-g4 was deprecated in Emscripten 2.0.17, and since we are well past
that version we can safely use -gsource-map instead.
Change-Id: I497155619d6320661f6be0e220b52fa7d6ca0b8e
Reviewed-by: David Skoland <david.skoland@qt.io>
Reviewed-by: Morten Johan Sørvig <morten.sorvig@qt.io>
[ChangeLog][wasm] Recommended emscripten version is now 3.0.0
Change-Id: I9ae082509415d8f00b8e9970b730ad4a472da310
Reviewed-by: Morten Johan Sørvig <morten.sorvig@qt.io>
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>
Building against a static debug MSVC Qt produced LNK4099 warnings (PDB
was not found with object file).
This was because we did not install the .pdb files for the object
libraries that are created for Qt resources. Now, these .pdb files are
named like the object library targets and are installed next to the
object files.
Pick-to: 6.2
Fixes: QTBUG-97699
Change-Id: I7e23f8392b7ac657be1d2fb3b33e051ae2e4d407
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
...if an Android platform < 10 is installed.
The existing platform detection code preferred android-9 over
android-31, because the sorting did not use natural comparison.
Natural comparison was added to CMake in version 3.18. We simulate this
feature for older CMake versions.
Pick-to: 6.2
Fixes: QTBUG-98726
Change-Id: Ib2eb87bd47220feb672275fa5203df4f2b6d7ca7
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
Like Q_NAMESPACE_EXPORT for Q_NAMESPACE, this variant of Q_GADGET
allows passing an export macro. This is useful to avoid exporting the
whole class just to get the staticMetaObject hidden therein exported.
Before anyone asks: No, we don't need Q_OBJECT_EXPORT, because QObject
subclasses, being polymorphic, always need to have a class-level
export macro (to export their vtable), but while that technique also
works for value classes (the Q_GADGET audience), it is not desirable
for them, because it makes inline functions exported in Windows debug
builds, which is not what we want, because it needlessly restricts
what you can to with the inline functions (e.g. remove).
[ChangeLog][QtCore] Added the Q_GADGET_EXPORT macro, which is like
Q_GADGET, but allows passing an export macro (like Q_NAMESPACE_EXPORT
for Q_NAMESPACE).
Fixes: QTBUG-55458
Change-Id: I546297de1e8aa45d83381991bcd3fbca61e1eef0
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
When building for and running the '_check' target in Android we need
to make sure that target apk is assembled and ready for deployment.
Pick-to: 6.2
Change-Id: Ifc63aac230f0bb8375d319ec443174c5a436485e
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
Large address aware is enabled by default in 64-bit
compilers, but not 32-bit compilers. But Qt users
may build 32-bit Qt themself, in this case large
address aware is disabled in fact, and it may cause
some issues. So we pass /LARGEADDRESSAWARE to the
linker unconditionally to make sure large address
aware is enabled for both 32-bit and 64-bit builds.
Microsoft Docs:
https://docs.microsoft.com/en-us/cpp/build/reference/largeaddressaware-handle-large-addresses?view=msvc-170
Change-Id: Idb2603d9ba0ba9ef4477ce1c3174b7c7e8ba76f6
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Oliver Wolff <oliver.wolff@qt.io>
Not a compile option, and also added to the linker
options above.
Change-Id: Iab2107d4b5ee4bc73e85b3cd5070f1a96b3a5a3b
Reviewed-by: David Skoland <david.skoland@qt.io>
Reviewed-by: Lorn Potter <lorn.potter@gmail.com>
- GHS compiler doesn't have -Wextra flag and it leads to huge count of warnings.
Pick-to: 6.2
Change-Id: Id2ba654a49fb163bebc75e3a22ecaa1895ecdbe8
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
In preparation for addition of new baseline tests, establish a new
test category, "baseline". This is similar to the category
"benchmarks" in that it contains tests that use the QTest framework,
but conceptually are not unit tests, in contrast to those under auto/.
Move the existing QPainter baseline test, tst_lancelot, into this new
category, and rename it accordingly.
Baseline tests use the QBaselineTest extension to QTest. Move that
extension too into the tests/baseline directory, allowing the clean
out of the baselineserver directory.
Pick-to: 6.2
Change-Id: I1b527f5867c953b1d22be73798fcf7d1494712ea
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Eirik Aavitsland <eirik.aavitsland@qt.io>
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
To allow using Android 12 APIs.
Pick-to: 6.2 5.15
Change-Id: I5e9da66c84457888ec723125d16876891232a99b
Reviewed-by: Ivan Solovev <ivan.solovev@qt.io>
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
This amends commit c5409964b0.
When detecting a variable assignment, we must ignore arguments that
start with a dash. Otherwise, arguments like --prefix=~/Qt are ignored.
Change-Id: I3b143113b94ca0d8af92679c1f567fbcec298349
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
The Qt bluetooth module uses ${FWCoreBluetooth} when building for iOS,
but that evaluated to empty.
Pick-to: 6.2
Change-Id: I93ff2c30697a912bed454087ae260fae9cc41b0f
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
We can technically build without Xcode, as long as the command line
tools are installed. We reflect that by an empty QT_MAC_XCODE_VERSION
variable.
Pick-to: 6.2
Change-Id: I40ffa28a628157e0a2c63656065ea7c41a680e01
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
And bump NTDDI_VERSION to 0x0A00000B (NTDDI_WIN10_CO) at the same time,
to unblock the developers from accessing the latest Windows APIs.
Pick-to: 6.2
Change-Id: Ifbc28c8f8b073866871685c020301f5f20dc9591
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
A top-level non-developer build is supposed to default to a prefix like
"/usr/local/Qt-6.3.0". That wasn't the case.
In QtSetup.cmake we check CMAKE_INSTALL_PREFIX_INITIALIZED_TO_DEFAULT,
set CMAKE_INSTALL_PREFIX and did the same thing later in
QtBuildInternalsExtra.cmake, with a different value. Make sure we run
the latter code only in per-repo builds.
Pick-to: 6.2
Fixes: QTBUG-98087
Change-Id: I4a1291dfd126aa11c21d69809f2cf7b075c98d2e
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
The qtpaths tool is supposed to replace "qmake -query", and it must be
available for cross-builds as a wrapper script like qmake.
Re-use the existing facility for creating the qmake wrapper script for
creating the qtpaths wrapper script.
Pick-to: 6.2
Fixes: QTBUG-97821
Change-Id: I460bae61a531994422e1c0fba09c79e4aa65713f
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
It's now possible to call configure with arbitrary variable assignments
"FOO=BAR" that get passed as "-DFOO=BAR" to CMake. There is no error
anymore for unknown variables. CMake already warns about those:
"Manually-specified variables were not used by the project: FOO".
[ChangeLog][configure] Users can directly assign CMake variables with
configure, for example "configure CMAKE_CXX_COMPILE=clang++-11".
Fixes: QTBUG-88210
Change-Id: Ib15e63a895df717919dd2b6623fa4d284209776f
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
Some apps may need more initial memory when linking, emscripten will
fail with:
wasm-ld: error: initial memory too small
This increases the default initial memory form 16MB to 20MB.
Larger apps might still get this error, in which case
QT_WASM_INITIAL_MEMORY can be set.
Pick-to: 6.2
Fixes: QTBUG-97457
Change-Id: Icdc56c41fb3a3852a9a83f7a3dc15820e83e1148
Reviewed-by: Morten Johan Sørvig <morten.sorvig@qt.io>
Misc. fixes, including:
Fix a couple of typos in the JavaScript code. Also, macros-
within-macros don’t work, (without resorting to preprocessor
token pasting), so remove the debug output for now.
Limit the exec() “simulateInfiniteLoop” workaround to
top-level application exec() only. This way, asyncify
can be used for nested QEventLoop::exec() calls. (Emscripten
supports one level of suspend only, so we don’t want
to use that for the top-level exec(), but instead use it
for dialogs and such).
Use the new QEventLoop::ProcessEventsFlag::ApplicationExec
enum value to detect the exec() call type.
Change-Id: Ic702bfc31faf2e9f84ac5d3ccf43d067c5c61bf0
Reviewed-by: Lorn Potter <lorn.potter@gmail.com>
The qt_finalize_framework_headers_copy function uses the module
QT_COPIED_FRAMEWORK_HEADERS property to generate the dependency list
for the ${target}_framework_headers target. In a common case elements
can be added to the QT_COPIED_FRAMEWORK_HEADERS property after the
qt_internal_add_module command call, that's why we need to make sure
that qt_finalize_framework_headers_copy is called after collecting all
headers assigned to the module.
Pick-to: 6.2
Change-Id: I2878fa6b8d4b11677c3f48345bf6e239221074c2
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
To be consistent with prepare_qdoc_args.
Pick-to: 6.2
Change-Id: Ibd40d6c26f6f6fa49ea78b6476f396153b7c4ee5
Reviewed-by: Paul Wicking <paul.wicking@qt.io>
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
If one of the dependencies in QtFooModuleDependencies.cmake is not
found, QtFoo_FOUND will be set to False by find_dependency.
In that case, we should not create imported targets that belong to
that package.
It would be misleading for projects that do target existence checks
instead of package_FOUND checks as well as generally being
incorrect. The created imported targets might be referencing other
targets that would not exist.
Pick-to: 6.2
Fixes: QTBUG-97896
Change-Id: I09198aa3f19be047b27c29329b6e62c30fa09dc0
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Craig Scott <craig.scott@qt.io>
The function qt_commandline_assignment and everything related to it is
removed from configure. It was only used in qtbase, and all usage has
been removed.
More general variable assignments will be added in a subsequent commit.
Task-number: QTBUG-88210
Change-Id: I7cfa782e89914f2b0dc0277c46e425c8a825557e
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
It's been deprecated since Mac OS X 10.5.
qmake templates fixed in 0768a28fbf.
Task-number: QTBUG-74872
Pick-to: 6.2
Change-Id: If7d988f27f07b6fa095b2ea51a87c306361d63d1
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
It's been there for ages, we may as well use it and remove unnecessary
complexity from CMake.
Pick-to: 6.2
Change-Id: I2bbf422288924c198645fffd16a9742567a7e4af
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Alexey Edelev <alexey.edelev@qt.io>
CMake's Modules/Platform/Android-Initialize.cmake was setting
CMAKE_FIND_USE_SYSTEM_ENVIRONMENT_PATH to OFF and host perl was not
picked up in PATH on Windows when building for Android.
Pick-to: 6.2
Change-Id: Ic98d8b38425231ef6aaf173fdfe958b573cc0c40
Reviewed-by: Craig Scott <craig.scott@qt.io>
One can now pass -DQT_REQUIRE_HOST_PATH_CHECK=OFF or
-DQT_REQUIRE_HOST_PATH_CHECK=ON to force disable or enable
the QT_HOST_PATH and QT_HOST_CMAKE_DIR_PATH validity checks.
One potential use case is if a project wants to manually find
QtFooTools packages by manipulating CMAKE_FIND_ROOT_PATH and such,
in which case they might not want to pass a QT_HOST_PATH.
Another potential use case is cross-building Qt with
-DQT_BUILD_TOOLS_WHEN_CROSSCOMPILING=ON and allowing the same
generated toolchain file to be used on the device image.
By default the generated toolchain file would require a host path
because it expects that it's still being used for cross-compiling.
But once the toolchain file is deployed to the device there's no need
to use the host tools anymore.
Another use case is building a desktop Qt using the host tools of
another desktop Qt. The new desktop Qt can use its newly built tools
just fine and wouldn't need the original desktop Qt tools.
QT_REQUIRE_HOST_PATH_CHECK needs to be added the list of vars to pass
through to try_compile calls.
Change-Id: I4b922b5d854828e6b9210dd8c07b4b1b8630aad1
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Craig Scott <craig.scott@qt.io>
The environment variable check to set extra env vars
was using invalid syntax. The condition always resolved to TRUE
which means the env vars were constantly re-assigned inside each
try_compile project.
To check for undefined-ness, one can use if(NOT DEFINED ENV{...})
To check for false-ness, one can use if(NOT "$ENV{...}")
To check for string emptiness, one can use
if(NOT "$ENV{...}" STREQUAL "")
In this particular case checking for false-ness is good enough.
The extra re-assigning had no visible effect, so this is just cleanup.
As a drive-by, clarify one comment.
Amends ca59c20939
Pick-to: 6.2
Change-Id: I8fd400101efa9e610a81268c33cac8c0cb33cba3
Reviewed-by: Alexey Edelev <alexey.edelev@qt.io>
Reviewed-by: Craig Scott <craig.scott@qt.io>
Instead of hard-coding a lib/cmake suffix to CMAKE_PREFIX_PATH
use the parent directory of the Qt6HostInfo package which is always
searched for when QT_HOST_PATH is set.
Amends d2359b2d4e
Pick-to: 6.2
Change-Id: I7f3c8fe17bc82f733580d68d928146434a57d849
Reviewed-by: Alexey Edelev <alexey.edelev@qt.io>
Previously when one wanted to use a cross-compiled Qt with a host Qt
installed in a non-default location, they'd have to provide both
QT_HOST_PATH and QT_HOST_PATH_CMAKE_DIR.
This change will now try to first check if ${QT_HOST_PATH}/lib/cmake
is a valid path on disk and use that. This is nicer to the user
because they don't need to specify 2 paths anymore.
Furthermore the path computation and sanity checks are now done after
any extra toolchain cmake files are loaded, to give an opportunity
to the files to set the paths first.
Finally, both variables need to be added to
__qt_toolchain_used_variables so they are passed along to try_compile
calls if the variables are specified manually.
Otherwise when the toolchain file is loaded by a try_compile project,
it will error out saying no host path found (as long as the initial
paths embedded in the toolchain are invalid).
Amends 93fc3afe71 and
ec90f9013b
Pick-to: 6.2
Change-Id: I433239b36b084f2d0a7a0dd043fdf87d77c138f3
Reviewed-by: Craig Scott <craig.scott@qt.io>
One can now set QT_FORCE_FIND_TOOLS to ON together with passing a
QT_HOST_PATH to ensure that a new desktop build uses
already existing tools from a different Qt host (desktop) installation.
Depends on a0e56294c1
to work, which is not in 6.2, but this change is still included in
6.2 because it cleans up the conditions a bit to make them more clear.
Amends 42d3b21c92
and 5a779a4ad3
Pick-to: 6.2
Fixes: QTBUG-95099
Task-number: QTBUG-97658
Change-Id: If6258fb1091c6c1e457f22ae5f468b811bd20d57
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
When cross-building, qt_internal_add_tool would accidentally find device
QtFooTools CMake packages as a result of calling find_package. It should
have found host Tools packages instead.
The reason was due to a combination of setting
CMAKE_FIND_ROOT_PATH_MODE_PACKAGE to BOTH, find_package preferring
the lookup of packages in CMAKE_FIND_ROOT_PATH before CMAKE_PREFIX_PATH
and there being a Tools package in the device sysroot.
Because qt_internal_add_tool didn't adjust CMAKE_FIND_ROOT_PATH to
contain the host path, the device package was picked up.
Change the implementation not to set CMAKE_FIND_ROOT_PATH_MODE_PACKAGE
and instead modify both CMAKE_FIND_ROOT_PATH and CMAKE_PREFIX_PATH
to prefer the host packages instead.
This aligns to the behavior that was introduced in
ec90f9013b
which is used in QtModuleDependencies when looking for Tools packages.
Pick-to: 6.2
Fixes: QTBUG-97599
Change-Id: I8e38284774ae97981ccfd5071465609f3de80f01
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
When building Qt repos, all find_package(Qt6) calls request a
PROJECT_VERSION version which is set in .cmake.conf via
QT_REPO_MODULE_VERSION.
This means trying to configure qtsvg from a 6.3 branch using a
6.2 qtbase won't work, because qtsvg will call find_package(Qt6 6.3)
and no such Qt6 package version exists.
There are certain scenarios where it might be useful to try to do
that though.
One of them is doing Qt development while locally mixing branches.
Another is building a 6.4 QtWebEngine against a 6.2 Qt.
Allow to opt out of the version check by configuring each Qt repo
with -DQT_NO_PACKAGE_VERSION_CHECK=TRUE. This setting is not
recorded and will have to be set again when configuring another
repo.
The version check will also be disabled by default when configuring
with the -developer-build feature. This will be recorded and embedded
into each ConfigVersion file.
If the version check is disabled, a warning will be shown mentioning
the incompatible version of a package that was found but that package
will still be accepted.
The warning will show both when building Qt or using Qt in a user
project.
The warnings can be disabled by passing
-DQT_NO_PACKAGE_VERSION_INCOMPATIBLE_WARNING=TRUE
Furthermore when building a Qt repo, another warning will show when an
incompatible package version is detected, to suggest to the Qt builder
whether they want to use the incompatible version by disabling the
version check.
Note that there are no compatibility promises when using mixed
non-matching versions. Things might not work. These options are only
provided for convenience and their users know what they are doing.
Pick-to: 6.2
Fixes: QTBUG-96458
Change-Id: I1a42e0b2a00b73513d776d89a76102ffd9136422
Reviewed-by: Craig Scott <craig.scott@qt.io>
Tests for QNX are run in qemu over ssh and setting WORKING_DIRECTORY
will cause created test wrapper to try change specified directory
which will fail (expects local directory) when test is invoked over
ssh.
Task-number: QTBUG-87628
Pick-to: 6.2
Change-Id: If991002398811cc9cb9cbac54da5d2cb0d8c3589
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
GNetworkMonitor can get information from 3 sources:
* org.freedesktop.portal.NetworkMonitor
* NetworkManager
* netlink
Change-Id: Icfafe6af42148f26360449f1262093ffc6b0613a
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
The previous implementation would pick up static Qt deps of a
shared Qt library as packages to find, but packages are not
created for these static libraries in a shared Qt build.
For example Qt::BundledSpirv_Cross is a static helper lib
that is linked directly into ShaderTools shared lib and no CMake
package is created for it, so we shouldn't look for it.
Separate the code path to filter out private dependencies
of a shared library target that don't have packages.
Amends 87215c70c0
Pick-to: 6.2
Fixes: QTBUG-97673
Task-number: QTBUG-86533
Change-Id: I43490b4d20c150256ccfa8b511a6e0e6b0f4b313
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Alexey Edelev <alexey.edelev@qt.io>
Private Qt module dependencies of a Qt module are recorded
in the IMPORTED_LINK_DEPENDENT_LIBRARIES property of a Qt module.
This property is used to compute the runtime dependency dir path
to be passed to the linker via the -rpath-link option.
If the referenced target does not exist in the scope where it's
used, no -rpath-link will be generated (or at least that specific
dir path won't be passed).
The linking operation will either fail saying the library is not found,
or a different version of the library might be silently picked up in
the sysroot or other implicit lib dir.
Make sure that QtFooModuleDependencies.cmake calls find_package() for
all Qt module private dependencies (or other Qt provided 3rd party
libs in the Qt6:: namespace) so that the targets are in scope and
IMPORTED_LINK_DEPENDENT_LIBRARIES does its job.
qmake also records the INTERFACE_LINK_LIBRARIES of a private Qt module
as the runtime dependencies of the module.
It's not clear why it does that. A private Qt module is an
INTERFACE_LIBRARY so it shouldn't add any new runtime dependencies.
Nevertheless, the find_package part of that has been recently addressed
in 2b6500cd15 for a different reason.
This change is basically the CMake equivalent of
326b91ea78
Pick-to: 6.2
Fixes: QTBUG-86533
Change-Id: Iaf514a14acaded4e8752149cca0c159a271be188
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Alexey Edelev <alexey.edelev@qt.io>
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
The constructor from a raw pointer should be
1) constexpr,
2) explicit, and
3) *private*.
We can do 1) without too much trouble.
2) is a (easy to fix) SIC in case of implicit conversions accidentally
relied upon from somewhere.
3) cannot be "easily" fixed by user code (they have to refactor), and
also, it's a BIC on Windows which encodes class members' access in
symbols. Someone may have been exporting some QList subclass, in turn
exporting the iterator classes, and therefore that someone now has the
constructors' symbols with a given access.
So, don't do 2+3 _just yet_ for user code, but set a deadline: Qt 6.5 is
the last that will support this. On Qt 6.6, we switch. All of this on
non-Windows, againt to avoid an ABI break. One can opt-in at any time
via a suitable define.
Given we have this define, use it to guard the other way around as well:
conversions from an iterator to a raw pointer should never be explicit
(there's std::to_address for this).
[ChangeLog][QtCore][QList] Converting a QList's iterator from and to a
raw pointer is deprecated, and will get removed in Qt 6.6. User code can
prepare for the change by defining QT_STRICT_QLIST_ITERATORS.
Change-Id: I0f34bfa3ac055c02af5a3ca159180304660dfc11
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Each included qml Qt6FooPluginTargets.cmake file checks whether
all the dependency targets that are referenced in the file already
exist by the time the file is included.
If one of the referenced targets is missing, the file sets
${CMAKE_FIND_PACKAGE_NAME}_NOT_FOUND_MESSAGE with a message mentioning
which targets are missing and also sets
${CMAKE_FIND_PACKAGE_NAME}_FOUND to FALSE.
All our qml package Config.cmake and Targets.cmake files are
include()d by Qt6QmlPlugins.cmake using a file(GLOB) which means the
order in which the files are loaded is implementation-defined.
Furthermore we didn't check the above set variables after each
inclusion, which means the values are overridden and the last plugin
to be loaded determines whether the Qml package is found or not.
If the last included file sets no error, it effectively silences any
previously set error.
Ever since we added dependencies between the qml plugin targets
themselves, we hit the above situation and depending on the platform,
no error was shown because the last file overrode any errors.
But we finally got a specific platform (wasm) which unearthed the
problem (QTBUG-97478). This can happen for any static Qt build though.
To fix this properly, we will most likely have to rewrite the whole
inclusion mechanism to use find_package() so that dependencies
can be resolved recursively as needed. This is a non-trivial change
that will have to be addressed in both qtbase and qtdeclarative.
In the mean time, a stop-gap solution that this change implements is
to include all the files while ignoring any error messages.
Then include the files one more time and check for error message after
each included file.
All qml plugin targets should have been brought into scope with the
first round of inclusions, thus circumventing "missing referenced
target" errors, while still catching any other possible errors.
Amends
6fd1216801f078f65be2cbc748cc459cb6912a4f
9fc302e6d146878103b3d105dce49c7695fcf93a
c368175a9e0a0c120b5bb8a0a02859bfc0cf42ba
in qtdeclarative.
Pick-to: 6.2 6.2.1
Fixes: QTBUG-97478
Task-number: QTBUG-95609
Task-number: QTBUG-97099
Change-Id: I157fa93fc979d726cd221d969b995b3642aeec77
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
Reviewed-by: Alexey Edelev <alexey.edelev@qt.io>
PRIVATE_MODULE_INTERFACE libraries are linked as interface libraries to
the module's Private target and exported as the dependencies of package
targets. We need to register these modules as public package dependencies
to call find_package when resolving module dependencies in user
projects.
Pick-to: 6.2
Fixes: QTBUG-96558
Change-Id: I4eef550aab306eaf357539ef7a0f76d69873f856
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
file(GENERATE) might fail if an unrelated configuration error happens,
and yet QT_TOOL_COMMAND_WRAPPER_PATH would already be set. Set the
cache variable only if generating was successful and replace the
QT_TOOL_COMMAND_WRAPPER_PATH valiable check with GLOBAL property to
protect the function from double call.
For CMake versions higher than or equal to 3.18 replace 'file(GENERATE'
call with 'file(CONFIGURE' to generate the wrapper at configure time
with the use of a plain semicolon character.
Pick-to: 6.2
Fixes: QTBUG-96870
Change-Id: Icf9c40f571d9c069043604f67ffcf2762966f6d0
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
When building Qt we need to set QT_ANDROID_ABIS value to
CMAKE_ANDROID_ARCH_ABI explictily since the automatical detecting of
android ABI is not executed. This fixes build of the qtbase in-tree
tests.
Fixes: QTBUG-97133
Change-Id: Ica7057bcfcc8f4fe4b5a921ca7449f74cdbca0f1
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
Reviewed-by: Assam Boudjelthia <assam.boudjelthia@qt.io>
