We have now had several requests for inlining previously-exported
member functions, but no standard mechanism to effect it.
As QT_REMOVED_SINCE has shown, there is great value in having such a
standard mechanism, so here is one.
With this change, to inline a previously exported (member) function,
simply
- mark the declaration with QT_<MODULE>_INLINE_SINCE
- move the definition into the header file (outside the class),
- wrap it in QT_<MODULE>_INLINE_IMPL_SINCE
- #include the header into the module's removed_api.cpp
Just including the header into removed_api.cpp is enough, so you may
want to add a comment:
#include "header.h" // uses QT_<MODULE>_INLINE_SINCE
The effect is as follows:
- A TU in a _different_ library will see an inline declaration,
followed by the definition, and so it will see a normal inline
function.
- A TU in the same library will, however, see a non-inline
declaration, to avoid the ODR violation that at least GCC/ld are
able to detect.
- When QT_<MODULE>_BUILD_REMOVED_API is in effect, the TU will also
see the definition, which is the same setup as before the change,
except in a different TU, and therefore export the member.
- When, OTOH, QT_<MODULE>_BUILD_REMOVED_API is _not_ in effect, the
TU will see no declaration, assuming (correctly), that the
definition will be supplied by a different TU.
This is, of course, an ODR violation, but not worse than what we do
elsewhere, as the definitions differ only between library and user.
The function is inline only for the users of the library, not the
library itself, which will still see the function as non-inline. If
inlining is critical within the library, too, the existing function
should call a new inline function, and calls in the same library should
be changed to call the new inline function instead.
Use the new mechanism to inline the QLocale ctor we intended to inline
for 6.3, but couldn't, because we hadn't found the magic incantation,
yet. Thiago found it a few weeks later, and this is what this patch is
based on.
Fixes: QTBUG-100452
Pick-to: 6.4
Change-Id: Ia0030cddc64b6b92edfed860170d5204aa74b953
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Changing a C++ source file can trigger rebuilds of a lot of other
source files that might include AUTOGEN'ed headers or sources.
See https://gitlab.kitware.com/cmake/cmake/-/issues/22531 for some
details. Fixed in CMake 3.21.2.
There are still files that are rebuilt even in 3.21.2, but it's less,
and it returns to the status quo of how it was in 3.21.0 or earlier
versions.
Pick-to: 6.2 6.3 6.4
Task-number: QTBUG-104352
Change-Id: Ie1c991d52df48442d4134e4ed22a8137a3c993c8
Reviewed-by: Alexey Edelev <alexey.edelev@qt.io>
Reviewed-by: Jörg Bornemann <joerg.bornemann@qt.io>
The test run is wrapped with a special TESTRUNNER script that ignores
failing tests (there are several tests failing when built with ASAN) and
also ignores LSAN errors (memory leaks - but still visible in the
output).
The test run only fails if a test reports ASAN errors or if it
crashes (or times out, which is like a crash caused by qtestlib's
watchdog timer).
Fixes: QTQAINFRA-5025
Change-Id: I861756ab49388ac4a52409d3a780684244e469b1
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
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>
This will hide the configuration summary and cmake feature summary and
found packages output upon reconfiguration.
Pick-to: 6.2 6.3 6.4
Task-number: QTBUG-104128
Change-Id: I42270b99e45076052ec176df4652661cae10ac0c
Reviewed-by: Jörg Bornemann <joerg.bornemann@qt.io>
Reviewed-by: Kai Koehne <kai.koehne@qt.io>
or when feature changes are detected, even when the log-level is set
to NOTICE (which is the default for non-developer-builds).
We want to show the summary during the first configuration so we don't
force users to look into the config.summary file.
We want not to show the summary upon reconfigurations, to keep regular
reconfigurations as quiet as possibe, so it's easy to notice any new
warnings.
Amends e2a0ddbb69
Amends 384dfceb53
Pick-to: 6.2 6.3 6.4
Fixes: QTBUG-104127
Change-Id: I506f33b4bae9da8957e04bb69c206bf00e3f7b0e
Reviewed-by: Jörg Bornemann <joerg.bornemann@qt.io>
Reviewed-by: Kai Koehne <kai.koehne@qt.io>
FindWrapSystemHarfbuzz.cmake relied on pkg-config to find system
harfbuzz. This patch makes it find system harfbuzz even if pkg-config
is not available or disabled.
Pick-to: 6.2 6.3 6.4
Task-number: QTBUG-103894
Change-Id: I2a8fc64c738c7604f47de89f387002e40a9fa5e0
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
We need to check whether pkg_check_modules returns success before
setting up target name etc.
Pick-to: 6.2 6.3 6.4
Fixes: QTBUG-103894
Change-Id: I12702639683723d976e93be95443099b88885869
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
Turning off pkg-config with the configure flag -no-pkg-config did not
work. There are different defaults for FEATURE_pkg_config on different
platforms (e.g. Linux: ON, Windows: OFF). The existing code that
calculated the initial FEATURE_pkg_config value assumed that the default
is OFF and never turned the feature off.
Pick-to: 6.2 6.3 6.4
Fixes: QTBUG-104123
Change-Id: I33b9687c55c60d4ec9224324951a8838741ee976
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
As of Emscripten 3.1.3, the --embind compiler argment has been
depreciated for -lembind
Change-Id: Iac5bc21602f27fda7c1ea6814a1c9525b9a5afab
Reviewed-by: Morten Johan Sørvig <morten.sorvig@qt.io>
Adding specialHTMLTargets to EXPORTED_RUNTIME_METHODS carries the
obligation to actually use it as well; failing to do so makes Emscripten
stop with a reference error on startup.
However, we can't guarantee that Qt will use it in all cases. The
current usage depends on QGuiApplication being used. Application code
could be using QCoreApplication, or no application object at all.
Detect if specialHTMLTargets is present instead, and then enable the code
paths which uses it if that's the case. This means that apps which want
to use e.g. multiple browser windows can opt into support by making sure
EXPORTED_RUNTIME_METHODS contains specialHTMLTargets.
Change-Id: I81105aa01946602fcf593f170e305d7dc9bad3be
Reviewed-by: Lorn Potter <lorn.potter@gmail.com>
Added a small helper function that also installs the script to
prefix builds. Additionally, leveraging the cmake option
CROSSCOMPILING_EMULATOR for the testrunner fits neatly in the rest
of our cmake code.
Change-Id: I75288e97c81b250ac3997f2e7a22bc7bd82b7b69
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
This ensures that the Xcode 'Display name' input under
${target} -> General -> Identity -> Display name is not empty.
Because adding ${PRODUCT_NAME} directly in the Info.plist.in template
will cause CMake to evaluate it as variable expansion, work around the
issue by putting the dollar sign into a separate cache variable that
after evaluation will result in ${PRODUCT_NAME} being in the file
verbatim, so that Xcode evaluate it at build time.
Amends 4d838dae5a
Pick-to: 6.2 6.3
Task-number: QTBUG-95838
Change-Id: I2d1090cc8e84b32442f7daca2d4ce5e3ad413c68
Reviewed-by: Jörg Bornemann <joerg.bornemann@qt.io>
Reviewed-by: Alexey Edelev <alexey.edelev@qt.io>
Currently our iOS CMake toolchain file sets the global
XCODE_EMIT_EFFECTIVE_PLATFORM_NAME property to OFF, to work
around a CMake issue regarding usage of object libraries in
conjunction with Xcode. The error was
The OBJECT library type may not be used for IMPORTED libraries under
Xcode with multiple architectures $(CURRENT_ARCH)
While this got rid of the error, it introduced a regression where
linking an executable against a static library in the same project
failed due to the library not being placed in a directory where Xcode
expects it to be.
clang: error: no such file or directory:
'~/build-untitled4-Qt_6_0_2_for_iOS/Debug/libuntitled4.a'
The actual library is created in Debug-iphoneos/libuntitled4.a
Note the -iphoneos suffix, which is related to the
EFFECTIVE_PLATFORM_NAME.
This could be further worked around by either explicitly setting the
library ARCHIVE_OUTPUT_DIRECTORY property in the project, or flipping
the value of the XCODE_EMIT_EFFECTIVE_PLATFORM_NAME to ON at the
end of the project. Both workarounds are not project-friendly.
In the mean time CMake got a fix for the original error
at https://gitlab.kitware.com/cmake/cmake/-/merge_requests/5771
which got released with CMake 3.20.0.
That was unfortunately not sufficient to remove our initial
workaround, because removing it would trigger a different
error at generation time
CMake Error:
Error evaluating generator expression:
$<TARGET_OBJECTS:Qt6::Quick_resources_1>
The evaluation of the TARGET_OBJECTS generator expression is only
suitable for consumption by CMake (limited under Xcode with multiple
architectures). It is not suitable for writing out elsewhere.
Fortunately, another fix landed in CMake 3.21.0 to address that
https://gitlab.kitware.com/cmake/cmake/-/merge_requests/6161
Because static builds (iOS) require CMake 3.21, with both of these
fixes we can now remove our workaround.
Even projects that set ARCHIVE_OUTPUT_DIRECTORY or flip
XCODE_EMIT_EFFECTIVE_PLATFORM_NAME to ON still continue to work
fine.
This also fixes installation of libraries, which also took into
account the effective platform name and thus caused a mismatch between
the expected output dir and the real one.
This reverts 1e1805ed36
Pick-to: 6.2 6.3
Fixes: QTBUG-93268
Fixes: QTBUG-95381
Task-number: QTBUG-87198
Change-Id: Ifcaf0f89e4328ae9859c596882ce32401fa491c3
Reviewed-by: Jörg Bornemann <joerg.bornemann@qt.io>
Reviewed-by: Alexey Edelev <alexey.edelev@qt.io>
When calling qt-cmake on the command line, we don't usually force
usage of a particular CMake generator and defer to the user's choice
or CMake's default for the host OS.
In the case of iOS, the generator that makes the most sense to use is
Xcode. One could also use Ninja / Unix Makefiles if the project
is only building static libraries, but for shared libraries and
executables the project likely needs the code signing
provided by xcodebuild.
When targeting iOS, use a different qt-cmake file template.
The iOS-specific shell script will set the CMAKE_GENERATOR environment
variable to 'Xcode'.
If no -G or -D CMAKE_GENERATOR is provided on the command line then
the project will use the Xcode generator. Otherwise the generator
given on the command line takes precedence.
The CMAKE_GENERATOR environment variable from the parent process is
completely ignored.
The logic is only done for iOS, to reduce the likeliness
of breaking the qt-cmake script for other platforms.
Note that Qt Creator does not use qt-cmake, but rather calls cmake
directly with additional options like the toolchain file,
architecture, sysroot, etc.
[ChangeLog][iOS][CMake] qt-cmake now defaults to using the Xcode
generator when targeting iOS projects.
Pick-to: 6.2 6.3
Fixes: QTBUG-100834
Change-Id: I39b3dce47cc9ee2f98678631e4bd035c59c65294
Reviewed-by: Jörg Bornemann <joerg.bornemann@qt.io>
Output names of static libraries might be different from target names.
For example, the library name of Qt6::DeviceDiscoverySupportPrivate is
"Qt6DeviceDiscoverySupport.lib", and the library name of
Qt6::QTlsBackendCertOnlyPlugin is "qcertonlybackend.lib".
This commit make pdb files names consistent with the library names.
And make sure we have set correct OUTPUT_NAME property before calling
qt_set_common_target_properties()/qt_internal_set_compile_pdb_names().
Change-Id: Idb3cacd7a46a4f298fd584b927b5d726956faea8
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Sometimes it is not desirable to include the libraries in the APK,
e.g. system and vendor apps could prefer having one set of libraries
installed on the device. If unbundled deployment is specified,
native libraries will not be included in the APK.
With unbundled deployment, optional arguments can be passed to
set the path to load the libraries on the device.
[ChangeLog][Android][Deployment Changes] Adds option for Unbundled
deployment, where native libraries are not packaged in the APK.
Task-number: QAA-771
Change-Id: Ica51ef83a24dad58c7586bf610a58abe21fc1100
Reviewed-by: Assam Boudjelthia <assam.boudjelthia@qt.io>
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>
If QT_GENERATE_WRAPPER_SCRIPTS_FOR_ALL_HOSTS is ON then we generate
Windows scripts on Unix and vice versa. We always used the host
platforms line endings for generating the scripts. This leads to
Windows line endings in Unix scripts and vice versa.
Explicitly specify the line endings style when generating wrapper
scripts.
Fixes: QTBUG-102747
Pick-to: 6.2 6.3
Change-Id: I1603add46f276a5d91bbf0f103a261cdd84c343b
Reviewed-by: Alexey Edelev <alexey.edelev@qt.io>
Replace the current license disclaimer in files by
a SPDX-License-Identifier.
Files that have to be modified by hand are modified.
License files are organized under LICENSES directory.
Task-number: QTBUG-67283
Change-Id: Id880c92784c40f3bbde861c0d93f58151c18b9f1
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Reviewed-by: Jörg Bornemann <joerg.bornemann@qt.io>
qmake adds QMAKE_PRL_LIBS values on the link line after the currently
processed library. Because CMake adds resource object files into
QMAKE_PRL_LIBS, they end up on the link line after the library.
This causes issues on Linux with GNU ld and ld.gold, because the
linker discards symbols from the library which are then later referenced
by the object files.
Work around that by placing the path to the library directly into
QMAKE_PRL_LIBS after the resource object files.
This ensures the linker doesn't discard the symbols required by the
resource object files.
This means that each library encountered in qmake's LIBS variable will
be temporarily referenced twice in qmake's project state: once from
LIBS / QMAKE_PRL_TARGET, and once when the QMAKE_PRL_LIBS values are
added. On the link line it will appear only once though, because qmake
does library deduplication during prl file processing, which only keeps
the last occurrence.
Amends 4ab5432081
Pick-to: 6.2 6.3
Fixes: QTBUG-103002
Change-Id: I97647b64de22b158424850915fee62b5fea5f995
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Alexey Edelev <alexey.edelev@qt.io>
JavaScripts's BigInt feature provides support for arbitrary-precision
integers. This makes it possible to represent 64-bit integers; the
standard JS Number type supports 32-bit integers only (or more
accurately 53-bit integers - see Number.MAX_SAFE_INTEGER).
Enable WASM_BIGINT which makes Emscripten map int64_t and uint64_t
to BigInt when interfacing with JavaScript code.
This removes one of the conditions which enables
wasm-emscripten-finalize.
Task-number: QTBUG-103352
Change-Id: Ia70d599daaf34c92695f5a2b61665e0c237e6b95
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Reviewed-by: Lorn Potter <lorn.potter@gmail.com>
Reviewed-by: David Skoland <david.skoland@qt.io>
This removes one of the conditions which enables
wasm-emscripten-finalize.
Task-number: QTBUG-103352
Change-Id: Id05db4b081dec360cdad2e611622e5baf09aeb23
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Reviewed-by: David Skoland <david.skoland@qt.io>
find_library does not always work because libatomic.so may be in a path
like /usr/lib/gcc/x86_64-linux-gnu/11/libatomic.so, which CMake does not
consider by default.
Pick-to: 6.3
Change-Id: I73a657c470efa4f84f8629bd531edfcac3b3a352
Reviewed-by: Alexey Edelev <alexey.edelev@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Jörg Bornemann <joerg.bornemann@qt.io>
Unsetting CMAKE_STRIP and including CMakeFindBinUtils to find it again
is not safe, because CMakeFindBinUtils has logic to search for
additional tool names depending on the currently processed language.
The currently processed language is set in _CMAKE_PROCESSING_LANGUAGE
only when CMake is doing it's language introspection via
CMakeCXXCompiler.cmake.
This resulted in the build system finding a regular host-OS strip,
rather than an android specific llvm-strip when doing an Android
build, which then silently failed to strip the Android libraries
and caused us to ship big binaries.
Improve the strip wrapper creation in a few ways:
- Save the original strip value on first configuration
- Restore it if needed, without using CMakeFindBinUtils
- Don't apply the strip wrapper creation logic to Android,
we currently don't need it there
- Add some informational messages about which CMAKE_STRIP
ends up being used even if log-level is not DEBUG
- Fix a typo
Amends 39f657032b
Pick-to: 6.2 6.3
Fixes: QTBUG-103356
Task-number: QTBUG-101653
Change-Id: I23d98180446d5bb7628e783668f46e4e546ed700
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Alexey Edelev <alexey.edelev@qt.io>
...when QT_BUILD_TOOLS_WHEN_CROSSCOMPILING is ON.
When QT_BUILD_TOOLS_WHEN_CROSSCOMPILING is ON, we want to set
QT_FORCE_BUILD_TOOLS. But this happened too late: after the
initialization of QT_BUILD_TOOLS_BY_DEFAULT. This value depends on
QT_FORCE_BUILD_TOOLS.
This amends commit acfbe3b779.
Change-Id: Ibdba54da943aea1b55618f10d2b8485f4390878a
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
Replace BOOTSTRAP option with the single value CORE_LIBRARY argument
in qt_internal_add_tool and qt_internal_add_executable functions.
The introduced argument now may accept 'Bootstap' and 'None' values.
Use 'Bootstap' to link Qt::Boostrap library instead Qt::Core or 'None'
to avoid any core library linking. This is useful for tools that need
to use the CMake deployment routines, but not require the Qt::Core
functionality.
Task-number: QTBUG-87480
Change-Id: I64a8b17f16ac5fe43c6b385252dc21def0c88d2c
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
Previously, this was only supported when cross-compiling, but that's an
unnecessary limitation. Instead, make it possible to build the "host"
tools (notably qmake) even when they've been found elsewhere due to
QT_FORCE_FIND_TOOLS=ON and a supplied QT_HOST_PATH.
The combination of QT_FORCE_FIND_TOOLS and QT_FORCE_BUILD_TOOLS set to
ON is useful for developers who touch content that ends up in the
bootstrap library.
QT_BUILD_TOOLS_WHEN_CROSSCOMPILING is deprecated in favor of
QT_FORCE_BUILD_TOOLS.
[ChangeLog][CMake] QT_BUILD_TOOLS_WHEN_CROSSCOMPILING has been
deprecated in favor of QT_FORCE_BUILD_TOOLS. The latter can be used in
combination with QT_FORCE_FIND_TOOLS and QT_HOST_PATH to use tools from
a host Qt even for a non-cross build and still build the tools.
Fixes: QTBUG-99683
Change-Id: I0e5f6bec596a4a78bd3bfffd16c8fb486181f9b6
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Otherwise they will just use default compiler
Pick-to: 6.3
Change-Id: Id5813b99fbbb6b0d8b0ee658e06312b637a097c1
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
If Qt is configured with -platform linux-g++-32 , make sure
to add the -m32 compile options for all built targets.
On 64 bit host OSes that provide both 32 and 64 bit libraries
we need to exclude the 64 bit libraries from being picked up.
The locations of the libraries are distro-specific.
This change by default excludes the Ubuntu x86_64 libraries
paths.
Opt outs are provided, which when used, forces Qt builders to
specify their own ignore paths in a custom CMake toolchain file.
The compile option and default path exclusions are added to the
Qt-generated CMake toolchain file as well, so they are reused
when building other Qt repositories.
Note that there is no foolproof way to tell CMake to ignore all
x86_64 packages / libraries, even if CMake 3.23
CMAKE_IGNORE_PREFIX_PATH is used, because there might not be
a single sysroot to exclude.
Both x86 and x86_64 libraries can co-exist in the same sysroot,
e.g in /usr
One would have to list each package / library directory in
CMAKE_IGNORE_PATH manually.
Additionally, the PKG_CONFIG_LIBDIR environment variable is also set
to Ubuntu specific prefixes, to ensure that pkg_check_modules ->
pkg-config don't pickup x86_64 libraries.
Fixes: QTBUG-101963
Change-Id: Ib17c8d2cd0ba33b2cf748772245bcd558de9120c
Reviewed-by: Jörg Bornemann <joerg.bornemann@qt.io>
Do build zlib as static 3rdparty library. This makes it easier to
disable warnings.
Pick-to: 6.3
Change-Id: I1db331b671b64e68d81c56b0df337983c3bbe7fa
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
message(STATUS) prints output to a buffered stdout, whereas
message(NOTICE) or just message() print to unbuffered stderr.
We use a mix of message() calls when printing the configuration
summary, which caused interleaved output.
Because CMake offers no message(FLUSH), we work around the issue
by calling execute_process(COMMAND -E echo " ") which does
call std::cout << s << std::flush;
We seem to have to do it twice, before and after the
detailed configuration summary is printed.
Pick-to: 6.2 6.3
Change-Id: Ibc075551fc0547073f0696477e54d9b9c1edca97
Reviewed-by: Jörg Bornemann <joerg.bornemann@qt.io>
When Qt is configured as relocatable,
QT_CONFIGURE_PREFIX_PATH_STR -> qt_configure_prefix_path_str ->
qt_prfxpath
is not used in relevant code paths.
Specifically qlibraryinfo.cpp getPrefix() uses getRelocatablePrefix()
instead of QT_CONFIGURE_PREFIX_PATH.
Thus, when Qt is configured as relocatable, set qt_prfxpath to an
empty string.
This avoids embedding a CI path like /home/qt/work/install into the
official packages, which makes reproducible builds a closer reality.
Change-Id: I9209b08e651ad0b7fdc4049df333e0978e05f1f5
Reviewed-by: Jörg Bornemann <joerg.bornemann@qt.io>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
CMake has logic to rewrite build rpaths that contain
CMAKE_STAGING_PREFIX to instead point to CMAKE_INSTALL_PREFIX.
This breaks running executables from the build directory, because
their build rpath will point to a location where the libraries might
not exist yet (we didn't install Qt yet).
Work around this by setting CMAKE_STAGING_PREFIX to a fake path, so
that CMake does not do the rewriting anymore.
CMAKE_STAGING_PREFIX needs to be set at subdirectory scope, not
function scope, which is why
qt_internal_apply_staging_prefix_build_rpath_workaround() is a macro
that is called from within each Qt internal function that creates
a target.
The workaround can be disabled by configuring with
-DQT_NO_STAGING_PREFIX_BUILD_RPATH_WORKAROUND=ON
The downside of this workaround is that it breaks per-subdirectory
install rules like 'ninja src/gui/install'.
Regular global installation like 'ninja install' works fine.
This is similar to what we do for tests in
qt_set_up_fake_standalone_tests_install_prefix()
introduced by 20292250d4
The reason it's not as good for other target types is because in
contrast to tests, we do want to install them.
In case if someone does call `ninja src/gui/install' they will most
likely get a permission error, telling them it's not possible to
install into
/qt_fake_staging_prefix/
check_qt_internal_apply_staging_prefix_build_rpath_workaround
Fixes: QTBUG-102592
Change-Id: I6ce78dde1924a8d830ef5c62808ff674c9639d65
Reviewed-by: Jörg Bornemann <joerg.bornemann@qt.io>
The change
0ec75f4b99
missed adding specialHTMLTargets for qmake
Also add warning to keep QtWasmHelpers in sync with qmake.conf
Change-Id: Idb363e77f0cecb4f125d3cb4f7507899149a3bac
Reviewed-by: Morten Johan Sørvig <morten.sorvig@qt.io>
There should be no need for CMake to add rpaths pointing to
directories outside of the build tree to the installed libraries.
All relevant install rpaths are handled by qt_apply_rpaths().
Change-Id: If554b1e3c790c2bb04a34e8b0524aab3febf5afc
Reviewed-by: Jörg Bornemann <joerg.bornemann@qt.io>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
There were a few things that were not ported correctly.
Make sure to disable rpath manipulation if the rpath feature is
disabled.
Fix if(IS_ABSOLUTE) conditions to actually take values.
Don't embed bogus relative rpaths if the platform does not support
it. QNX is such a platform, it does not support $ORIGIN (at least from
my scouring of QNX documentation and manual testing via QEMU).
Handle the extra rpath case where they are relative, but the platform
does not support relative rpaths, by transforming them into absolute
ones.
Amends 67ee92f4d8
Change-Id: I04168633ec51b3cc5d580b738a7dc280fe6e0d2d
Reviewed-by: Jörg Bornemann <joerg.bornemann@qt.io>
Previously if
-extprefix /tmp/sysroot (CMAKE_STAGING_PREFIX)
-developer-build (FEATURE_developer_build)
were specified, but
-prefix (CMAKE_INSTALL_PREFIX)
was not,
the build system would set the CMAKE_INSTALL_PREFIX to the
qtbase build dir.
Then, if targeting desktop, this would be considered a non-prefix
build (ninja install would refuse to work), whereas in a cross-build
it would be considered an installable build.
In both cases though, the rpath of installed binaries would point to
the qtbase build dir (because CMAKE_INSTALL_PREFIX would be set to the
qtbase build dir).
This is quite confusing behavior, in more than one way.
Change the build system to consider that an explicit -extprefix should
cause Qt to always be installed, even if -developer-build is
specified.
This means the installed rpaths and on-device install prefix
(CMAKE_INSTALL_PREFIX) will now use the default computed install
prefix, e.g. /usr/local/Qt
It also means that to get a non-installable developer + custom staging
and install (on-device) prefix build, users will have to be explicit
and set all the options
-extprefix ~/qt/qtbase_build_dir
-prefix /usr
-developer-build
Pick-to: 6.2 6.3
Change-Id: Ib560452a4b4778860e0fd7666c76f8a6745773ee
Reviewed-by: Jörg Bornemann <joerg.bornemann@qt.io>
