are happy to live with. The changes here preserve the previous intent of updating the cache, but it's still a bit fragile. Fixes: QTBUG-89821 Task-number: QTBUG-85992 Task-number: QTBUG-86866 Change-Id: I8a16753e159bde338e6d2e5dd2ce91fc8ac5c39d Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2021-01-07 05:53:43 +00:00
endforeach()
CMake: Refactor optimization flag handling and add optimize_full Introduce a bunch of helper functions to manipulate compiler flags and linker flags for the CMAKE_<LANG>_FLAGS_<CONFIG> and CMAKE_<LINK_TYPE>_LINKER_FLAGS_<CONFIG> CMake variables. These variables can be assigned and modified either in the cache or for a specific subdirectory scope, which will apply the flags only to targets in that scope. Add qt_internal_add_optimize_full_flags() function which mimics qmake's CONFIG += optimize_full behavior. Calling it will force usage of the '-O3' optimization flag on supported platforms (falling back '-O2' where not supported). Use the function for the Core and Gui subdirectories, to enable full optimization for the respective Qt modules as it is done in the qmake projects. To ensure that the global qmake-like compiler flags are assigned eveywhere, qt_internal_set_up_config_optimizations_like_in_qmake() needs to be called after Qt global features like optimize_size and optimize_full are available. This means that qtbase and its standalone tests need some special handling in regards to when to call that function. Task-number: QTBUG-86866 Change-Id: Ic7ac23de0265561cb06a0ba55089b6c0d3347441 Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-10-02 13:38:47 +00:00
endfunction()
# Convenience function that replaces all optimization flags with the equivalent of '-O3'
# (optimize_full) flag for all release configs.
#
# This is the equivalent of qmake's CONFIG += optimize_full.
# It is meant to be called in a subdirectory scope to enable full optimizations for a particular
# Qt module, like Core or Gui.
function(qt_internal_add_optimize_full_flags)
Retire the qt_parse_all_arguments This replaces the qt_parse_all_arguments macro with the built-in `cmake_parse_arguments(PARSE_ARGV`. In addition, a new function, _qt_internal_validate_all_args_are_parsed, can be used to check whether any _UNPARSED_ARGUMENTS have been passed to the function. Fixes: QTBUG-99238 Change-Id: I8cee83dc92dc6acdaaf747ea6ff9084c11dc649b Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
2023-01-11 14:36:18 +00:00
cmake_parse_arguments(PARSE_ARGV 0 arg
Prevent compiler and linker flags from growing every cmake run Functions in QtFlagHandlingHelpers.cmake try to update the CMake cache variables for compiler and linker flags. These were using the current value of those variables and writing the modified ones back to the cache every time CMake ran. If a toolchain file sets or modifies any of these variables, that updated value is used and written back into the cache instead of the original cache variable's value. The next time CMake executes, the toolchain file re-applies the same change and the variable grows longer each time with flags repeated. With Ninja, this causes a complete rebuild every time CMake is re-run. The Android NDK toolchain file is one example where this behavior is triggered (the fault is shared, one could argue that the NDK should only be setting ..._INIT variables, but that's out of our control). Another related bug in the previous implementation was that the flags used to build after the first CMake execution could be different to those used for all builds after the second and later CMake runs. This is because the CMake cache was being updated, but not always the calling scope of the functions that modified them. If a toolchain file set any of the compiler or linker flag variables as non-cache variables, then updating the cache variable would have no effect on the calling scope. The non-cache variable would continue to take precedence for that scope for that run. The next time CMake executes though, the updated cache variable would now have been used by the toolchain file and the change *will* be part of the non-cache variable's value. The above are examples of why you should try to avoid updating these cache variables from project code. We could leave the cache alone and always update only non-cache variables, but then a developer looking at the cache may wonder why the values they see there don't match the values being used in builds. Or worse, they think the cache values are being used and don't realize the builds are using something different. Ultimately, we have to choose which downside we are happy to live with. The changes here preserve the previous intent of updating the cache, but it's still a bit fragile. Fixes: QTBUG-89821 Task-number: QTBUG-85992 Task-number: QTBUG-86866 Change-Id: I8a16753e159bde338e6d2e5dd2ce91fc8ac5c39d Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2021-01-07 05:53:43 +00:00
"IN_CACHE"
CMake: Refactor optimization flag handling and add optimize_full Introduce a bunch of helper functions to manipulate compiler flags and linker flags for the CMAKE_<LANG>_FLAGS_<CONFIG> and CMAKE_<LINK_TYPE>_LINKER_FLAGS_<CONFIG> CMake variables. These variables can be assigned and modified either in the cache or for a specific subdirectory scope, which will apply the flags only to targets in that scope. Add qt_internal_add_optimize_full_flags() function which mimics qmake's CONFIG += optimize_full behavior. Calling it will force usage of the '-O3' optimization flag on supported platforms (falling back '-O2' where not supported). Use the function for the Core and Gui subdirectories, to enable full optimization for the respective Qt modules as it is done in the qmake projects. To ensure that the global qmake-like compiler flags are assigned eveywhere, qt_internal_set_up_config_optimizations_like_in_qmake() needs to be called after Qt global features like optimize_size and optimize_full are available. This means that qtbase and its standalone tests need some special handling in regards to when to call that function. Task-number: QTBUG-86866 Change-Id: Ic7ac23de0265561cb06a0ba55089b6c0d3347441 Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-10-02 13:38:47 +00:00
""
Retire the qt_parse_all_arguments This replaces the qt_parse_all_arguments macro with the built-in `cmake_parse_arguments(PARSE_ARGV`. In addition, a new function, _qt_internal_validate_all_args_are_parsed, can be used to check whether any _UNPARSED_ARGUMENTS have been passed to the function. Fixes: QTBUG-99238 Change-Id: I8cee83dc92dc6acdaaf747ea6ff9084c11dc649b Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
2023-01-11 14:36:18 +00:00
"")
_qt_internal_validate_all_args_are_parsed(arg)
CMake: Refactor optimization flag handling and add optimize_full Introduce a bunch of helper functions to manipulate compiler flags and linker flags for the CMAKE_<LANG>_FLAGS_<CONFIG> and CMAKE_<LINK_TYPE>_LINKER_FLAGS_<CONFIG> CMake variables. These variables can be assigned and modified either in the cache or for a specific subdirectory scope, which will apply the flags only to targets in that scope. Add qt_internal_add_optimize_full_flags() function which mimics qmake's CONFIG += optimize_full behavior. Calling it will force usage of the '-O3' optimization flag on supported platforms (falling back '-O2' where not supported). Use the function for the Core and Gui subdirectories, to enable full optimization for the respective Qt modules as it is done in the qmake projects. To ensure that the global qmake-like compiler flags are assigned eveywhere, qt_internal_set_up_config_optimizations_like_in_qmake() needs to be called after Qt global features like optimize_size and optimize_full are available. This means that qtbase and its standalone tests need some special handling in regards to when to call that function. Task-number: QTBUG-86866 Change-Id: Ic7ac23de0265561cb06a0ba55089b6c0d3347441 Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-10-02 13:38:47 +00:00
Disable forced full optimization if QT_USE_DEFAULT_CMAKE_OPTIMIZATION_FLAGS=ON Pick-to: 6.2 Task-number: QTBUG-94753 Change-Id: Ib9dda1a96f3123394e3b3d4f84feac8cedd18b89 Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
2021-06-24 12:18:28 +00:00
# QT_USE_DEFAULT_CMAKE_OPTIMIZATION_FLAGS disables forced full optimization.
if(QT_USE_DEFAULT_CMAKE_OPTIMIZATION_FLAGS)
return()
endif()
Tune compilation flags to reduce libraries size Add '_WINDLL' definition for the MSVC compilers family. Check the 'FEATURE_optimize_size' value before replacing compiler flags in the qt_internal_add_optimize_full_flags function. This is required, because Qt::Core and Qt::Gui modules lost their ability to shrink, when selecting the appropriate build type or features. Fixes: QTBUG-89952 Change-Id: I982c25ea84e4793b4006ead0ee516b3f3eb2a054 Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2021-01-20 18:34:26 +00:00
# Assume that FEATURE_optimize_full has higher priority. But if FEATURE_optimize_full is OFF,
# flags are set by FEATURE_optimize_size should remain unchanged.
if(QT_FEATURE_optimize_size AND NOT QT_FEATURE_optimize_full)
return()
endif()
CMake: Refactor optimization flag handling and add optimize_full Introduce a bunch of helper functions to manipulate compiler flags and linker flags for the CMAKE_<LANG>_FLAGS_<CONFIG> and CMAKE_<LINK_TYPE>_LINKER_FLAGS_<CONFIG> CMake variables. These variables can be assigned and modified either in the cache or for a specific subdirectory scope, which will apply the flags only to targets in that scope. Add qt_internal_add_optimize_full_flags() function which mimics qmake's CONFIG += optimize_full behavior. Calling it will force usage of the '-O3' optimization flag on supported platforms (falling back '-O2' where not supported). Use the function for the Core and Gui subdirectories, to enable full optimization for the respective Qt modules as it is done in the qmake projects. To ensure that the global qmake-like compiler flags are assigned eveywhere, qt_internal_set_up_config_optimizations_like_in_qmake() needs to be called after Qt global features like optimize_size and optimize_full are available. This means that qtbase and its standalone tests need some special handling in regards to when to call that function. Task-number: QTBUG-86866 Change-Id: Ic7ac23de0265561cb06a0ba55089b6c0d3347441 Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-10-02 13:38:47 +00:00
set(args "")
if(arg_IN_CACHE)
list(APPEND args IN_CACHE)
endif()
qt_internal_get_enabled_languages_for_flag_manipulation(enabled_languages)
Tune compilation flags to reduce libraries size Add '_WINDLL' definition for the MSVC compilers family. Check the 'FEATURE_optimize_size' value before replacing compiler flags in the qt_internal_add_optimize_full_flags function. This is required, because Qt::Core and Qt::Gui modules lost their ability to shrink, when selecting the appropriate build type or features. Fixes: QTBUG-89952 Change-Id: I982c25ea84e4793b4006ead0ee516b3f3eb2a054 Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2021-01-20 18:34:26 +00:00
set(configs RELEASE RELWITHDEBINFO)
if(QT_FEATURE_optimize_full) # Assume that FEATURE_optimize_full has higher priority.
list(APPEND configs MINSIZEREL)
endif()
CMake: Refactor optimization flag handling and add optimize_full Introduce a bunch of helper functions to manipulate compiler flags and linker flags for the CMAKE_<LANG>_FLAGS_<CONFIG> and CMAKE_<LINK_TYPE>_LINKER_FLAGS_<CONFIG> CMake variables. These variables can be assigned and modified either in the cache or for a specific subdirectory scope, which will apply the flags only to targets in that scope. Add qt_internal_add_optimize_full_flags() function which mimics qmake's CONFIG += optimize_full behavior. Calling it will force usage of the '-O3' optimization flag on supported platforms (falling back '-O2' where not supported). Use the function for the Core and Gui subdirectories, to enable full optimization for the respective Qt modules as it is done in the qmake projects. To ensure that the global qmake-like compiler flags are assigned eveywhere, qt_internal_set_up_config_optimizations_like_in_qmake() needs to be called after Qt global features like optimize_size and optimize_full are available. This means that qtbase and its standalone tests need some special handling in regards to when to call that function. Task-number: QTBUG-86866 Change-Id: Ic7ac23de0265561cb06a0ba55089b6c0d3347441 Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-10-02 13:38:47 +00:00
qt_internal_remove_known_optimization_flags(${args} CONFIGS ${configs})
# If the respective compiler doesn't have optimize_full flags, use regular optimization flags.
# Mainly MSVC.
qt_internal_get_optimize_full_flags(optimize_full_flags)
list(APPEND args FLAGS "${optimize_full_flags}")
qt_internal_add_compiler_flags_for_release_configs(${args})
Prevent compiler and linker flags from growing every cmake run Functions in QtFlagHandlingHelpers.cmake try to update the CMake cache variables for compiler and linker flags. These were using the current value of those variables and writing the modified ones back to the cache every time CMake ran. If a toolchain file sets or modifies any of these variables, that updated value is used and written back into the cache instead of the original cache variable's value. The next time CMake executes, the toolchain file re-applies the same change and the variable grows longer each time with flags repeated. With Ninja, this causes a complete rebuild every time CMake is re-run. The Android NDK toolchain file is one example where this behavior is triggered (the fault is shared, one could argue that the NDK should only be setting ..._INIT variables, but that's out of our control). Another related bug in the previous implementation was that the flags used to build after the first CMake execution could be different to those used for all builds after the second and later CMake runs. This is because the CMake cache was being updated, but not always the calling scope of the functions that modified them. If a toolchain file set any of the compiler or linker flag variables as non-cache variables, then updating the cache variable would have no effect on the calling scope. The non-cache variable would continue to take precedence for that scope for that run. The next time CMake executes though, the updated cache variable would now have been used by the toolchain file and the change *will* be part of the non-cache variable's value. The above are examples of why you should try to avoid updating these cache variables from project code. We could leave the cache alone and always update only non-cache variables, but then a developer looking at the cache may wonder why the values they see there don't match the values being used in builds. Or worse, they think the cache values are being used and don't realize the builds are using something different. Ultimately, we have to choose which downside we are happy to live with. The changes here preserve the previous intent of updating the cache, but it's still a bit fragile. Fixes: QTBUG-89821 Task-number: QTBUG-85992 Task-number: QTBUG-86866 Change-Id: I8a16753e159bde338e6d2e5dd2ce91fc8ac5c39d Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2021-01-07 05:53:43 +00:00
foreach(lang ${enabled_languages})
foreach(config ${configs})
set(flag_var_name "CMAKE_${lang}_FLAGS_${config}")
set("${flag_var_name}" "${${flag_var_name}}" PARENT_SCOPE)
CMake: Refactor optimization flag handling and add optimize_full Introduce a bunch of helper functions to manipulate compiler flags and linker flags for the CMAKE_<LANG>_FLAGS_<CONFIG> and CMAKE_<LINK_TYPE>_LINKER_FLAGS_<CONFIG> CMake variables. These variables can be assigned and modified either in the cache or for a specific subdirectory scope, which will apply the flags only to targets in that scope. Add qt_internal_add_optimize_full_flags() function which mimics qmake's CONFIG += optimize_full behavior. Calling it will force usage of the '-O3' optimization flag on supported platforms (falling back '-O2' where not supported). Use the function for the Core and Gui subdirectories, to enable full optimization for the respective Qt modules as it is done in the qmake projects. To ensure that the global qmake-like compiler flags are assigned eveywhere, qt_internal_set_up_config_optimizations_like_in_qmake() needs to be called after Qt global features like optimize_size and optimize_full are available. This means that qtbase and its standalone tests need some special handling in regards to when to call that function. Task-number: QTBUG-86866 Change-Id: Ic7ac23de0265561cb06a0ba55089b6c0d3347441 Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-10-02 13:38:47 +00:00
endforeach()
Prevent compiler and linker flags from growing every cmake run Functions in QtFlagHandlingHelpers.cmake try to update the CMake cache variables for compiler and linker flags. These were using the current value of those variables and writing the modified ones back to the cache every time CMake ran. If a toolchain file sets or modifies any of these variables, that updated value is used and written back into the cache instead of the original cache variable's value. The next time CMake executes, the toolchain file re-applies the same change and the variable grows longer each time with flags repeated. With Ninja, this causes a complete rebuild every time CMake is re-run. The Android NDK toolchain file is one example where this behavior is triggered (the fault is shared, one could argue that the NDK should only be setting ..._INIT variables, but that's out of our control). Another related bug in the previous implementation was that the flags used to build after the first CMake execution could be different to those used for all builds after the second and later CMake runs. This is because the CMake cache was being updated, but not always the calling scope of the functions that modified them. If a toolchain file set any of the compiler or linker flag variables as non-cache variables, then updating the cache variable would have no effect on the calling scope. The non-cache variable would continue to take precedence for that scope for that run. The next time CMake executes though, the updated cache variable would now have been used by the toolchain file and the change *will* be part of the non-cache variable's value. The above are examples of why you should try to avoid updating these cache variables from project code. We could leave the cache alone and always update only non-cache variables, but then a developer looking at the cache may wonder why the values they see there don't match the values being used in builds. Or worse, they think the cache values are being used and don't realize the builds are using something different. Ultimately, we have to choose which downside we are happy to live with. The changes here preserve the previous intent of updating the cache, but it's still a bit fragile. Fixes: QTBUG-89821 Task-number: QTBUG-85992 Task-number: QTBUG-86866 Change-Id: I8a16753e159bde338e6d2e5dd2ce91fc8ac5c39d Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2021-01-07 05:53:43 +00:00
endforeach()
CMake: Refactor optimization flag handling and add optimize_full Introduce a bunch of helper functions to manipulate compiler flags and linker flags for the CMAKE_<LANG>_FLAGS_<CONFIG> and CMAKE_<LINK_TYPE>_LINKER_FLAGS_<CONFIG> CMake variables. These variables can be assigned and modified either in the cache or for a specific subdirectory scope, which will apply the flags only to targets in that scope. Add qt_internal_add_optimize_full_flags() function which mimics qmake's CONFIG += optimize_full behavior. Calling it will force usage of the '-O3' optimization flag on supported platforms (falling back '-O2' where not supported). Use the function for the Core and Gui subdirectories, to enable full optimization for the respective Qt modules as it is done in the qmake projects. To ensure that the global qmake-like compiler flags are assigned eveywhere, qt_internal_set_up_config_optimizations_like_in_qmake() needs to be called after Qt global features like optimize_size and optimize_full are available. This means that qtbase and its standalone tests need some special handling in regards to when to call that function. Task-number: QTBUG-86866 Change-Id: Ic7ac23de0265561cb06a0ba55089b6c0d3347441 Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-10-02 13:38:47 +00:00
endfunction()
# Convenience function to replace a compiler flag with another one, for the given configurations
# for all enabled 'safe' languages.
# Essentially a glorified string(REPLACE).
# Can be used to remove compiler flags.
Prevent compiler and linker flags from growing every cmake run Functions in QtFlagHandlingHelpers.cmake try to update the CMake cache variables for compiler and linker flags. These were using the current value of those variables and writing the modified ones back to the cache every time CMake ran. If a toolchain file sets or modifies any of these variables, that updated value is used and written back into the cache instead of the original cache variable's value. The next time CMake executes, the toolchain file re-applies the same change and the variable grows longer each time with flags repeated. With Ninja, this causes a complete rebuild every time CMake is re-run. The Android NDK toolchain file is one example where this behavior is triggered (the fault is shared, one could argue that the NDK should only be setting ..._INIT variables, but that's out of our control). Another related bug in the previous implementation was that the flags used to build after the first CMake execution could be different to those used for all builds after the second and later CMake runs. This is because the CMake cache was being updated, but not always the calling scope of the functions that modified them. If a toolchain file set any of the compiler or linker flag variables as non-cache variables, then updating the cache variable would have no effect on the calling scope. The non-cache variable would continue to take precedence for that scope for that run. The next time CMake executes though, the updated cache variable would now have been used by the toolchain file and the change *will* be part of the non-cache variable's value. The above are examples of why you should try to avoid updating these cache variables from project code. We could leave the cache alone and always update only non-cache variables, but then a developer looking at the cache may wonder why the values they see there don't match the values being used in builds. Or worse, they think the cache values are being used and don't realize the builds are using something different. Ultimately, we have to choose which downside we are happy to live with. The changes here preserve the previous intent of updating the cache, but it's still a bit fragile. Fixes: QTBUG-89821 Task-number: QTBUG-85992 Task-number: QTBUG-86866 Change-Id: I8a16753e159bde338e6d2e5dd2ce91fc8ac5c39d Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2021-01-07 05:53:43 +00:00
# The flag variables are always updated in the calling scope, even if they did not
# exist beforehand.
CMake: Refactor optimization flag handling and add optimize_full Introduce a bunch of helper functions to manipulate compiler flags and linker flags for the CMAKE_<LANG>_FLAGS_<CONFIG> and CMAKE_<LINK_TYPE>_LINKER_FLAGS_<CONFIG> CMake variables. These variables can be assigned and modified either in the cache or for a specific subdirectory scope, which will apply the flags only to targets in that scope. Add qt_internal_add_optimize_full_flags() function which mimics qmake's CONFIG += optimize_full behavior. Calling it will force usage of the '-O3' optimization flag on supported platforms (falling back '-O2' where not supported). Use the function for the Core and Gui subdirectories, to enable full optimization for the respective Qt modules as it is done in the qmake projects. To ensure that the global qmake-like compiler flags are assigned eveywhere, qt_internal_set_up_config_optimizations_like_in_qmake() needs to be called after Qt global features like optimize_size and optimize_full are available. This means that qtbase and its standalone tests need some special handling in regards to when to call that function. Task-number: QTBUG-86866 Change-Id: Ic7ac23de0265561cb06a0ba55089b6c0d3347441 Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-10-02 13:38:47 +00:00
#
Prevent compiler and linker flags from growing every cmake run Functions in QtFlagHandlingHelpers.cmake try to update the CMake cache variables for compiler and linker flags. These were using the current value of those variables and writing the modified ones back to the cache every time CMake ran. If a toolchain file sets or modifies any of these variables, that updated value is used and written back into the cache instead of the original cache variable's value. The next time CMake executes, the toolchain file re-applies the same change and the variable grows longer each time with flags repeated. With Ninja, this causes a complete rebuild every time CMake is re-run. The Android NDK toolchain file is one example where this behavior is triggered (the fault is shared, one could argue that the NDK should only be setting ..._INIT variables, but that's out of our control). Another related bug in the previous implementation was that the flags used to build after the first CMake execution could be different to those used for all builds after the second and later CMake runs. This is because the CMake cache was being updated, but not always the calling scope of the functions that modified them. If a toolchain file set any of the compiler or linker flag variables as non-cache variables, then updating the cache variable would have no effect on the calling scope. The non-cache variable would continue to take precedence for that scope for that run. The next time CMake executes though, the updated cache variable would now have been used by the toolchain file and the change *will* be part of the non-cache variable's value. The above are examples of why you should try to avoid updating these cache variables from project code. We could leave the cache alone and always update only non-cache variables, but then a developer looking at the cache may wonder why the values they see there don't match the values being used in builds. Or worse, they think the cache values are being used and don't realize the builds are using something different. Ultimately, we have to choose which downside we are happy to live with. The changes here preserve the previous intent of updating the cache, but it's still a bit fragile. Fixes: QTBUG-89821 Task-number: QTBUG-85992 Task-number: QTBUG-86866 Change-Id: I8a16753e159bde338e6d2e5dd2ce91fc8ac5c39d Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2021-01-07 05:53:43 +00:00
# match_string - string to match
# replace_string - replacement string
# IN_CACHE - replace them in the corresponding cache variable too. Note that the cache
# variable may have a different value to the non-cache variable.
# CONFIGS - should be a list of upper case configs like DEBUG, RELEASE, RELWITHDEBINFO.
# LANGUAGES - optional list of languages like 'C', 'CXX', for which to replace the flags
# if not provided, defaults to the list of enabled C-like languages
CMake: Refactor optimization flag handling and add optimize_full Introduce a bunch of helper functions to manipulate compiler flags and linker flags for the CMAKE_<LANG>_FLAGS_<CONFIG> and CMAKE_<LINK_TYPE>_LINKER_FLAGS_<CONFIG> CMake variables. These variables can be assigned and modified either in the cache or for a specific subdirectory scope, which will apply the flags only to targets in that scope. Add qt_internal_add_optimize_full_flags() function which mimics qmake's CONFIG += optimize_full behavior. Calling it will force usage of the '-O3' optimization flag on supported platforms (falling back '-O2' where not supported). Use the function for the Core and Gui subdirectories, to enable full optimization for the respective Qt modules as it is done in the qmake projects. To ensure that the global qmake-like compiler flags are assigned eveywhere, qt_internal_set_up_config_optimizations_like_in_qmake() needs to be called after Qt global features like optimize_size and optimize_full are available. This means that qtbase and its standalone tests need some special handling in regards to when to call that function. Task-number: QTBUG-86866 Change-Id: Ic7ac23de0265561cb06a0ba55089b6c0d3347441 Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-10-02 13:38:47 +00:00
function(qt_internal_replace_compiler_flags match_string replace_string)
Retire the qt_parse_all_arguments This replaces the qt_parse_all_arguments macro with the built-in `cmake_parse_arguments(PARSE_ARGV`. In addition, a new function, _qt_internal_validate_all_args_are_parsed, can be used to check whether any _UNPARSED_ARGUMENTS have been passed to the function. Fixes: QTBUG-99238 Change-Id: I8cee83dc92dc6acdaaf747ea6ff9084c11dc649b Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
2023-01-11 14:36:18 +00:00
cmake_parse_arguments(PARSE_ARGV 2 arg
Prevent compiler and linker flags from growing every cmake run Functions in QtFlagHandlingHelpers.cmake try to update the CMake cache variables for compiler and linker flags. These were using the current value of those variables and writing the modified ones back to the cache every time CMake ran. If a toolchain file sets or modifies any of these variables, that updated value is used and written back into the cache instead of the original cache variable's value. The next time CMake executes, the toolchain file re-applies the same change and the variable grows longer each time with flags repeated. With Ninja, this causes a complete rebuild every time CMake is re-run. The Android NDK toolchain file is one example where this behavior is triggered (the fault is shared, one could argue that the NDK should only be setting ..._INIT variables, but that's out of our control). Another related bug in the previous implementation was that the flags used to build after the first CMake execution could be different to those used for all builds after the second and later CMake runs. This is because the CMake cache was being updated, but not always the calling scope of the functions that modified them. If a toolchain file set any of the compiler or linker flag variables as non-cache variables, then updating the cache variable would have no effect on the calling scope. The non-cache variable would continue to take precedence for that scope for that run. The next time CMake executes though, the updated cache variable would now have been used by the toolchain file and the change *will* be part of the non-cache variable's value. The above are examples of why you should try to avoid updating these cache variables from project code. We could leave the cache alone and always update only non-cache variables, but then a developer looking at the cache may wonder why the values they see there don't match the values being used in builds. Or worse, they think the cache values are being used and don't realize the builds are using something different. Ultimately, we have to choose which downside we are happy to live with. The changes here preserve the previous intent of updating the cache, but it's still a bit fragile. Fixes: QTBUG-89821 Task-number: QTBUG-85992 Task-number: QTBUG-86866 Change-Id: I8a16753e159bde338e6d2e5dd2ce91fc8ac5c39d Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2021-01-07 05:53:43 +00:00
"IN_CACHE"
CMake: Refactor optimization flag handling and add optimize_full Introduce a bunch of helper functions to manipulate compiler flags and linker flags for the CMAKE_<LANG>_FLAGS_<CONFIG> and CMAKE_<LINK_TYPE>_LINKER_FLAGS_<CONFIG> CMake variables. These variables can be assigned and modified either in the cache or for a specific subdirectory scope, which will apply the flags only to targets in that scope. Add qt_internal_add_optimize_full_flags() function which mimics qmake's CONFIG += optimize_full behavior. Calling it will force usage of the '-O3' optimization flag on supported platforms (falling back '-O2' where not supported). Use the function for the Core and Gui subdirectories, to enable full optimization for the respective Qt modules as it is done in the qmake projects. To ensure that the global qmake-like compiler flags are assigned eveywhere, qt_internal_set_up_config_optimizations_like_in_qmake() needs to be called after Qt global features like optimize_size and optimize_full are available. This means that qtbase and its standalone tests need some special handling in regards to when to call that function. Task-number: QTBUG-86866 Change-Id: Ic7ac23de0265561cb06a0ba55089b6c0d3347441 Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-10-02 13:38:47 +00:00
""
Retire the qt_parse_all_arguments This replaces the qt_parse_all_arguments macro with the built-in `cmake_parse_arguments(PARSE_ARGV`. In addition, a new function, _qt_internal_validate_all_args_are_parsed, can be used to check whether any _UNPARSED_ARGUMENTS have been passed to the function. Fixes: QTBUG-99238 Change-Id: I8cee83dc92dc6acdaaf747ea6ff9084c11dc649b Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
2023-01-11 14:36:18 +00:00
"CONFIGS;LANGUAGES")
_qt_internal_validate_all_args_are_parsed(arg)
CMake: Refactor optimization flag handling and add optimize_full Introduce a bunch of helper functions to manipulate compiler flags and linker flags for the CMAKE_<LANG>_FLAGS_<CONFIG> and CMAKE_<LINK_TYPE>_LINKER_FLAGS_<CONFIG> CMake variables. These variables can be assigned and modified either in the cache or for a specific subdirectory scope, which will apply the flags only to targets in that scope. Add qt_internal_add_optimize_full_flags() function which mimics qmake's CONFIG += optimize_full behavior. Calling it will force usage of the '-O3' optimization flag on supported platforms (falling back '-O2' where not supported). Use the function for the Core and Gui subdirectories, to enable full optimization for the respective Qt modules as it is done in the qmake projects. To ensure that the global qmake-like compiler flags are assigned eveywhere, qt_internal_set_up_config_optimizations_like_in_qmake() needs to be called after Qt global features like optimize_size and optimize_full are available. This means that qtbase and its standalone tests need some special handling in regards to when to call that function. Task-number: QTBUG-86866 Change-Id: Ic7ac23de0265561cb06a0ba55089b6c0d3347441 Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-10-02 13:38:47 +00:00
if(NOT arg_CONFIGS)
message(FATAL_ERROR
"You must specify at least one configuration for which to replace the flags.")
endif()
if(arg_LANGUAGES)
set(enabled_languages "${arg_LANGUAGES}")
else()
qt_internal_get_enabled_languages_for_flag_manipulation(enabled_languages)
endif()
set(configs ${arg_CONFIGS})
foreach(lang ${enabled_languages})
foreach(config ${configs})
set(flag_var_name "CMAKE_${lang}_FLAGS_${config}")
Prevent compiler and linker flags from growing every cmake run Functions in QtFlagHandlingHelpers.cmake try to update the CMake cache variables for compiler and linker flags. These were using the current value of those variables and writing the modified ones back to the cache every time CMake ran. If a toolchain file sets or modifies any of these variables, that updated value is used and written back into the cache instead of the original cache variable's value. The next time CMake executes, the toolchain file re-applies the same change and the variable grows longer each time with flags repeated. With Ninja, this causes a complete rebuild every time CMake is re-run. The Android NDK toolchain file is one example where this behavior is triggered (the fault is shared, one could argue that the NDK should only be setting ..._INIT variables, but that's out of our control). Another related bug in the previous implementation was that the flags used to build after the first CMake execution could be different to those used for all builds after the second and later CMake runs. This is because the CMake cache was being updated, but not always the calling scope of the functions that modified them. If a toolchain file set any of the compiler or linker flag variables as non-cache variables, then updating the cache variable would have no effect on the calling scope. The non-cache variable would continue to take precedence for that scope for that run. The next time CMake executes though, the updated cache variable would now have been used by the toolchain file and the change *will* be part of the non-cache variable's value. The above are examples of why you should try to avoid updating these cache variables from project code. We could leave the cache alone and always update only non-cache variables, but then a developer looking at the cache may wonder why the values they see there don't match the values being used in builds. Or worse, they think the cache values are being used and don't realize the builds are using something different. Ultimately, we have to choose which downside we are happy to live with. The changes here preserve the previous intent of updating the cache, but it's still a bit fragile. Fixes: QTBUG-89821 Task-number: QTBUG-85992 Task-number: QTBUG-86866 Change-Id: I8a16753e159bde338e6d2e5dd2ce91fc8ac5c39d Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2021-01-07 05:53:43 +00:00
qt_internal_replace_flags_impl(${flag_var_name}
"${match_string}" "${replace_string}" "${arg_IN_CACHE}")
set(${flag_var_name} "${${flag_var_name}}" PARENT_SCOPE)
CMake: Refactor optimization flag handling and add optimize_full Introduce a bunch of helper functions to manipulate compiler flags and linker flags for the CMAKE_<LANG>_FLAGS_<CONFIG> and CMAKE_<LINK_TYPE>_LINKER_FLAGS_<CONFIG> CMake variables. These variables can be assigned and modified either in the cache or for a specific subdirectory scope, which will apply the flags only to targets in that scope. Add qt_internal_add_optimize_full_flags() function which mimics qmake's CONFIG += optimize_full behavior. Calling it will force usage of the '-O3' optimization flag on supported platforms (falling back '-O2' where not supported). Use the function for the Core and Gui subdirectories, to enable full optimization for the respective Qt modules as it is done in the qmake projects. To ensure that the global qmake-like compiler flags are assigned eveywhere, qt_internal_set_up_config_optimizations_like_in_qmake() needs to be called after Qt global features like optimize_size and optimize_full are available. This means that qtbase and its standalone tests need some special handling in regards to when to call that function. Task-number: QTBUG-86866 Change-Id: Ic7ac23de0265561cb06a0ba55089b6c0d3347441 Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-10-02 13:38:47 +00:00
endforeach()
endforeach()
endfunction()
# Convenience function to add linker flags, for the given configurations and target link types.
Prevent compiler and linker flags from growing every cmake run Functions in QtFlagHandlingHelpers.cmake try to update the CMake cache variables for compiler and linker flags. These were using the current value of those variables and writing the modified ones back to the cache every time CMake ran. If a toolchain file sets or modifies any of these variables, that updated value is used and written back into the cache instead of the original cache variable's value. The next time CMake executes, the toolchain file re-applies the same change and the variable grows longer each time with flags repeated. With Ninja, this causes a complete rebuild every time CMake is re-run. The Android NDK toolchain file is one example where this behavior is triggered (the fault is shared, one could argue that the NDK should only be setting ..._INIT variables, but that's out of our control). Another related bug in the previous implementation was that the flags used to build after the first CMake execution could be different to those used for all builds after the second and later CMake runs. This is because the CMake cache was being updated, but not always the calling scope of the functions that modified them. If a toolchain file set any of the compiler or linker flag variables as non-cache variables, then updating the cache variable would have no effect on the calling scope. The non-cache variable would continue to take precedence for that scope for that run. The next time CMake executes though, the updated cache variable would now have been used by the toolchain file and the change *will* be part of the non-cache variable's value. The above are examples of why you should try to avoid updating these cache variables from project code. We could leave the cache alone and always update only non-cache variables, but then a developer looking at the cache may wonder why the values they see there don't match the values being used in builds. Or worse, they think the cache values are being used and don't realize the builds are using something different. Ultimately, we have to choose which downside we are happy to live with. The changes here preserve the previous intent of updating the cache, but it's still a bit fragile. Fixes: QTBUG-89821 Task-number: QTBUG-85992 Task-number: QTBUG-86866 Change-Id: I8a16753e159bde338e6d2e5dd2ce91fc8ac5c39d Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2021-01-07 05:53:43 +00:00
# The flag variables are always updated in the calling scope, even if they did not exist beforehand.
CMake: Refactor optimization flag handling and add optimize_full Introduce a bunch of helper functions to manipulate compiler flags and linker flags for the CMAKE_<LANG>_FLAGS_<CONFIG> and CMAKE_<LINK_TYPE>_LINKER_FLAGS_<CONFIG> CMake variables. These variables can be assigned and modified either in the cache or for a specific subdirectory scope, which will apply the flags only to targets in that scope. Add qt_internal_add_optimize_full_flags() function which mimics qmake's CONFIG += optimize_full behavior. Calling it will force usage of the '-O3' optimization flag on supported platforms (falling back '-O2' where not supported). Use the function for the Core and Gui subdirectories, to enable full optimization for the respective Qt modules as it is done in the qmake projects. To ensure that the global qmake-like compiler flags are assigned eveywhere, qt_internal_set_up_config_optimizations_like_in_qmake() needs to be called after Qt global features like optimize_size and optimize_full are available. This means that qtbase and its standalone tests need some special handling in regards to when to call that function. Task-number: QTBUG-86866 Change-Id: Ic7ac23de0265561cb06a0ba55089b6c0d3347441 Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-10-02 13:38:47 +00:00
#
Prevent compiler and linker flags from growing every cmake run Functions in QtFlagHandlingHelpers.cmake try to update the CMake cache variables for compiler and linker flags. These were using the current value of those variables and writing the modified ones back to the cache every time CMake ran. If a toolchain file sets or modifies any of these variables, that updated value is used and written back into the cache instead of the original cache variable's value. The next time CMake executes, the toolchain file re-applies the same change and the variable grows longer each time with flags repeated. With Ninja, this causes a complete rebuild every time CMake is re-run. The Android NDK toolchain file is one example where this behavior is triggered (the fault is shared, one could argue that the NDK should only be setting ..._INIT variables, but that's out of our control). Another related bug in the previous implementation was that the flags used to build after the first CMake execution could be different to those used for all builds after the second and later CMake runs. This is because the CMake cache was being updated, but not always the calling scope of the functions that modified them. If a toolchain file set any of the compiler or linker flag variables as non-cache variables, then updating the cache variable would have no effect on the calling scope. The non-cache variable would continue to take precedence for that scope for that run. The next time CMake executes though, the updated cache variable would now have been used by the toolchain file and the change *will* be part of the non-cache variable's value. The above are examples of why you should try to avoid updating these cache variables from project code. We could leave the cache alone and always update only non-cache variables, but then a developer looking at the cache may wonder why the values they see there don't match the values being used in builds. Or worse, they think the cache values are being used and don't realize the builds are using something different. Ultimately, we have to choose which downside we are happy to live with. The changes here preserve the previous intent of updating the cache, but it's still a bit fragile. Fixes: QTBUG-89821 Task-number: QTBUG-85992 Task-number: QTBUG-86866 Change-Id: I8a16753e159bde338e6d2e5dd2ce91fc8ac5c39d Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2021-01-07 05:53:43 +00:00
# FLAGS - should be a single string of flags separated by spaces.
# IN_CACHE - add them to the corresponding cache variable too. Note that the cache
# variable may have a different value to the non-cache variable.
# CONFIGS - should be a list of upper case configs like DEBUG, RELEASE, RELWITHDEBINFO.
# TYPES - should be a list of target link types as expected by CMake's
# CMAKE_<LINKER_TYPE>_LINKER_FLAGS_<CONFIG> cache variable.
# e.g EXE, MODULE, SHARED, STATIC.
CMake: Refactor optimization flag handling and add optimize_full Introduce a bunch of helper functions to manipulate compiler flags and linker flags for the CMAKE_<LANG>_FLAGS_<CONFIG> and CMAKE_<LINK_TYPE>_LINKER_FLAGS_<CONFIG> CMake variables. These variables can be assigned and modified either in the cache or for a specific subdirectory scope, which will apply the flags only to targets in that scope. Add qt_internal_add_optimize_full_flags() function which mimics qmake's CONFIG += optimize_full behavior. Calling it will force usage of the '-O3' optimization flag on supported platforms (falling back '-O2' where not supported). Use the function for the Core and Gui subdirectories, to enable full optimization for the respective Qt modules as it is done in the qmake projects. To ensure that the global qmake-like compiler flags are assigned eveywhere, qt_internal_set_up_config_optimizations_like_in_qmake() needs to be called after Qt global features like optimize_size and optimize_full are available. This means that qtbase and its standalone tests need some special handling in regards to when to call that function. Task-number: QTBUG-86866 Change-Id: Ic7ac23de0265561cb06a0ba55089b6c0d3347441 Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-10-02 13:38:47 +00:00
function(qt_internal_add_linker_flags)
Retire the qt_parse_all_arguments This replaces the qt_parse_all_arguments macro with the built-in `cmake_parse_arguments(PARSE_ARGV`. In addition, a new function, _qt_internal_validate_all_args_are_parsed, can be used to check whether any _UNPARSED_ARGUMENTS have been passed to the function. Fixes: QTBUG-99238 Change-Id: I8cee83dc92dc6acdaaf747ea6ff9084c11dc649b Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
2023-01-11 14:36:18 +00:00
cmake_parse_arguments(PARSE_ARGV 0 arg
Prevent compiler and linker flags from growing every cmake run Functions in QtFlagHandlingHelpers.cmake try to update the CMake cache variables for compiler and linker flags. These were using the current value of those variables and writing the modified ones back to the cache every time CMake ran. If a toolchain file sets or modifies any of these variables, that updated value is used and written back into the cache instead of the original cache variable's value. The next time CMake executes, the toolchain file re-applies the same change and the variable grows longer each time with flags repeated. With Ninja, this causes a complete rebuild every time CMake is re-run. The Android NDK toolchain file is one example where this behavior is triggered (the fault is shared, one could argue that the NDK should only be setting ..._INIT variables, but that's out of our control). Another related bug in the previous implementation was that the flags used to build after the first CMake execution could be different to those used for all builds after the second and later CMake runs. This is because the CMake cache was being updated, but not always the calling scope of the functions that modified them. If a toolchain file set any of the compiler or linker flag variables as non-cache variables, then updating the cache variable would have no effect on the calling scope. The non-cache variable would continue to take precedence for that scope for that run. The next time CMake executes though, the updated cache variable would now have been used by the toolchain file and the change *will* be part of the non-cache variable's value. The above are examples of why you should try to avoid updating these cache variables from project code. We could leave the cache alone and always update only non-cache variables, but then a developer looking at the cache may wonder why the values they see there don't match the values being used in builds. Or worse, they think the cache values are being used and don't realize the builds are using something different. Ultimately, we have to choose which downside we are happy to live with. The changes here preserve the previous intent of updating the cache, but it's still a bit fragile. Fixes: QTBUG-89821 Task-number: QTBUG-85992 Task-number: QTBUG-86866 Change-Id: I8a16753e159bde338e6d2e5dd2ce91fc8ac5c39d Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2021-01-07 05:53:43 +00:00
"IN_CACHE"
CMake: Refactor optimization flag handling and add optimize_full Introduce a bunch of helper functions to manipulate compiler flags and linker flags for the CMAKE_<LANG>_FLAGS_<CONFIG> and CMAKE_<LINK_TYPE>_LINKER_FLAGS_<CONFIG> CMake variables. These variables can be assigned and modified either in the cache or for a specific subdirectory scope, which will apply the flags only to targets in that scope. Add qt_internal_add_optimize_full_flags() function which mimics qmake's CONFIG += optimize_full behavior. Calling it will force usage of the '-O3' optimization flag on supported platforms (falling back '-O2' where not supported). Use the function for the Core and Gui subdirectories, to enable full optimization for the respective Qt modules as it is done in the qmake projects. To ensure that the global qmake-like compiler flags are assigned eveywhere, qt_internal_set_up_config_optimizations_like_in_qmake() needs to be called after Qt global features like optimize_size and optimize_full are available. This means that qtbase and its standalone tests need some special handling in regards to when to call that function. Task-number: QTBUG-86866 Change-Id: Ic7ac23de0265561cb06a0ba55089b6c0d3347441 Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-10-02 13:38:47 +00:00
"FLAGS"
Retire the qt_parse_all_arguments This replaces the qt_parse_all_arguments macro with the built-in `cmake_parse_arguments(PARSE_ARGV`. In addition, a new function, _qt_internal_validate_all_args_are_parsed, can be used to check whether any _UNPARSED_ARGUMENTS have been passed to the function. Fixes: QTBUG-99238 Change-Id: I8cee83dc92dc6acdaaf747ea6ff9084c11dc649b Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
2023-01-11 14:36:18 +00:00
"CONFIGS;TYPES")
_qt_internal_validate_all_args_are_parsed(arg)
CMake: Refactor optimization flag handling and add optimize_full Introduce a bunch of helper functions to manipulate compiler flags and linker flags for the CMAKE_<LANG>_FLAGS_<CONFIG> and CMAKE_<LINK_TYPE>_LINKER_FLAGS_<CONFIG> CMake variables. These variables can be assigned and modified either in the cache or for a specific subdirectory scope, which will apply the flags only to targets in that scope. Add qt_internal_add_optimize_full_flags() function which mimics qmake's CONFIG += optimize_full behavior. Calling it will force usage of the '-O3' optimization flag on supported platforms (falling back '-O2' where not supported). Use the function for the Core and Gui subdirectories, to enable full optimization for the respective Qt modules as it is done in the qmake projects. To ensure that the global qmake-like compiler flags are assigned eveywhere, qt_internal_set_up_config_optimizations_like_in_qmake() needs to be called after Qt global features like optimize_size and optimize_full are available. This means that qtbase and its standalone tests need some special handling in regards to when to call that function. Task-number: QTBUG-86866 Change-Id: Ic7ac23de0265561cb06a0ba55089b6c0d3347441 Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-10-02 13:38:47 +00:00
if(NOT arg_TYPES)
message(FATAL_ERROR
"You must specify at least one linker target type for which to add the flags.")
endif()
if(NOT arg_CONFIGS)
message(FATAL_ERROR
"You must specify at least one configuration for which to add the flags.")
endif()
if(NOT arg_FLAGS)
message(FATAL_ERROR "You must specify at least one flag to add.")
endif()
set(configs ${arg_CONFIGS})
set(target_link_types ${arg_TYPES})
foreach(config ${configs})
foreach(t ${target_link_types})
set(flag_var_name "CMAKE_${t}_LINKER_FLAGS_${config}")
Prevent compiler and linker flags from growing every cmake run Functions in QtFlagHandlingHelpers.cmake try to update the CMake cache variables for compiler and linker flags. These were using the current value of those variables and writing the modified ones back to the cache every time CMake ran. If a toolchain file sets or modifies any of these variables, that updated value is used and written back into the cache instead of the original cache variable's value. The next time CMake executes, the toolchain file re-applies the same change and the variable grows longer each time with flags repeated. With Ninja, this causes a complete rebuild every time CMake is re-run. The Android NDK toolchain file is one example where this behavior is triggered (the fault is shared, one could argue that the NDK should only be setting ..._INIT variables, but that's out of our control). Another related bug in the previous implementation was that the flags used to build after the first CMake execution could be different to those used for all builds after the second and later CMake runs. This is because the CMake cache was being updated, but not always the calling scope of the functions that modified them. If a toolchain file set any of the compiler or linker flag variables as non-cache variables, then updating the cache variable would have no effect on the calling scope. The non-cache variable would continue to take precedence for that scope for that run. The next time CMake executes though, the updated cache variable would now have been used by the toolchain file and the change *will* be part of the non-cache variable's value. The above are examples of why you should try to avoid updating these cache variables from project code. We could leave the cache alone and always update only non-cache variables, but then a developer looking at the cache may wonder why the values they see there don't match the values being used in builds. Or worse, they think the cache values are being used and don't realize the builds are using something different. Ultimately, we have to choose which downside we are happy to live with. The changes here preserve the previous intent of updating the cache, but it's still a bit fragile. Fixes: QTBUG-89821 Task-number: QTBUG-85992 Task-number: QTBUG-86866 Change-Id: I8a16753e159bde338e6d2e5dd2ce91fc8ac5c39d Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2021-01-07 05:53:43 +00:00
qt_internal_add_flags_impl(${flag_var_name} "${arg_FLAGS}" "${arg_IN_CACHE}")
set(${flag_var_name} "${${flag_var_name}}" PARENT_SCOPE)
CMake: Refactor optimization flag handling and add optimize_full Introduce a bunch of helper functions to manipulate compiler flags and linker flags for the CMAKE_<LANG>_FLAGS_<CONFIG> and CMAKE_<LINK_TYPE>_LINKER_FLAGS_<CONFIG> CMake variables. These variables can be assigned and modified either in the cache or for a specific subdirectory scope, which will apply the flags only to targets in that scope. Add qt_internal_add_optimize_full_flags() function which mimics qmake's CONFIG += optimize_full behavior. Calling it will force usage of the '-O3' optimization flag on supported platforms (falling back '-O2' where not supported). Use the function for the Core and Gui subdirectories, to enable full optimization for the respective Qt modules as it is done in the qmake projects. To ensure that the global qmake-like compiler flags are assigned eveywhere, qt_internal_set_up_config_optimizations_like_in_qmake() needs to be called after Qt global features like optimize_size and optimize_full are available. This means that qtbase and its standalone tests need some special handling in regards to when to call that function. Task-number: QTBUG-86866 Change-Id: Ic7ac23de0265561cb06a0ba55089b6c0d3347441 Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-10-02 13:38:47 +00:00
endforeach()
endforeach()
endfunction()
# Convenience function to replace a linker flag with another one, for the given configurations
# and target link types.
# Essentially a glorified string(REPLACE).
# Can be used to remove linker flags.
Prevent compiler and linker flags from growing every cmake run Functions in QtFlagHandlingHelpers.cmake try to update the CMake cache variables for compiler and linker flags. These were using the current value of those variables and writing the modified ones back to the cache every time CMake ran. If a toolchain file sets or modifies any of these variables, that updated value is used and written back into the cache instead of the original cache variable's value. The next time CMake executes, the toolchain file re-applies the same change and the variable grows longer each time with flags repeated. With Ninja, this causes a complete rebuild every time CMake is re-run. The Android NDK toolchain file is one example where this behavior is triggered (the fault is shared, one could argue that the NDK should only be setting ..._INIT variables, but that's out of our control). Another related bug in the previous implementation was that the flags used to build after the first CMake execution could be different to those used for all builds after the second and later CMake runs. This is because the CMake cache was being updated, but not always the calling scope of the functions that modified them. If a toolchain file set any of the compiler or linker flag variables as non-cache variables, then updating the cache variable would have no effect on the calling scope. The non-cache variable would continue to take precedence for that scope for that run. The next time CMake executes though, the updated cache variable would now have been used by the toolchain file and the change *will* be part of the non-cache variable's value. The above are examples of why you should try to avoid updating these cache variables from project code. We could leave the cache alone and always update only non-cache variables, but then a developer looking at the cache may wonder why the values they see there don't match the values being used in builds. Or worse, they think the cache values are being used and don't realize the builds are using something different. Ultimately, we have to choose which downside we are happy to live with. The changes here preserve the previous intent of updating the cache, but it's still a bit fragile. Fixes: QTBUG-89821 Task-number: QTBUG-85992 Task-number: QTBUG-86866 Change-Id: I8a16753e159bde338e6d2e5dd2ce91fc8ac5c39d Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2021-01-07 05:53:43 +00:00
# The flag variables are always updated in the calling scope, even if they did not exist beforehand.
CMake: Refactor optimization flag handling and add optimize_full Introduce a bunch of helper functions to manipulate compiler flags and linker flags for the CMAKE_<LANG>_FLAGS_<CONFIG> and CMAKE_<LINK_TYPE>_LINKER_FLAGS_<CONFIG> CMake variables. These variables can be assigned and modified either in the cache or for a specific subdirectory scope, which will apply the flags only to targets in that scope. Add qt_internal_add_optimize_full_flags() function which mimics qmake's CONFIG += optimize_full behavior. Calling it will force usage of the '-O3' optimization flag on supported platforms (falling back '-O2' where not supported). Use the function for the Core and Gui subdirectories, to enable full optimization for the respective Qt modules as it is done in the qmake projects. To ensure that the global qmake-like compiler flags are assigned eveywhere, qt_internal_set_up_config_optimizations_like_in_qmake() needs to be called after Qt global features like optimize_size and optimize_full are available. This means that qtbase and its standalone tests need some special handling in regards to when to call that function. Task-number: QTBUG-86866 Change-Id: Ic7ac23de0265561cb06a0ba55089b6c0d3347441 Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-10-02 13:38:47 +00:00
#
Prevent compiler and linker flags from growing every cmake run Functions in QtFlagHandlingHelpers.cmake try to update the CMake cache variables for compiler and linker flags. These were using the current value of those variables and writing the modified ones back to the cache every time CMake ran. If a toolchain file sets or modifies any of these variables, that updated value is used and written back into the cache instead of the original cache variable's value. The next time CMake executes, the toolchain file re-applies the same change and the variable grows longer each time with flags repeated. With Ninja, this causes a complete rebuild every time CMake is re-run. The Android NDK toolchain file is one example where this behavior is triggered (the fault is shared, one could argue that the NDK should only be setting ..._INIT variables, but that's out of our control). Another related bug in the previous implementation was that the flags used to build after the first CMake execution could be different to those used for all builds after the second and later CMake runs. This is because the CMake cache was being updated, but not always the calling scope of the functions that modified them. If a toolchain file set any of the compiler or linker flag variables as non-cache variables, then updating the cache variable would have no effect on the calling scope. The non-cache variable would continue to take precedence for that scope for that run. The next time CMake executes though, the updated cache variable would now have been used by the toolchain file and the change *will* be part of the non-cache variable's value. The above are examples of why you should try to avoid updating these cache variables from project code. We could leave the cache alone and always update only non-cache variables, but then a developer looking at the cache may wonder why the values they see there don't match the values being used in builds. Or worse, they think the cache values are being used and don't realize the builds are using something different. Ultimately, we have to choose which downside we are happy to live with. The changes here preserve the previous intent of updating the cache, but it's still a bit fragile. Fixes: QTBUG-89821 Task-number: QTBUG-85992 Task-number: QTBUG-86866 Change-Id: I8a16753e159bde338e6d2e5dd2ce91fc8ac5c39d Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2021-01-07 05:53:43 +00:00
# match_string - string to match
# replace_string - replacement string
# IN_CACHE - replace them in the corresponding cache variable too. Note that the cache
# variable may have a different value to the non-cache variable.
# CONFIGS - should be a list of upper case configs like DEBUG, RELEASE, RELWITHDEBINFO.
# TYPES - should be a list of target link types as expected by CMake's
# CMAKE_<LINKER_TYPE>_LINKER_FLAGS_<CONFIG> cache variable.
# e.g EXE, MODULE, SHARED, STATIC.
CMake: Refactor optimization flag handling and add optimize_full Introduce a bunch of helper functions to manipulate compiler flags and linker flags for the CMAKE_<LANG>_FLAGS_<CONFIG> and CMAKE_<LINK_TYPE>_LINKER_FLAGS_<CONFIG> CMake variables. These variables can be assigned and modified either in the cache or for a specific subdirectory scope, which will apply the flags only to targets in that scope. Add qt_internal_add_optimize_full_flags() function which mimics qmake's CONFIG += optimize_full behavior. Calling it will force usage of the '-O3' optimization flag on supported platforms (falling back '-O2' where not supported). Use the function for the Core and Gui subdirectories, to enable full optimization for the respective Qt modules as it is done in the qmake projects. To ensure that the global qmake-like compiler flags are assigned eveywhere, qt_internal_set_up_config_optimizations_like_in_qmake() needs to be called after Qt global features like optimize_size and optimize_full are available. This means that qtbase and its standalone tests need some special handling in regards to when to call that function. Task-number: QTBUG-86866 Change-Id: Ic7ac23de0265561cb06a0ba55089b6c0d3347441 Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-10-02 13:38:47 +00:00
function(qt_internal_replace_linker_flags match_string replace_string)
Retire the qt_parse_all_arguments This replaces the qt_parse_all_arguments macro with the built-in `cmake_parse_arguments(PARSE_ARGV`. In addition, a new function, _qt_internal_validate_all_args_are_parsed, can be used to check whether any _UNPARSED_ARGUMENTS have been passed to the function. Fixes: QTBUG-99238 Change-Id: I8cee83dc92dc6acdaaf747ea6ff9084c11dc649b Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
2023-01-11 14:36:18 +00:00
cmake_parse_arguments(PARSE_ARGV 2 arg
Prevent compiler and linker flags from growing every cmake run Functions in QtFlagHandlingHelpers.cmake try to update the CMake cache variables for compiler and linker flags. These were using the current value of those variables and writing the modified ones back to the cache every time CMake ran. If a toolchain file sets or modifies any of these variables, that updated value is used and written back into the cache instead of the original cache variable's value. The next time CMake executes, the toolchain file re-applies the same change and the variable grows longer each time with flags repeated. With Ninja, this causes a complete rebuild every time CMake is re-run. The Android NDK toolchain file is one example where this behavior is triggered (the fault is shared, one could argue that the NDK should only be setting ..._INIT variables, but that's out of our control). Another related bug in the previous implementation was that the flags used to build after the first CMake execution could be different to those used for all builds after the second and later CMake runs. This is because the CMake cache was being updated, but not always the calling scope of the functions that modified them. If a toolchain file set any of the compiler or linker flag variables as non-cache variables, then updating the cache variable would have no effect on the calling scope. The non-cache variable would continue to take precedence for that scope for that run. The next time CMake executes though, the updated cache variable would now have been used by the toolchain file and the change *will* be part of the non-cache variable's value. The above are examples of why you should try to avoid updating these cache variables from project code. We could leave the cache alone and always update only non-cache variables, but then a developer looking at the cache may wonder why the values they see there don't match the values being used in builds. Or worse, they think the cache values are being used and don't realize the builds are using something different. Ultimately, we have to choose which downside we are happy to live with. The changes here preserve the previous intent of updating the cache, but it's still a bit fragile. Fixes: QTBUG-89821 Task-number: QTBUG-85992 Task-number: QTBUG-86866 Change-Id: I8a16753e159bde338e6d2e5dd2ce91fc8ac5c39d Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2021-01-07 05:53:43 +00:00
"IN_CACHE"
CMake: Refactor optimization flag handling and add optimize_full Introduce a bunch of helper functions to manipulate compiler flags and linker flags for the CMAKE_<LANG>_FLAGS_<CONFIG> and CMAKE_<LINK_TYPE>_LINKER_FLAGS_<CONFIG> CMake variables. These variables can be assigned and modified either in the cache or for a specific subdirectory scope, which will apply the flags only to targets in that scope. Add qt_internal_add_optimize_full_flags() function which mimics qmake's CONFIG += optimize_full behavior. Calling it will force usage of the '-O3' optimization flag on supported platforms (falling back '-O2' where not supported). Use the function for the Core and Gui subdirectories, to enable full optimization for the respective Qt modules as it is done in the qmake projects. To ensure that the global qmake-like compiler flags are assigned eveywhere, qt_internal_set_up_config_optimizations_like_in_qmake() needs to be called after Qt global features like optimize_size and optimize_full are available. This means that qtbase and its standalone tests need some special handling in regards to when to call that function. Task-number: QTBUG-86866 Change-Id: Ic7ac23de0265561cb06a0ba55089b6c0d3347441 Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-10-02 13:38:47 +00:00
""
Retire the qt_parse_all_arguments This replaces the qt_parse_all_arguments macro with the built-in `cmake_parse_arguments(PARSE_ARGV`. In addition, a new function, _qt_internal_validate_all_args_are_parsed, can be used to check whether any _UNPARSED_ARGUMENTS have been passed to the function. Fixes: QTBUG-99238 Change-Id: I8cee83dc92dc6acdaaf747ea6ff9084c11dc649b Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
2023-01-11 14:36:18 +00:00
"CONFIGS;TYPES")
_qt_internal_validate_all_args_are_parsed(arg)
CMake: Refactor optimization flag handling and add optimize_full Introduce a bunch of helper functions to manipulate compiler flags and linker flags for the CMAKE_<LANG>_FLAGS_<CONFIG> and CMAKE_<LINK_TYPE>_LINKER_FLAGS_<CONFIG> CMake variables. These variables can be assigned and modified either in the cache or for a specific subdirectory scope, which will apply the flags only to targets in that scope. Add qt_internal_add_optimize_full_flags() function which mimics qmake's CONFIG += optimize_full behavior. Calling it will force usage of the '-O3' optimization flag on supported platforms (falling back '-O2' where not supported). Use the function for the Core and Gui subdirectories, to enable full optimization for the respective Qt modules as it is done in the qmake projects. To ensure that the global qmake-like compiler flags are assigned eveywhere, qt_internal_set_up_config_optimizations_like_in_qmake() needs to be called after Qt global features like optimize_size and optimize_full are available. This means that qtbase and its standalone tests need some special handling in regards to when to call that function. Task-number: QTBUG-86866 Change-Id: Ic7ac23de0265561cb06a0ba55089b6c0d3347441 Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-10-02 13:38:47 +00:00
if(NOT arg_TYPES)
message(FATAL_ERROR
"You must specify at least one linker target type for which to replace the flags.")
endif()
if(NOT arg_CONFIGS)
message(FATAL_ERROR
"You must specify at least one configuration for which to replace the flags.")
endif()
set(configs ${arg_CONFIGS})
set(target_link_types ${arg_TYPES})
foreach(config ${configs})
foreach(t ${target_link_types})
set(flag_var_name "CMAKE_${t}_LINKER_FLAGS_${config}")
Prevent compiler and linker flags from growing every cmake run Functions in QtFlagHandlingHelpers.cmake try to update the CMake cache variables for compiler and linker flags. These were using the current value of those variables and writing the modified ones back to the cache every time CMake ran. If a toolchain file sets or modifies any of these variables, that updated value is used and written back into the cache instead of the original cache variable's value. The next time CMake executes, the toolchain file re-applies the same change and the variable grows longer each time with flags repeated. With Ninja, this causes a complete rebuild every time CMake is re-run. The Android NDK toolchain file is one example where this behavior is triggered (the fault is shared, one could argue that the NDK should only be setting ..._INIT variables, but that's out of our control). Another related bug in the previous implementation was that the flags used to build after the first CMake execution could be different to those used for all builds after the second and later CMake runs. This is because the CMake cache was being updated, but not always the calling scope of the functions that modified them. If a toolchain file set any of the compiler or linker flag variables as non-cache variables, then updating the cache variable would have no effect on the calling scope. The non-cache variable would continue to take precedence for that scope for that run. The next time CMake executes though, the updated cache variable would now have been used by the toolchain file and the change *will* be part of the non-cache variable's value. The above are examples of why you should try to avoid updating these cache variables from project code. We could leave the cache alone and always update only non-cache variables, but then a developer looking at the cache may wonder why the values they see there don't match the values being used in builds. Or worse, they think the cache values are being used and don't realize the builds are using something different. Ultimately, we have to choose which downside we are happy to live with. The changes here preserve the previous intent of updating the cache, but it's still a bit fragile. Fixes: QTBUG-89821 Task-number: QTBUG-85992 Task-number: QTBUG-86866 Change-Id: I8a16753e159bde338e6d2e5dd2ce91fc8ac5c39d Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2021-01-07 05:53:43 +00:00
qt_internal_replace_flags_impl(${flag_var_name}
"${match_string}" "${replace_string}" "${arg_IN_CACHE}")
set(${flag_var_name} "${${flag_var_name}}" PARENT_SCOPE)
CMake: Refactor optimization flag handling and add optimize_full Introduce a bunch of helper functions to manipulate compiler flags and linker flags for the CMAKE_<LANG>_FLAGS_<CONFIG> and CMAKE_<LINK_TYPE>_LINKER_FLAGS_<CONFIG> CMake variables. These variables can be assigned and modified either in the cache or for a specific subdirectory scope, which will apply the flags only to targets in that scope. Add qt_internal_add_optimize_full_flags() function which mimics qmake's CONFIG += optimize_full behavior. Calling it will force usage of the '-O3' optimization flag on supported platforms (falling back '-O2' where not supported). Use the function for the Core and Gui subdirectories, to enable full optimization for the respective Qt modules as it is done in the qmake projects. To ensure that the global qmake-like compiler flags are assigned eveywhere, qt_internal_set_up_config_optimizations_like_in_qmake() needs to be called after Qt global features like optimize_size and optimize_full are available. This means that qtbase and its standalone tests need some special handling in regards to when to call that function. Task-number: QTBUG-86866 Change-Id: Ic7ac23de0265561cb06a0ba55089b6c0d3347441 Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-10-02 13:38:47 +00:00
endforeach()
endforeach()
endfunction()
# This function finds the optimization flags set by the default CMake platform modules or toolchain
# files and replaces them with flags that Qt qmake builds expect, for all the usual
# CMAKE_BUILD_TYPE configurations.
# This normalizes things like using -O2 for both Release and RelWithDebInfo, among other compilation
# flags. Also some linker flags specific to MSVC.
# See QTBUG-85992 for details.
Prevent compiler and linker flags from growing every cmake run Functions in QtFlagHandlingHelpers.cmake try to update the CMake cache variables for compiler and linker flags. These were using the current value of those variables and writing the modified ones back to the cache every time CMake ran. If a toolchain file sets or modifies any of these variables, that updated value is used and written back into the cache instead of the original cache variable's value. The next time CMake executes, the toolchain file re-applies the same change and the variable grows longer each time with flags repeated. With Ninja, this causes a complete rebuild every time CMake is re-run. The Android NDK toolchain file is one example where this behavior is triggered (the fault is shared, one could argue that the NDK should only be setting ..._INIT variables, but that's out of our control). Another related bug in the previous implementation was that the flags used to build after the first CMake execution could be different to those used for all builds after the second and later CMake runs. This is because the CMake cache was being updated, but not always the calling scope of the functions that modified them. If a toolchain file set any of the compiler or linker flag variables as non-cache variables, then updating the cache variable would have no effect on the calling scope. The non-cache variable would continue to take precedence for that scope for that run. The next time CMake executes though, the updated cache variable would now have been used by the toolchain file and the change *will* be part of the non-cache variable's value. The above are examples of why you should try to avoid updating these cache variables from project code. We could leave the cache alone and always update only non-cache variables, but then a developer looking at the cache may wonder why the values they see there don't match the values being used in builds. Or worse, they think the cache values are being used and don't realize the builds are using something different. Ultimately, we have to choose which downside we are happy to live with. The changes here preserve the previous intent of updating the cache, but it's still a bit fragile. Fixes: QTBUG-89821 Task-number: QTBUG-85992 Task-number: QTBUG-86866 Change-Id: I8a16753e159bde338e6d2e5dd2ce91fc8ac5c39d Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2021-01-07 05:53:43 +00:00
#
# Note that both the calling scope and the CMake cache are updated.
CMake: Refactor optimization flag handling and add optimize_full Introduce a bunch of helper functions to manipulate compiler flags and linker flags for the CMAKE_<LANG>_FLAGS_<CONFIG> and CMAKE_<LINK_TYPE>_LINKER_FLAGS_<CONFIG> CMake variables. These variables can be assigned and modified either in the cache or for a specific subdirectory scope, which will apply the flags only to targets in that scope. Add qt_internal_add_optimize_full_flags() function which mimics qmake's CONFIG += optimize_full behavior. Calling it will force usage of the '-O3' optimization flag on supported platforms (falling back '-O2' where not supported). Use the function for the Core and Gui subdirectories, to enable full optimization for the respective Qt modules as it is done in the qmake projects. To ensure that the global qmake-like compiler flags are assigned eveywhere, qt_internal_set_up_config_optimizations_like_in_qmake() needs to be called after Qt global features like optimize_size and optimize_full are available. This means that qtbase and its standalone tests need some special handling in regards to when to call that function. Task-number: QTBUG-86866 Change-Id: Ic7ac23de0265561cb06a0ba55089b6c0d3347441 Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-10-02 13:38:47 +00:00
function(qt_internal_set_up_config_optimizations_like_in_qmake)
# Allow opt out.
if(QT_USE_DEFAULT_CMAKE_OPTIMIZATION_FLAGS)
return()
endif()
qt_internal_get_enabled_languages_for_flag_manipulation(enabled_languages)
qt_internal_get_configs_for_flag_manipulation(configs)
qt_internal_get_target_link_types_for_flag_manipulation(target_link_types)
# You can set QT_DEBUG_OPTIMIZATION_FLAGS to see the before and after results.
if(QT_DEBUG_OPTIMIZATION_FLAGS)
message(STATUS "")
message(STATUS "DEBUG: Original CMake optimization flags.\n")
qt_internal_print_optimization_flags_values_helper("${enabled_languages}" "${configs}"
"${target_link_types}")
endif()
# Remove known optimization flags.
qt_internal_remove_known_optimization_flags(IN_CACHE CONFIGS ${configs})
CMake: Adjust compiler flag optimizations to qmake mkspec ones There are inconsistencies in the default optimization flags added by CMake across configurations like Release and RelWithDebInfo. In particular Release uses -O3, whereas RelWithDebInfo uses -O2, as well as usage of /INCREMENTAL in release configs with MSVC, etc. To make sure that the Qt 6 binaries built with CMake are consistent across configs, as well as consistent with the flags we used when building Qt 5 with qmake, add a horrible search and replace mechanism to replaces the CMake flags with what our mkspecs indicate to use. Ideally this would be done by providing custom CMake toolchain files for each platform we support, and we might revisit that later if the need really arises. To implement the replacing, we first need the flags that should be added. Port the QMAKE_CFLAGS_OPTIMIZE variables to CMake, which is done in QtCompilerOptimization.cmake. Then a new function called qt_internal_set_up_config_optimizations_like_in_qmake will look for any kind of optimization flags set in the CMAKE_<LANG>_FLAGS_<CONFIG> style variables, remove them, and add the appropriate flags that qmake mkspecs provide. On some platforms (like Windows MSVC) the function also alters the linker CMAKE_${TYPE}_LINKER_FLAGS_<CONFIG> style variables. The mechanism allows opting out of this replacing by setting the QT_USE_DEFAULT_CMAKE_OPTIMIZATION_FLAGS value. It also allows opting into removal of flags for custom configs by providing QT_ADDITIONAL_OPTIMIZATION_FLAG_CONFIGS. It's only removal, because we wouldn't know what kind of config it is, and thus what flags to add. The currently modified configs are: Release, RelWithDebInfo, MinSizeRel, Debug aka the usual default CMake provided ones. The mechanism is only applied to C-like languages. ASM is not handled to be on the safe side due to not knowing what kind of compiler flags the platform assembler might take. It's also important to skip RC on MSVC platforms. Task-number: QTBUG-85992 Change-Id: I3712d5cd5a34fceab54f56a6fa46b4e678952362 Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org> Reviewed-by: Cristian Adam <cristian.adam@qt.io>
2020-08-17 15:32:28 +00:00
# Re-add optimization flags as per qmake mkspecs.
foreach(lang ${enabled_languages})
foreach(config ${configs})
set(flag_var_name "CMAKE_${lang}_FLAGS_${config}")
CMake: Refactor optimization flag handling and add optimize_full Introduce a bunch of helper functions to manipulate compiler flags and linker flags for the CMAKE_<LANG>_FLAGS_<CONFIG> and CMAKE_<LINK_TYPE>_LINKER_FLAGS_<CONFIG> CMake variables. These variables can be assigned and modified either in the cache or for a specific subdirectory scope, which will apply the flags only to targets in that scope. Add qt_internal_add_optimize_full_flags() function which mimics qmake's CONFIG += optimize_full behavior. Calling it will force usage of the '-O3' optimization flag on supported platforms (falling back '-O2' where not supported). Use the function for the Core and Gui subdirectories, to enable full optimization for the respective Qt modules as it is done in the qmake projects. To ensure that the global qmake-like compiler flags are assigned eveywhere, qt_internal_set_up_config_optimizations_like_in_qmake() needs to be called after Qt global features like optimize_size and optimize_full are available. This means that qtbase and its standalone tests need some special handling in regards to when to call that function. Task-number: QTBUG-86866 Change-Id: Ic7ac23de0265561cb06a0ba55089b6c0d3347441 Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-10-02 13:38:47 +00:00
set(value_to_append "")
CMake: Adjust compiler flag optimizations to qmake mkspec ones There are inconsistencies in the default optimization flags added by CMake across configurations like Release and RelWithDebInfo. In particular Release uses -O3, whereas RelWithDebInfo uses -O2, as well as usage of /INCREMENTAL in release configs with MSVC, etc. To make sure that the Qt 6 binaries built with CMake are consistent across configs, as well as consistent with the flags we used when building Qt 5 with qmake, add a horrible search and replace mechanism to replaces the CMake flags with what our mkspecs indicate to use. Ideally this would be done by providing custom CMake toolchain files for each platform we support, and we might revisit that later if the need really arises. To implement the replacing, we first need the flags that should be added. Port the QMAKE_CFLAGS_OPTIMIZE variables to CMake, which is done in QtCompilerOptimization.cmake. Then a new function called qt_internal_set_up_config_optimizations_like_in_qmake will look for any kind of optimization flags set in the CMAKE_<LANG>_FLAGS_<CONFIG> style variables, remove them, and add the appropriate flags that qmake mkspecs provide. On some platforms (like Windows MSVC) the function also alters the linker CMAKE_${TYPE}_LINKER_FLAGS_<CONFIG> style variables. The mechanism allows opting out of this replacing by setting the QT_USE_DEFAULT_CMAKE_OPTIMIZATION_FLAGS value. It also allows opting into removal of flags for custom configs by providing QT_ADDITIONAL_OPTIMIZATION_FLAG_CONFIGS. It's only removal, because we wouldn't know what kind of config it is, and thus what flags to add. The currently modified configs are: Release, RelWithDebInfo, MinSizeRel, Debug aka the usual default CMake provided ones. The mechanism is only applied to C-like languages. ASM is not handled to be on the safe side due to not knowing what kind of compiler flags the platform assembler might take. It's also important to skip RC on MSVC platforms. Task-number: QTBUG-85992 Change-Id: I3712d5cd5a34fceab54f56a6fa46b4e678952362 Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org> Reviewed-by: Cristian Adam <cristian.adam@qt.io>
2020-08-17 15:32:28 +00:00
# Release and RelWithDebInfo should get the same base optimization flags.
if(config STREQUAL "RELEASE" AND QT_CFLAGS_OPTIMIZE)
CMake: Refactor optimization flag handling and add optimize_full Introduce a bunch of helper functions to manipulate compiler flags and linker flags for the CMAKE_<LANG>_FLAGS_<CONFIG> and CMAKE_<LINK_TYPE>_LINKER_FLAGS_<CONFIG> CMake variables. These variables can be assigned and modified either in the cache or for a specific subdirectory scope, which will apply the flags only to targets in that scope. Add qt_internal_add_optimize_full_flags() function which mimics qmake's CONFIG += optimize_full behavior. Calling it will force usage of the '-O3' optimization flag on supported platforms (falling back '-O2' where not supported). Use the function for the Core and Gui subdirectories, to enable full optimization for the respective Qt modules as it is done in the qmake projects. To ensure that the global qmake-like compiler flags are assigned eveywhere, qt_internal_set_up_config_optimizations_like_in_qmake() needs to be called after Qt global features like optimize_size and optimize_full are available. This means that qtbase and its standalone tests need some special handling in regards to when to call that function. Task-number: QTBUG-86866 Change-Id: Ic7ac23de0265561cb06a0ba55089b6c0d3347441 Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-10-02 13:38:47 +00:00
set(value_to_append "${QT_CFLAGS_OPTIMIZE}")
CMake: Adjust compiler flag optimizations to qmake mkspec ones There are inconsistencies in the default optimization flags added by CMake across configurations like Release and RelWithDebInfo. In particular Release uses -O3, whereas RelWithDebInfo uses -O2, as well as usage of /INCREMENTAL in release configs with MSVC, etc. To make sure that the Qt 6 binaries built with CMake are consistent across configs, as well as consistent with the flags we used when building Qt 5 with qmake, add a horrible search and replace mechanism to replaces the CMake flags with what our mkspecs indicate to use. Ideally this would be done by providing custom CMake toolchain files for each platform we support, and we might revisit that later if the need really arises. To implement the replacing, we first need the flags that should be added. Port the QMAKE_CFLAGS_OPTIMIZE variables to CMake, which is done in QtCompilerOptimization.cmake. Then a new function called qt_internal_set_up_config_optimizations_like_in_qmake will look for any kind of optimization flags set in the CMAKE_<LANG>_FLAGS_<CONFIG> style variables, remove them, and add the appropriate flags that qmake mkspecs provide. On some platforms (like Windows MSVC) the function also alters the linker CMAKE_${TYPE}_LINKER_FLAGS_<CONFIG> style variables. The mechanism allows opting out of this replacing by setting the QT_USE_DEFAULT_CMAKE_OPTIMIZATION_FLAGS value. It also allows opting into removal of flags for custom configs by providing QT_ADDITIONAL_OPTIMIZATION_FLAG_CONFIGS. It's only removal, because we wouldn't know what kind of config it is, and thus what flags to add. The currently modified configs are: Release, RelWithDebInfo, MinSizeRel, Debug aka the usual default CMake provided ones. The mechanism is only applied to C-like languages. ASM is not handled to be on the safe side due to not knowing what kind of compiler flags the platform assembler might take. It's also important to skip RC on MSVC platforms. Task-number: QTBUG-85992 Change-Id: I3712d5cd5a34fceab54f56a6fa46b4e678952362 Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org> Reviewed-by: Cristian Adam <cristian.adam@qt.io>
2020-08-17 15:32:28 +00:00
elseif(config STREQUAL "RELWITHDEBINFO" AND QT_CFLAGS_OPTIMIZE)
CMake: Refactor optimization flag handling and add optimize_full Introduce a bunch of helper functions to manipulate compiler flags and linker flags for the CMAKE_<LANG>_FLAGS_<CONFIG> and CMAKE_<LINK_TYPE>_LINKER_FLAGS_<CONFIG> CMake variables. These variables can be assigned and modified either in the cache or for a specific subdirectory scope, which will apply the flags only to targets in that scope. Add qt_internal_add_optimize_full_flags() function which mimics qmake's CONFIG += optimize_full behavior. Calling it will force usage of the '-O3' optimization flag on supported platforms (falling back '-O2' where not supported). Use the function for the Core and Gui subdirectories, to enable full optimization for the respective Qt modules as it is done in the qmake projects. To ensure that the global qmake-like compiler flags are assigned eveywhere, qt_internal_set_up_config_optimizations_like_in_qmake() needs to be called after Qt global features like optimize_size and optimize_full are available. This means that qtbase and its standalone tests need some special handling in regards to when to call that function. Task-number: QTBUG-86866 Change-Id: Ic7ac23de0265561cb06a0ba55089b6c0d3347441 Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-10-02 13:38:47 +00:00
set(value_to_append "${QT_CFLAGS_OPTIMIZE}")
CMake: Adjust compiler flag optimizations to qmake mkspec ones There are inconsistencies in the default optimization flags added by CMake across configurations like Release and RelWithDebInfo. In particular Release uses -O3, whereas RelWithDebInfo uses -O2, as well as usage of /INCREMENTAL in release configs with MSVC, etc. To make sure that the Qt 6 binaries built with CMake are consistent across configs, as well as consistent with the flags we used when building Qt 5 with qmake, add a horrible search and replace mechanism to replaces the CMake flags with what our mkspecs indicate to use. Ideally this would be done by providing custom CMake toolchain files for each platform we support, and we might revisit that later if the need really arises. To implement the replacing, we first need the flags that should be added. Port the QMAKE_CFLAGS_OPTIMIZE variables to CMake, which is done in QtCompilerOptimization.cmake. Then a new function called qt_internal_set_up_config_optimizations_like_in_qmake will look for any kind of optimization flags set in the CMAKE_<LANG>_FLAGS_<CONFIG> style variables, remove them, and add the appropriate flags that qmake mkspecs provide. On some platforms (like Windows MSVC) the function also alters the linker CMAKE_${TYPE}_LINKER_FLAGS_<CONFIG> style variables. The mechanism allows opting out of this replacing by setting the QT_USE_DEFAULT_CMAKE_OPTIMIZATION_FLAGS value. It also allows opting into removal of flags for custom configs by providing QT_ADDITIONAL_OPTIMIZATION_FLAG_CONFIGS. It's only removal, because we wouldn't know what kind of config it is, and thus what flags to add. The currently modified configs are: Release, RelWithDebInfo, MinSizeRel, Debug aka the usual default CMake provided ones. The mechanism is only applied to C-like languages. ASM is not handled to be on the safe side due to not knowing what kind of compiler flags the platform assembler might take. It's also important to skip RC on MSVC platforms. Task-number: QTBUG-85992 Change-Id: I3712d5cd5a34fceab54f56a6fa46b4e678952362 Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org> Reviewed-by: Cristian Adam <cristian.adam@qt.io>
2020-08-17 15:32:28 +00:00
# MinSizeRel should get the optimize size flag if available, otherwise the regular
# release flag.
elseif(config STREQUAL "MINSIZEREL")
if(QT_CFLAGS_OPTIMIZE_SIZE)
CMake: Refactor optimization flag handling and add optimize_full Introduce a bunch of helper functions to manipulate compiler flags and linker flags for the CMAKE_<LANG>_FLAGS_<CONFIG> and CMAKE_<LINK_TYPE>_LINKER_FLAGS_<CONFIG> CMake variables. These variables can be assigned and modified either in the cache or for a specific subdirectory scope, which will apply the flags only to targets in that scope. Add qt_internal_add_optimize_full_flags() function which mimics qmake's CONFIG += optimize_full behavior. Calling it will force usage of the '-O3' optimization flag on supported platforms (falling back '-O2' where not supported). Use the function for the Core and Gui subdirectories, to enable full optimization for the respective Qt modules as it is done in the qmake projects. To ensure that the global qmake-like compiler flags are assigned eveywhere, qt_internal_set_up_config_optimizations_like_in_qmake() needs to be called after Qt global features like optimize_size and optimize_full are available. This means that qtbase and its standalone tests need some special handling in regards to when to call that function. Task-number: QTBUG-86866 Change-Id: Ic7ac23de0265561cb06a0ba55089b6c0d3347441 Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-10-02 13:38:47 +00:00
set(value_to_append "${QT_CFLAGS_OPTIMIZE_SIZE}")
CMake: Adjust compiler flag optimizations to qmake mkspec ones There are inconsistencies in the default optimization flags added by CMake across configurations like Release and RelWithDebInfo. In particular Release uses -O3, whereas RelWithDebInfo uses -O2, as well as usage of /INCREMENTAL in release configs with MSVC, etc. To make sure that the Qt 6 binaries built with CMake are consistent across configs, as well as consistent with the flags we used when building Qt 5 with qmake, add a horrible search and replace mechanism to replaces the CMake flags with what our mkspecs indicate to use. Ideally this would be done by providing custom CMake toolchain files for each platform we support, and we might revisit that later if the need really arises. To implement the replacing, we first need the flags that should be added. Port the QMAKE_CFLAGS_OPTIMIZE variables to CMake, which is done in QtCompilerOptimization.cmake. Then a new function called qt_internal_set_up_config_optimizations_like_in_qmake will look for any kind of optimization flags set in the CMAKE_<LANG>_FLAGS_<CONFIG> style variables, remove them, and add the appropriate flags that qmake mkspecs provide. On some platforms (like Windows MSVC) the function also alters the linker CMAKE_${TYPE}_LINKER_FLAGS_<CONFIG> style variables. The mechanism allows opting out of this replacing by setting the QT_USE_DEFAULT_CMAKE_OPTIMIZATION_FLAGS value. It also allows opting into removal of flags for custom configs by providing QT_ADDITIONAL_OPTIMIZATION_FLAG_CONFIGS. It's only removal, because we wouldn't know what kind of config it is, and thus what flags to add. The currently modified configs are: Release, RelWithDebInfo, MinSizeRel, Debug aka the usual default CMake provided ones. The mechanism is only applied to C-like languages. ASM is not handled to be on the safe side due to not knowing what kind of compiler flags the platform assembler might take. It's also important to skip RC on MSVC platforms. Task-number: QTBUG-85992 Change-Id: I3712d5cd5a34fceab54f56a6fa46b4e678952362 Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org> Reviewed-by: Cristian Adam <cristian.adam@qt.io>
2020-08-17 15:32:28 +00:00
else()
CMake: Refactor optimization flag handling and add optimize_full Introduce a bunch of helper functions to manipulate compiler flags and linker flags for the CMAKE_<LANG>_FLAGS_<CONFIG> and CMAKE_<LINK_TYPE>_LINKER_FLAGS_<CONFIG> CMake variables. These variables can be assigned and modified either in the cache or for a specific subdirectory scope, which will apply the flags only to targets in that scope. Add qt_internal_add_optimize_full_flags() function which mimics qmake's CONFIG += optimize_full behavior. Calling it will force usage of the '-O3' optimization flag on supported platforms (falling back '-O2' where not supported). Use the function for the Core and Gui subdirectories, to enable full optimization for the respective Qt modules as it is done in the qmake projects. To ensure that the global qmake-like compiler flags are assigned eveywhere, qt_internal_set_up_config_optimizations_like_in_qmake() needs to be called after Qt global features like optimize_size and optimize_full are available. This means that qtbase and its standalone tests need some special handling in regards to when to call that function. Task-number: QTBUG-86866 Change-Id: Ic7ac23de0265561cb06a0ba55089b6c0d3347441 Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-10-02 13:38:47 +00:00
set(value_to_append "${QT_CFLAGS_OPTIMIZE}")
CMake: Adjust compiler flag optimizations to qmake mkspec ones There are inconsistencies in the default optimization flags added by CMake across configurations like Release and RelWithDebInfo. In particular Release uses -O3, whereas RelWithDebInfo uses -O2, as well as usage of /INCREMENTAL in release configs with MSVC, etc. To make sure that the Qt 6 binaries built with CMake are consistent across configs, as well as consistent with the flags we used when building Qt 5 with qmake, add a horrible search and replace mechanism to replaces the CMake flags with what our mkspecs indicate to use. Ideally this would be done by providing custom CMake toolchain files for each platform we support, and we might revisit that later if the need really arises. To implement the replacing, we first need the flags that should be added. Port the QMAKE_CFLAGS_OPTIMIZE variables to CMake, which is done in QtCompilerOptimization.cmake. Then a new function called qt_internal_set_up_config_optimizations_like_in_qmake will look for any kind of optimization flags set in the CMAKE_<LANG>_FLAGS_<CONFIG> style variables, remove them, and add the appropriate flags that qmake mkspecs provide. On some platforms (like Windows MSVC) the function also alters the linker CMAKE_${TYPE}_LINKER_FLAGS_<CONFIG> style variables. The mechanism allows opting out of this replacing by setting the QT_USE_DEFAULT_CMAKE_OPTIMIZATION_FLAGS value. It also allows opting into removal of flags for custom configs by providing QT_ADDITIONAL_OPTIMIZATION_FLAG_CONFIGS. It's only removal, because we wouldn't know what kind of config it is, and thus what flags to add. The currently modified configs are: Release, RelWithDebInfo, MinSizeRel, Debug aka the usual default CMake provided ones. The mechanism is only applied to C-like languages. ASM is not handled to be on the safe side due to not knowing what kind of compiler flags the platform assembler might take. It's also important to skip RC on MSVC platforms. Task-number: QTBUG-85992 Change-Id: I3712d5cd5a34fceab54f56a6fa46b4e678952362 Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org> Reviewed-by: Cristian Adam <cristian.adam@qt.io>
2020-08-17 15:32:28 +00:00
endif()
endif()
CMake: Implement configure -optimize-debug Task-number: QTBUG-85373 Change-Id: I3aba1fcf96e36544dd75a5e6adc5f770a9a45726 Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
2020-08-20 08:19:32 +00:00
# Debug should get the OPTIMIZE_DEBUG flag if the respective feature is ON.
if(config STREQUAL "DEBUG" AND QT_FEATURE_optimize_debug)
CMake: Refactor optimization flag handling and add optimize_full Introduce a bunch of helper functions to manipulate compiler flags and linker flags for the CMAKE_<LANG>_FLAGS_<CONFIG> and CMAKE_<LINK_TYPE>_LINKER_FLAGS_<CONFIG> CMake variables. These variables can be assigned and modified either in the cache or for a specific subdirectory scope, which will apply the flags only to targets in that scope. Add qt_internal_add_optimize_full_flags() function which mimics qmake's CONFIG += optimize_full behavior. Calling it will force usage of the '-O3' optimization flag on supported platforms (falling back '-O2' where not supported). Use the function for the Core and Gui subdirectories, to enable full optimization for the respective Qt modules as it is done in the qmake projects. To ensure that the global qmake-like compiler flags are assigned eveywhere, qt_internal_set_up_config_optimizations_like_in_qmake() needs to be called after Qt global features like optimize_size and optimize_full are available. This means that qtbase and its standalone tests need some special handling in regards to when to call that function. Task-number: QTBUG-86866 Change-Id: Ic7ac23de0265561cb06a0ba55089b6c0d3347441 Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-10-02 13:38:47 +00:00
set(value_to_append "${QT_CFLAGS_OPTIMIZE_DEBUG}")
CMake: Implement configure -optimize-debug Task-number: QTBUG-85373 Change-Id: I3aba1fcf96e36544dd75a5e6adc5f770a9a45726 Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
2020-08-20 08:19:32 +00:00
endif()
CMake: Implement configure -optimize-size This configure switch controls the feature 'optimize_size'. This isn't merely a mapping to CMAKE_BUILD_TYPE=MinSizeRel, because we potentially want to combine -optimize-size with -force-debug-info, which maps to CMAKE_BUILD_CONFIG=RelWithDebInfo. Task-number: QTBUG-85373 Change-Id: I1a9343ebc54816f52e90e9d33ea3df4c99b1ec9f Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
2020-08-21 06:40:54 +00:00
set(configs_for_optimize_size RELEASE RELWITHDEBINFO)
if(QT_FEATURE_optimize_size AND config IN_LIST configs_for_optimize_size)
CMake: Refactor optimization flag handling and add optimize_full Introduce a bunch of helper functions to manipulate compiler flags and linker flags for the CMAKE_<LANG>_FLAGS_<CONFIG> and CMAKE_<LINK_TYPE>_LINKER_FLAGS_<CONFIG> CMake variables. These variables can be assigned and modified either in the cache or for a specific subdirectory scope, which will apply the flags only to targets in that scope. Add qt_internal_add_optimize_full_flags() function which mimics qmake's CONFIG += optimize_full behavior. Calling it will force usage of the '-O3' optimization flag on supported platforms (falling back '-O2' where not supported). Use the function for the Core and Gui subdirectories, to enable full optimization for the respective Qt modules as it is done in the qmake projects. To ensure that the global qmake-like compiler flags are assigned eveywhere, qt_internal_set_up_config_optimizations_like_in_qmake() needs to be called after Qt global features like optimize_size and optimize_full are available. This means that qtbase and its standalone tests need some special handling in regards to when to call that function. Task-number: QTBUG-86866 Change-Id: Ic7ac23de0265561cb06a0ba55089b6c0d3347441 Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-10-02 13:38:47 +00:00
set(value_to_append "${QT_CFLAGS_OPTIMIZE_SIZE}")
CMake: Implement configure -optimize-size This configure switch controls the feature 'optimize_size'. This isn't merely a mapping to CMAKE_BUILD_TYPE=MinSizeRel, because we potentially want to combine -optimize-size with -force-debug-info, which maps to CMAKE_BUILD_CONFIG=RelWithDebInfo. Task-number: QTBUG-85373 Change-Id: I1a9343ebc54816f52e90e9d33ea3df4c99b1ec9f Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
2020-08-21 06:40:54 +00:00
endif()
CMake: Introduce CMake-only optimize_full feature Enabling it will force usage of '-O3' flag when building Qt. If the platform has no '-O3' flag, use '-O2' as a fallback. Task-number: QTBUG-86866 Change-Id: If13f7de954ba5c01dc9634f06a85529828fe90a9 Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-10-05 07:09:13 +00:00
# Check if the fake 'optimize_full' feature is enabled.
# Use the max optimization level flag for all release configs, effectively
# overriding any previous setting.
set(configs_for_optimize RELEASE RELWITHDEBINFO MINSIZEREL)
if(QT_FEATURE_optimize_full AND config IN_LIST configs_for_optimize)
qt_internal_get_optimize_full_flags(optimize_full_flags)
set(value_to_append "${optimize_full_flags}")
endif()
CMake: Adjust compiler flag optimizations to qmake mkspec ones There are inconsistencies in the default optimization flags added by CMake across configurations like Release and RelWithDebInfo. In particular Release uses -O3, whereas RelWithDebInfo uses -O2, as well as usage of /INCREMENTAL in release configs with MSVC, etc. To make sure that the Qt 6 binaries built with CMake are consistent across configs, as well as consistent with the flags we used when building Qt 5 with qmake, add a horrible search and replace mechanism to replaces the CMake flags with what our mkspecs indicate to use. Ideally this would be done by providing custom CMake toolchain files for each platform we support, and we might revisit that later if the need really arises. To implement the replacing, we first need the flags that should be added. Port the QMAKE_CFLAGS_OPTIMIZE variables to CMake, which is done in QtCompilerOptimization.cmake. Then a new function called qt_internal_set_up_config_optimizations_like_in_qmake will look for any kind of optimization flags set in the CMAKE_<LANG>_FLAGS_<CONFIG> style variables, remove them, and add the appropriate flags that qmake mkspecs provide. On some platforms (like Windows MSVC) the function also alters the linker CMAKE_${TYPE}_LINKER_FLAGS_<CONFIG> style variables. The mechanism allows opting out of this replacing by setting the QT_USE_DEFAULT_CMAKE_OPTIMIZATION_FLAGS value. It also allows opting into removal of flags for custom configs by providing QT_ADDITIONAL_OPTIMIZATION_FLAG_CONFIGS. It's only removal, because we wouldn't know what kind of config it is, and thus what flags to add. The currently modified configs are: Release, RelWithDebInfo, MinSizeRel, Debug aka the usual default CMake provided ones. The mechanism is only applied to C-like languages. ASM is not handled to be on the safe side due to not knowing what kind of compiler flags the platform assembler might take. It's also important to skip RC on MSVC platforms. Task-number: QTBUG-85992 Change-Id: I3712d5cd5a34fceab54f56a6fa46b4e678952362 Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org> Reviewed-by: Cristian Adam <cristian.adam@qt.io>
2020-08-17 15:32:28 +00:00
# Assign value to the cache entry.
CMake: Refactor optimization flag handling and add optimize_full Introduce a bunch of helper functions to manipulate compiler flags and linker flags for the CMAKE_<LANG>_FLAGS_<CONFIG> and CMAKE_<LINK_TYPE>_LINKER_FLAGS_<CONFIG> CMake variables. These variables can be assigned and modified either in the cache or for a specific subdirectory scope, which will apply the flags only to targets in that scope. Add qt_internal_add_optimize_full_flags() function which mimics qmake's CONFIG += optimize_full behavior. Calling it will force usage of the '-O3' optimization flag on supported platforms (falling back '-O2' where not supported). Use the function for the Core and Gui subdirectories, to enable full optimization for the respective Qt modules as it is done in the qmake projects. To ensure that the global qmake-like compiler flags are assigned eveywhere, qt_internal_set_up_config_optimizations_like_in_qmake() needs to be called after Qt global features like optimize_size and optimize_full are available. This means that qtbase and its standalone tests need some special handling in regards to when to call that function. Task-number: QTBUG-86866 Change-Id: Ic7ac23de0265561cb06a0ba55089b6c0d3347441 Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-10-02 13:38:47 +00:00
if(value_to_append)
Prevent compiler and linker flags from growing every cmake run Functions in QtFlagHandlingHelpers.cmake try to update the CMake cache variables for compiler and linker flags. These were using the current value of those variables and writing the modified ones back to the cache every time CMake ran. If a toolchain file sets or modifies any of these variables, that updated value is used and written back into the cache instead of the original cache variable's value. The next time CMake executes, the toolchain file re-applies the same change and the variable grows longer each time with flags repeated. With Ninja, this causes a complete rebuild every time CMake is re-run. The Android NDK toolchain file is one example where this behavior is triggered (the fault is shared, one could argue that the NDK should only be setting ..._INIT variables, but that's out of our control). Another related bug in the previous implementation was that the flags used to build after the first CMake execution could be different to those used for all builds after the second and later CMake runs. This is because the CMake cache was being updated, but not always the calling scope of the functions that modified them. If a toolchain file set any of the compiler or linker flag variables as non-cache variables, then updating the cache variable would have no effect on the calling scope. The non-cache variable would continue to take precedence for that scope for that run. The next time CMake executes though, the updated cache variable would now have been used by the toolchain file and the change *will* be part of the non-cache variable's value. The above are examples of why you should try to avoid updating these cache variables from project code. We could leave the cache alone and always update only non-cache variables, but then a developer looking at the cache may wonder why the values they see there don't match the values being used in builds. Or worse, they think the cache values are being used and don't realize the builds are using something different. Ultimately, we have to choose which downside we are happy to live with. The changes here preserve the previous intent of updating the cache, but it's still a bit fragile. Fixes: QTBUG-89821 Task-number: QTBUG-85992 Task-number: QTBUG-86866 Change-Id: I8a16753e159bde338e6d2e5dd2ce91fc8ac5c39d Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2021-01-07 05:53:43 +00:00
qt_internal_add_flags_impl(${flag_var_name} "${value_to_append}" TRUE)
# Delay updating the calling scope's variables to the end of this function
CMake: Refactor optimization flag handling and add optimize_full Introduce a bunch of helper functions to manipulate compiler flags and linker flags for the CMAKE_<LANG>_FLAGS_<CONFIG> and CMAKE_<LINK_TYPE>_LINKER_FLAGS_<CONFIG> CMake variables. These variables can be assigned and modified either in the cache or for a specific subdirectory scope, which will apply the flags only to targets in that scope. Add qt_internal_add_optimize_full_flags() function which mimics qmake's CONFIG += optimize_full behavior. Calling it will force usage of the '-O3' optimization flag on supported platforms (falling back '-O2' where not supported). Use the function for the Core and Gui subdirectories, to enable full optimization for the respective Qt modules as it is done in the qmake projects. To ensure that the global qmake-like compiler flags are assigned eveywhere, qt_internal_set_up_config_optimizations_like_in_qmake() needs to be called after Qt global features like optimize_size and optimize_full are available. This means that qtbase and its standalone tests need some special handling in regards to when to call that function. Task-number: QTBUG-86866 Change-Id: Ic7ac23de0265561cb06a0ba55089b6c0d3347441 Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-10-02 13:38:47 +00:00
endif()
CMake: Adjust compiler flag optimizations to qmake mkspec ones There are inconsistencies in the default optimization flags added by CMake across configurations like Release and RelWithDebInfo. In particular Release uses -O3, whereas RelWithDebInfo uses -O2, as well as usage of /INCREMENTAL in release configs with MSVC, etc. To make sure that the Qt 6 binaries built with CMake are consistent across configs, as well as consistent with the flags we used when building Qt 5 with qmake, add a horrible search and replace mechanism to replaces the CMake flags with what our mkspecs indicate to use. Ideally this would be done by providing custom CMake toolchain files for each platform we support, and we might revisit that later if the need really arises. To implement the replacing, we first need the flags that should be added. Port the QMAKE_CFLAGS_OPTIMIZE variables to CMake, which is done in QtCompilerOptimization.cmake. Then a new function called qt_internal_set_up_config_optimizations_like_in_qmake will look for any kind of optimization flags set in the CMAKE_<LANG>_FLAGS_<CONFIG> style variables, remove them, and add the appropriate flags that qmake mkspecs provide. On some platforms (like Windows MSVC) the function also alters the linker CMAKE_${TYPE}_LINKER_FLAGS_<CONFIG> style variables. The mechanism allows opting out of this replacing by setting the QT_USE_DEFAULT_CMAKE_OPTIMIZATION_FLAGS value. It also allows opting into removal of flags for custom configs by providing QT_ADDITIONAL_OPTIMIZATION_FLAG_CONFIGS. It's only removal, because we wouldn't know what kind of config it is, and thus what flags to add. The currently modified configs are: Release, RelWithDebInfo, MinSizeRel, Debug aka the usual default CMake provided ones. The mechanism is only applied to C-like languages. ASM is not handled to be on the safe side due to not knowing what kind of compiler flags the platform assembler might take. It's also important to skip RC on MSVC platforms. Task-number: QTBUG-85992 Change-Id: I3712d5cd5a34fceab54f56a6fa46b4e678952362 Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org> Reviewed-by: Cristian Adam <cristian.adam@qt.io>
2020-08-17 15:32:28 +00:00
endforeach()
endforeach()
if(MSVC)
# Handle MSVC /INCREMENTAL flag which should not be enabled for Release configurations.
# First remove them from all configs, and re-add INCREMENTAL for Debug only.
set(flag_values "/INCREMENTAL:YES" "/INCREMENTAL:NO" "/INCREMENTAL")
CMake: Refactor optimization flag handling and add optimize_full Introduce a bunch of helper functions to manipulate compiler flags and linker flags for the CMAKE_<LANG>_FLAGS_<CONFIG> and CMAKE_<LINK_TYPE>_LINKER_FLAGS_<CONFIG> CMake variables. These variables can be assigned and modified either in the cache or for a specific subdirectory scope, which will apply the flags only to targets in that scope. Add qt_internal_add_optimize_full_flags() function which mimics qmake's CONFIG += optimize_full behavior. Calling it will force usage of the '-O3' optimization flag on supported platforms (falling back '-O2' where not supported). Use the function for the Core and Gui subdirectories, to enable full optimization for the respective Qt modules as it is done in the qmake projects. To ensure that the global qmake-like compiler flags are assigned eveywhere, qt_internal_set_up_config_optimizations_like_in_qmake() needs to be called after Qt global features like optimize_size and optimize_full are available. This means that qtbase and its standalone tests need some special handling in regards to when to call that function. Task-number: QTBUG-86866 Change-Id: Ic7ac23de0265561cb06a0ba55089b6c0d3347441 Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-10-02 13:38:47 +00:00
foreach(flag_value ${flag_values})
qt_internal_replace_linker_flags(
"${flag_value}" ""
CONFIGS ${configs}
TYPES ${target_link_types}
IN_CACHE)
CMake: Adjust compiler flag optimizations to qmake mkspec ones There are inconsistencies in the default optimization flags added by CMake across configurations like Release and RelWithDebInfo. In particular Release uses -O3, whereas RelWithDebInfo uses -O2, as well as usage of /INCREMENTAL in release configs with MSVC, etc. To make sure that the Qt 6 binaries built with CMake are consistent across configs, as well as consistent with the flags we used when building Qt 5 with qmake, add a horrible search and replace mechanism to replaces the CMake flags with what our mkspecs indicate to use. Ideally this would be done by providing custom CMake toolchain files for each platform we support, and we might revisit that later if the need really arises. To implement the replacing, we first need the flags that should be added. Port the QMAKE_CFLAGS_OPTIMIZE variables to CMake, which is done in QtCompilerOptimization.cmake. Then a new function called qt_internal_set_up_config_optimizations_like_in_qmake will look for any kind of optimization flags set in the CMAKE_<LANG>_FLAGS_<CONFIG> style variables, remove them, and add the appropriate flags that qmake mkspecs provide. On some platforms (like Windows MSVC) the function also alters the linker CMAKE_${TYPE}_LINKER_FLAGS_<CONFIG> style variables. The mechanism allows opting out of this replacing by setting the QT_USE_DEFAULT_CMAKE_OPTIMIZATION_FLAGS value. It also allows opting into removal of flags for custom configs by providing QT_ADDITIONAL_OPTIMIZATION_FLAG_CONFIGS. It's only removal, because we wouldn't know what kind of config it is, and thus what flags to add. The currently modified configs are: Release, RelWithDebInfo, MinSizeRel, Debug aka the usual default CMake provided ones. The mechanism is only applied to C-like languages. ASM is not handled to be on the safe side due to not knowing what kind of compiler flags the platform assembler might take. It's also important to skip RC on MSVC platforms. Task-number: QTBUG-85992 Change-Id: I3712d5cd5a34fceab54f56a6fa46b4e678952362 Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org> Reviewed-by: Cristian Adam <cristian.adam@qt.io>
2020-08-17 15:32:28 +00:00
endforeach()
CMake: Refactor optimization flag handling and add optimize_full Introduce a bunch of helper functions to manipulate compiler flags and linker flags for the CMAKE_<LANG>_FLAGS_<CONFIG> and CMAKE_<LINK_TYPE>_LINKER_FLAGS_<CONFIG> CMake variables. These variables can be assigned and modified either in the cache or for a specific subdirectory scope, which will apply the flags only to targets in that scope. Add qt_internal_add_optimize_full_flags() function which mimics qmake's CONFIG += optimize_full behavior. Calling it will force usage of the '-O3' optimization flag on supported platforms (falling back '-O2' where not supported). Use the function for the Core and Gui subdirectories, to enable full optimization for the respective Qt modules as it is done in the qmake projects. To ensure that the global qmake-like compiler flags are assigned eveywhere, qt_internal_set_up_config_optimizations_like_in_qmake() needs to be called after Qt global features like optimize_size and optimize_full are available. This means that qtbase and its standalone tests need some special handling in regards to when to call that function. Task-number: QTBUG-86866 Change-Id: Ic7ac23de0265561cb06a0ba55089b6c0d3347441 Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-10-02 13:38:47 +00:00
set(flag_value "/INCREMENTAL:NO")
qt_internal_add_linker_flags(
FLAGS "${flag_value}"
CONFIGS RELEASE RELWITHDEBINFO MINSIZEREL
Ensure consistent flags for ASM language and MODULE targets By not including ASM in the languages for which compiler flags are adjusted, we can end up with inconsistent optimization and debug settings between different languages. Make sure they are treated the same. Similarly, linker flags for MODULE target types should also have their debugging and incremental linking flags adjusted, just like EXE and SHARED targets. When building with MSVC, MODULE targets were having the relevant flags stripped, but not then replaced with the desired ones, leaving them empty in typical scenarios. This would primarily affect plugins, which are built as MODULE rather than SHARED libraries. Fixes: QTBUG-90237 Change-Id: I648ea74be1654d24cbecc592ce0ca4b59b2ae839 Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io> Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
2021-01-14 05:47:19 +00:00
TYPES EXE SHARED MODULE # when linking static libraries, link.exe can't recognize this parameter, clang-cl will error out.
CMake: Refactor optimization flag handling and add optimize_full Introduce a bunch of helper functions to manipulate compiler flags and linker flags for the CMAKE_<LANG>_FLAGS_<CONFIG> and CMAKE_<LINK_TYPE>_LINKER_FLAGS_<CONFIG> CMake variables. These variables can be assigned and modified either in the cache or for a specific subdirectory scope, which will apply the flags only to targets in that scope. Add qt_internal_add_optimize_full_flags() function which mimics qmake's CONFIG += optimize_full behavior. Calling it will force usage of the '-O3' optimization flag on supported platforms (falling back '-O2' where not supported). Use the function for the Core and Gui subdirectories, to enable full optimization for the respective Qt modules as it is done in the qmake projects. To ensure that the global qmake-like compiler flags are assigned eveywhere, qt_internal_set_up_config_optimizations_like_in_qmake() needs to be called after Qt global features like optimize_size and optimize_full are available. This means that qtbase and its standalone tests need some special handling in regards to when to call that function. Task-number: QTBUG-86866 Change-Id: Ic7ac23de0265561cb06a0ba55089b6c0d3347441 Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-10-02 13:38:47 +00:00
IN_CACHE)
Clear non-relevant CMake compiler flags Add internal function to cleanup compiler flags out of the CMAKE_xxx_FLAGS_xxx variables. Use introduced interface to clear the '/EHsc' flag for the MSVC compiler family. This adjusts the CMake behavior to the qmake one. Change the 'EXCEPTIONS' option handling in helper functions. Add ability to add enabling and disabling exception flags. Previously it was only possible to add disabling exception flags. Fixes: QTBUG-89952 Change-Id: I60d47660a97ae9b5a1d1f4107d352c9e97890144 Reviewed-by: Craig Scott <craig.scott@qt.io> Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2021-01-21 10:17:58 +00:00
qt_internal_remove_compiler_flags("(^| )/EH[scra-]*( |$)" LANGUAGES CXX CONFIGS ${configs} IN_CACHE REGEX)
CMake: Adjust compiler flag optimizations to qmake mkspec ones There are inconsistencies in the default optimization flags added by CMake across configurations like Release and RelWithDebInfo. In particular Release uses -O3, whereas RelWithDebInfo uses -O2, as well as usage of /INCREMENTAL in release configs with MSVC, etc. To make sure that the Qt 6 binaries built with CMake are consistent across configs, as well as consistent with the flags we used when building Qt 5 with qmake, add a horrible search and replace mechanism to replaces the CMake flags with what our mkspecs indicate to use. Ideally this would be done by providing custom CMake toolchain files for each platform we support, and we might revisit that later if the need really arises. To implement the replacing, we first need the flags that should be added. Port the QMAKE_CFLAGS_OPTIMIZE variables to CMake, which is done in QtCompilerOptimization.cmake. Then a new function called qt_internal_set_up_config_optimizations_like_in_qmake will look for any kind of optimization flags set in the CMAKE_<LANG>_FLAGS_<CONFIG> style variables, remove them, and add the appropriate flags that qmake mkspecs provide. On some platforms (like Windows MSVC) the function also alters the linker CMAKE_${TYPE}_LINKER_FLAGS_<CONFIG> style variables. The mechanism allows opting out of this replacing by setting the QT_USE_DEFAULT_CMAKE_OPTIMIZATION_FLAGS value. It also allows opting into removal of flags for custom configs by providing QT_ADDITIONAL_OPTIMIZATION_FLAG_CONFIGS. It's only removal, because we wouldn't know what kind of config it is, and thus what flags to add. The currently modified configs are: Release, RelWithDebInfo, MinSizeRel, Debug aka the usual default CMake provided ones. The mechanism is only applied to C-like languages. ASM is not handled to be on the safe side due to not knowing what kind of compiler flags the platform assembler might take. It's also important to skip RC on MSVC platforms. Task-number: QTBUG-85992 Change-Id: I3712d5cd5a34fceab54f56a6fa46b4e678952362 Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org> Reviewed-by: Cristian Adam <cristian.adam@qt.io>
2020-08-17 15:32:28 +00:00
endif()
CMake: Allow embedding debug info in object files with MSVC Introduce a new feature called 'FEATURE_msvc_obj_debug_info' which should allow usage of sccache for Windows MSVC configs. Enabling the feature will replace the default '/Zi' compile flag that CMake sets in Windows-MSVC.cmake with '/Z7'. This ensures that the debug info is placed into the compiled object file instead of placing it in the compilation .pdb file via mspdbsrv.exe. The final pdb file will still be created for shared libraries and executables due to the linker '/debug' flag. Static libraries will not have a .pdb file, instead the debug info being embeeded in the .lib file. The downsides of using '/Z7' are more disk space usage, slower link times, more memory used for linking because the linker needs to deal with more symbols. The upside is that caching each indepedent cl.exe call is possible. This is what Chromium uses with goma, and Firefox with sccache. We're not enabling the feature by default to allow easier testing (and switching it off) in the CI if it proves not to be feasible. It will probably have to be disabled at least for building qtwebengine due to high memory requirements. Task-number: QTQAINFRA-3934 Change-Id: I89e0e57995e938a780be318d1faec700194aa93f Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org> Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-10-02 16:24:32 +00:00
# Allow opting into generating debug info in object files with a fake feature.
# This would allow us to enable caching with sccache.
# See QTQAINFRA-3934 for details.
if(MSVC AND QT_FEATURE_msvc_obj_debug_info)
qt_internal_replace_compiler_flags(
"/Zi" "/Z7"
CONFIGS RELWITHDEBINFO DEBUG
TYPES ${target_link_types}
IN_CACHE)
endif()
Remove debug symbols from Android's release build Android's toolchain file, ie., android-legacy.toolchain.cmake assumes that the default build is a Debug build, and it adds the `-g` flag to CMAKE_<LANG>_FLAGS, as a result, our release Android build always contains debug symbols. In this patch, I basically move the `-g` flag from CMAKE_<LANG>_FLAGS to CMAKE_<LANG>_FLAGS_DEBUG, and CMAKE_<LANG>_FLAGS_RELWITHDEBINFO. Pick-to: 6.5 6.6 Fixes: QTBUG-111901 Change-Id: I31eadb07d9172c923e8beaf0ac6c6e34fe1ebefb Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2023-05-05 14:22:42 +00:00
# Legacy Android toolchain file adds the `-g` flag to CMAKE_<LANG>_FLAGS, as a
# result, our release build ends up containing debug symbols. To avoid that, we
# remove the flag from CMAKE_<LANGL>_FLAGS and add
# it to CMAKE_<LANG>_FLAGS_DEBUG.
#
# Note:
# The new `android.toolchain.cmake` file does not have this problem, but
# it has other issues, eg., https://github.com/android/ndk/issues/1693, so we
# cannot force it. While we do load the new toolchain, it automatically falls
# back to the legacy toolchain, ie., `android-legacy.toolchain.cmake` which
# has the problem described above.
#
# Todo:
# When the new toolchain is fixed, and it doesn't fall back to the legacy
# anymore by default, then we should be able to remove this workaround.
if(ANDROID AND ANDROID_COMPILER_FLAGS MATCHES "(^| )-g")
qt_internal_remove_compiler_flags("-g")
qt_internal_add_compiler_flags(FLAGS "-g" CONFIGS DEBUG RELWITHDEBINFO)
endif()
Prevent compiler and linker flags from growing every cmake run Functions in QtFlagHandlingHelpers.cmake try to update the CMake cache variables for compiler and linker flags. These were using the current value of those variables and writing the modified ones back to the cache every time CMake ran. If a toolchain file sets or modifies any of these variables, that updated value is used and written back into the cache instead of the original cache variable's value. The next time CMake executes, the toolchain file re-applies the same change and the variable grows longer each time with flags repeated. With Ninja, this causes a complete rebuild every time CMake is re-run. The Android NDK toolchain file is one example where this behavior is triggered (the fault is shared, one could argue that the NDK should only be setting ..._INIT variables, but that's out of our control). Another related bug in the previous implementation was that the flags used to build after the first CMake execution could be different to those used for all builds after the second and later CMake runs. This is because the CMake cache was being updated, but not always the calling scope of the functions that modified them. If a toolchain file set any of the compiler or linker flag variables as non-cache variables, then updating the cache variable would have no effect on the calling scope. The non-cache variable would continue to take precedence for that scope for that run. The next time CMake executes though, the updated cache variable would now have been used by the toolchain file and the change *will* be part of the non-cache variable's value. The above are examples of why you should try to avoid updating these cache variables from project code. We could leave the cache alone and always update only non-cache variables, but then a developer looking at the cache may wonder why the values they see there don't match the values being used in builds. Or worse, they think the cache values are being used and don't realize the builds are using something different. Ultimately, we have to choose which downside we are happy to live with. The changes here preserve the previous intent of updating the cache, but it's still a bit fragile. Fixes: QTBUG-89821 Task-number: QTBUG-85992 Task-number: QTBUG-86866 Change-Id: I8a16753e159bde338e6d2e5dd2ce91fc8ac5c39d Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2021-01-07 05:53:43 +00:00
# Update all relevant flags in the calling scope
Set CMAKE_<LANG>_FLAGS in PARENT_SCOPE as well Previously, we were not setting the CMAKE_<LANG>_FLAGS, instead we were only processing the CMAKE_<LANG>_FLAGS_<CONFIG>. This patch amends that so we also update the default flags list as well. Pick-to: 6.5 Change-Id: I7ceb87850a806fe8ad4eac6f8147b312eba01bc6 Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
2023-06-01 09:50:32 +00:00
foreach(lang ${enabled_languages})
set(flag_var_name "CMAKE_${lang}_FLAGS")
set(${flag_var_name} "${${flag_var_name}}" PARENT_SCOPE)
foreach(config ${configs})
Prevent compiler and linker flags from growing every cmake run Functions in QtFlagHandlingHelpers.cmake try to update the CMake cache variables for compiler and linker flags. These were using the current value of those variables and writing the modified ones back to the cache every time CMake ran. If a toolchain file sets or modifies any of these variables, that updated value is used and written back into the cache instead of the original cache variable's value. The next time CMake executes, the toolchain file re-applies the same change and the variable grows longer each time with flags repeated. With Ninja, this causes a complete rebuild every time CMake is re-run. The Android NDK toolchain file is one example where this behavior is triggered (the fault is shared, one could argue that the NDK should only be setting ..._INIT variables, but that's out of our control). Another related bug in the previous implementation was that the flags used to build after the first CMake execution could be different to those used for all builds after the second and later CMake runs. This is because the CMake cache was being updated, but not always the calling scope of the functions that modified them. If a toolchain file set any of the compiler or linker flag variables as non-cache variables, then updating the cache variable would have no effect on the calling scope. The non-cache variable would continue to take precedence for that scope for that run. The next time CMake executes though, the updated cache variable would now have been used by the toolchain file and the change *will* be part of the non-cache variable's value. The above are examples of why you should try to avoid updating these cache variables from project code. We could leave the cache alone and always update only non-cache variables, but then a developer looking at the cache may wonder why the values they see there don't match the values being used in builds. Or worse, they think the cache values are being used and don't realize the builds are using something different. Ultimately, we have to choose which downside we are happy to live with. The changes here preserve the previous intent of updating the cache, but it's still a bit fragile. Fixes: QTBUG-89821 Task-number: QTBUG-85992 Task-number: QTBUG-86866 Change-Id: I8a16753e159bde338e6d2e5dd2ce91fc8ac5c39d Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2021-01-07 05:53:43 +00:00
set(flag_var_name "CMAKE_${lang}_FLAGS_${config}")
set(${flag_var_name} "${${flag_var_name}}" PARENT_SCOPE)
endforeach()
Set CMAKE_<LANG>_FLAGS in PARENT_SCOPE as well Previously, we were not setting the CMAKE_<LANG>_FLAGS, instead we were only processing the CMAKE_<LANG>_FLAGS_<CONFIG>. This patch amends that so we also update the default flags list as well. Pick-to: 6.5 Change-Id: I7ceb87850a806fe8ad4eac6f8147b312eba01bc6 Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
2023-06-01 09:50:32 +00:00
endforeach()
Prevent compiler and linker flags from growing every cmake run Functions in QtFlagHandlingHelpers.cmake try to update the CMake cache variables for compiler and linker flags. These were using the current value of those variables and writing the modified ones back to the cache every time CMake ran. If a toolchain file sets or modifies any of these variables, that updated value is used and written back into the cache instead of the original cache variable's value. The next time CMake executes, the toolchain file re-applies the same change and the variable grows longer each time with flags repeated. With Ninja, this causes a complete rebuild every time CMake is re-run. The Android NDK toolchain file is one example where this behavior is triggered (the fault is shared, one could argue that the NDK should only be setting ..._INIT variables, but that's out of our control). Another related bug in the previous implementation was that the flags used to build after the first CMake execution could be different to those used for all builds after the second and later CMake runs. This is because the CMake cache was being updated, but not always the calling scope of the functions that modified them. If a toolchain file set any of the compiler or linker flag variables as non-cache variables, then updating the cache variable would have no effect on the calling scope. The non-cache variable would continue to take precedence for that scope for that run. The next time CMake executes though, the updated cache variable would now have been used by the toolchain file and the change *will* be part of the non-cache variable's value. The above are examples of why you should try to avoid updating these cache variables from project code. We could leave the cache alone and always update only non-cache variables, but then a developer looking at the cache may wonder why the values they see there don't match the values being used in builds. Or worse, they think the cache values are being used and don't realize the builds are using something different. Ultimately, we have to choose which downside we are happy to live with. The changes here preserve the previous intent of updating the cache, but it's still a bit fragile. Fixes: QTBUG-89821 Task-number: QTBUG-85992 Task-number: QTBUG-86866 Change-Id: I8a16753e159bde338e6d2e5dd2ce91fc8ac5c39d Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2021-01-07 05:53:43 +00:00
Set CMAKE_<LANG>_FLAGS in PARENT_SCOPE as well Previously, we were not setting the CMAKE_<LANG>_FLAGS, instead we were only processing the CMAKE_<LANG>_FLAGS_<CONFIG>. This patch amends that so we also update the default flags list as well. Pick-to: 6.5 Change-Id: I7ceb87850a806fe8ad4eac6f8147b312eba01bc6 Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
2023-06-01 09:50:32 +00:00
foreach(t ${target_link_types})
set(flag_var_name "CMAKE_${t}_LINKER_FLAGS")
set(${flag_var_name} "${${flag_var_name}}" PARENT_SCOPE)
foreach(config ${configs})
Prevent compiler and linker flags from growing every cmake run Functions in QtFlagHandlingHelpers.cmake try to update the CMake cache variables for compiler and linker flags. These were using the current value of those variables and writing the modified ones back to the cache every time CMake ran. If a toolchain file sets or modifies any of these variables, that updated value is used and written back into the cache instead of the original cache variable's value. The next time CMake executes, the toolchain file re-applies the same change and the variable grows longer each time with flags repeated. With Ninja, this causes a complete rebuild every time CMake is re-run. The Android NDK toolchain file is one example where this behavior is triggered (the fault is shared, one could argue that the NDK should only be setting ..._INIT variables, but that's out of our control). Another related bug in the previous implementation was that the flags used to build after the first CMake execution could be different to those used for all builds after the second and later CMake runs. This is because the CMake cache was being updated, but not always the calling scope of the functions that modified them. If a toolchain file set any of the compiler or linker flag variables as non-cache variables, then updating the cache variable would have no effect on the calling scope. The non-cache variable would continue to take precedence for that scope for that run. The next time CMake executes though, the updated cache variable would now have been used by the toolchain file and the change *will* be part of the non-cache variable's value. The above are examples of why you should try to avoid updating these cache variables from project code. We could leave the cache alone and always update only non-cache variables, but then a developer looking at the cache may wonder why the values they see there don't match the values being used in builds. Or worse, they think the cache values are being used and don't realize the builds are using something different. Ultimately, we have to choose which downside we are happy to live with. The changes here preserve the previous intent of updating the cache, but it's still a bit fragile. Fixes: QTBUG-89821 Task-number: QTBUG-85992 Task-number: QTBUG-86866 Change-Id: I8a16753e159bde338e6d2e5dd2ce91fc8ac5c39d Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2021-01-07 05:53:43 +00:00
set(flag_var_name "CMAKE_${t}_LINKER_FLAGS_${config}")
set(${flag_var_name} "${${flag_var_name}}" PARENT_SCOPE)
endforeach()
endforeach()
CMake: Adjust compiler flag optimizations to qmake mkspec ones There are inconsistencies in the default optimization flags added by CMake across configurations like Release and RelWithDebInfo. In particular Release uses -O3, whereas RelWithDebInfo uses -O2, as well as usage of /INCREMENTAL in release configs with MSVC, etc. To make sure that the Qt 6 binaries built with CMake are consistent across configs, as well as consistent with the flags we used when building Qt 5 with qmake, add a horrible search and replace mechanism to replaces the CMake flags with what our mkspecs indicate to use. Ideally this would be done by providing custom CMake toolchain files for each platform we support, and we might revisit that later if the need really arises. To implement the replacing, we first need the flags that should be added. Port the QMAKE_CFLAGS_OPTIMIZE variables to CMake, which is done in QtCompilerOptimization.cmake. Then a new function called qt_internal_set_up_config_optimizations_like_in_qmake will look for any kind of optimization flags set in the CMAKE_<LANG>_FLAGS_<CONFIG> style variables, remove them, and add the appropriate flags that qmake mkspecs provide. On some platforms (like Windows MSVC) the function also alters the linker CMAKE_${TYPE}_LINKER_FLAGS_<CONFIG> style variables. The mechanism allows opting out of this replacing by setting the QT_USE_DEFAULT_CMAKE_OPTIMIZATION_FLAGS value. It also allows opting into removal of flags for custom configs by providing QT_ADDITIONAL_OPTIMIZATION_FLAG_CONFIGS. It's only removal, because we wouldn't know what kind of config it is, and thus what flags to add. The currently modified configs are: Release, RelWithDebInfo, MinSizeRel, Debug aka the usual default CMake provided ones. The mechanism is only applied to C-like languages. ASM is not handled to be on the safe side due to not knowing what kind of compiler flags the platform assembler might take. It's also important to skip RC on MSVC platforms. Task-number: QTBUG-85992 Change-Id: I3712d5cd5a34fceab54f56a6fa46b4e678952362 Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org> Reviewed-by: Cristian Adam <cristian.adam@qt.io>
2020-08-17 15:32:28 +00:00
if(QT_DEBUG_OPTIMIZATION_FLAGS)
message(STATUS "")
message(STATUS "DEBUG: Modified optimization flags to mirror qmake mkspecs.\n")
CMake: Refactor optimization flag handling and add optimize_full Introduce a bunch of helper functions to manipulate compiler flags and linker flags for the CMAKE_<LANG>_FLAGS_<CONFIG> and CMAKE_<LINK_TYPE>_LINKER_FLAGS_<CONFIG> CMake variables. These variables can be assigned and modified either in the cache or for a specific subdirectory scope, which will apply the flags only to targets in that scope. Add qt_internal_add_optimize_full_flags() function which mimics qmake's CONFIG += optimize_full behavior. Calling it will force usage of the '-O3' optimization flag on supported platforms (falling back '-O2' where not supported). Use the function for the Core and Gui subdirectories, to enable full optimization for the respective Qt modules as it is done in the qmake projects. To ensure that the global qmake-like compiler flags are assigned eveywhere, qt_internal_set_up_config_optimizations_like_in_qmake() needs to be called after Qt global features like optimize_size and optimize_full are available. This means that qtbase and its standalone tests need some special handling in regards to when to call that function. Task-number: QTBUG-86866 Change-Id: Ic7ac23de0265561cb06a0ba55089b6c0d3347441 Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
2020-10-02 13:38:47 +00:00
qt_internal_print_optimization_flags_values_helper("${enabled_languages}" "${configs}"
"${target_link_types}")
CMake: Adjust compiler flag optimizations to qmake mkspec ones There are inconsistencies in the default optimization flags added by CMake across configurations like Release and RelWithDebInfo. In particular Release uses -O3, whereas RelWithDebInfo uses -O2, as well as usage of /INCREMENTAL in release configs with MSVC, etc. To make sure that the Qt 6 binaries built with CMake are consistent across configs, as well as consistent with the flags we used when building Qt 5 with qmake, add a horrible search and replace mechanism to replaces the CMake flags with what our mkspecs indicate to use. Ideally this would be done by providing custom CMake toolchain files for each platform we support, and we might revisit that later if the need really arises. To implement the replacing, we first need the flags that should be added. Port the QMAKE_CFLAGS_OPTIMIZE variables to CMake, which is done in QtCompilerOptimization.cmake. Then a new function called qt_internal_set_up_config_optimizations_like_in_qmake will look for any kind of optimization flags set in the CMAKE_<LANG>_FLAGS_<CONFIG> style variables, remove them, and add the appropriate flags that qmake mkspecs provide. On some platforms (like Windows MSVC) the function also alters the linker CMAKE_${TYPE}_LINKER_FLAGS_<CONFIG> style variables. The mechanism allows opting out of this replacing by setting the QT_USE_DEFAULT_CMAKE_OPTIMIZATION_FLAGS value. It also allows opting into removal of flags for custom configs by providing QT_ADDITIONAL_OPTIMIZATION_FLAG_CONFIGS. It's only removal, because we wouldn't know what kind of config it is, and thus what flags to add. The currently modified configs are: Release, RelWithDebInfo, MinSizeRel, Debug aka the usual default CMake provided ones. The mechanism is only applied to C-like languages. ASM is not handled to be on the safe side due to not knowing what kind of compiler flags the platform assembler might take. It's also important to skip RC on MSVC platforms. Task-number: QTBUG-85992 Change-Id: I3712d5cd5a34fceab54f56a6fa46b4e678952362 Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org> Reviewed-by: Cristian Adam <cristian.adam@qt.io>
2020-08-17 15:32:28 +00:00
endif()
endfunction()
Reference in New Issue Copy Permalink