summaryrefslogtreecommitdiffstats
path: root/cmake/QtBuild.cmake
Commit message (Collapse)AuthorAgeFilesLines
* Write out a module description fileTobias Hunger2019-10-211-0/+20
| | | | | | | | | Write out a file with some JSON data to describe a module. This file contains information on how that module has been built. Change-Id: I8a604692663cbb7b76b96b97124130e30b822e4b Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
* Add ALLOW_UNDEFINED_SYMBOLS to add_qt_pluginLeander Beernaert2019-10-211-4/+6
| | | | | | | | Allow plugins to be built with undefined symbols when ALLOW_UNDEFINED_SYMBOLS is specified. Change-Id: I6bc809e3e5257302157bf8484f850d8319674a4a Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
* Rename builtin test data resource nameLeander Beernaert2019-10-181-1/+1
| | | | | | | | | | Rename the builtin test data resource name to something more unique as there are name clashes when building tests which also add a resource named 'testdata'. Change-Id: Icc1bbff3134e1dbc6ea4f6a87a1715b936c723cc Reviewed-by: Qt CMake Build Bot Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
* Pick the latest available C/C++ standard when compiling QtSimon Hausmann2019-10-171-0/+19
| | | | | | | | | | This maps the behavior of mkspecs/features/qt_common.prf and enables the use of C++17 for example in Android, where the toolchain supports it anyway. Change-Id: I41f4bdb160a3929e2fb78f36efb1ad5f2ad391a5 Reviewed-by: Qt CMake Build Bot Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
* Fix precompile header usageAlexandru Croitor2019-10-151-1/+1
| | | | | | | | | | | | | | Latest CMake will try to compile the given precompiled headers as C files, due to enabling the C language in the project + also having c files. Make sure to wrap the precompiled headers in a CXX language only generator expression. For details, see https://gitlab.kitware.com/cmake/cmake/issues/19839 Change-Id: Ib3508ad920092455c300b1a566ba6152bab032db Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
* Fix output property for add_cmake_library()Leander Beernaert2019-10-151-1/+3
| | | | | | | OUTPUT_DIRECTORY by itself is not a valid cmake target property. Change-Id: Ic3a2a81b8b982ad7ccf0551000a157f9a4d3ef22 Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
* Remove compatibility code for old call sites of qlalr processingSimon Hausmann2019-10-151-8/+0
| | | | | | Change-Id: I48801c336f7546b9366c70d877d20222c16dd4ff Reviewed-by: Qt CMake Build Bot Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
* Regenerate files after dev -> wip/cmake mergeAlexandru Croitor2019-10-141-0/+39
| | | | | | | | | | Note the following bigger things that had to be done: Handle GSS library / feature with a new custom find module. Implement rudimentary support for relocatability (does not currently handle extprefix). Change-Id: Ic6cd27dda7ebca9829f51cb42ea76fff6d1767ef Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
* Don't include header directories that don't existAlexandru Croitor2019-10-141-1/+1
| | | | | | | | | | | | If a module is marked as NO_MODULE_HEADERS, we shouldn't try to include the non-existing include directory which is usually generated by syncqt. Note it seems that NO_SYNC_QT and NO_MODULE_HEADERS seems to converge, so it might make sense to merge them in the future. Change-Id: Iab4e2907ed68776632337b37496b015535da8784 Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
* Fix Android build from multi-arch qmake changesLeander Beernaert2019-10-141-2/+2
| | | | | | | | | | Add condition replacements for the android ABIs. Add a replacement for QT_ARCH to ANDROID_ABI, since QT_ARCH is only used with the android build for now. Change-Id: I553d7910546de32236f723ec2e9a05a18da76130 Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
* Fix handling of OUTPUT_DIRECTORY in add_cmake_libraryLeander Beernaert2019-10-111-2/+7
| | | | | | | | extend_target() does not handle the OUTPUT_DIRECTORY argument, so we must handle it ourselves. Change-Id: I31880a516ae185f3255b2a51f41d61ee6b1d9838 Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
* Minor fix to qt_process_qlalr APISimon Hausmann2019-10-111-1/+9
| | | | | | | | | It's probably best to make it follow the usual calling convention that the associated (consuming) target is the first parameter of the function. So first this change accepts both formats. Change-Id: I1f20706b23d5e819e0eb689eecedb3afb49df3b7 Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
* Add SKIP_INSTALL to add_qt_plugin()Leander Beernaert2019-10-111-43/+47
| | | | | | | | Provide a SKIP_INSTALL argument to add_qt_plugin for test cases with plugins lacking install information. Change-Id: Iddb3843fab1790d69d64686530a46057a2ff0477 Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
* Add 'add_cmake_library' to QtBuild.cmakeLeander Beernaert2019-10-101-0/+62
| | | | | | | | | | | Add add_cmake_library to allow us to create normal cmake targets using all the information we have collected via the conversion script. This function is only meant for tests. For an example, see tests/auto/corelib/plugin/qpluginloader/lib/lib.pro. Change-Id: I738cb8ac241b8da1a1da3ef957c24dc7a754d43f Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
* Add install directory for plugins without typeLeander Beernaert2019-10-101-3/+3
| | | | | | | | | | | | | | | | When we run into a plugin that does not have a type and is not a qml plugin, we try to see if we can find the target installation path and provide INSTALL_DIRECTORY AND ARCHIVE_INSTALL_DIRECTORY to the add_qt_plugin call. We run into this frequently with the unit tests. This patch also changes add_qt_plugin() to use the value provided in INSTALL_DIRECTORY for ARCHIVE_INSTALL_DIRECTORY if no value is provided for the latter. Change-Id: I61278904a4d2d72308079cd362bd085b4e2f540c Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
* Match qt_import_plugin API with qt5'sJean-Michaël Celerier2019-10-091-3/+37
| | | | | | | | | | Some work was needed to make the plug-in types, and which plug-ins are available for each type in client code. Change-Id: Ib71feca31069deca3d3f54c8613054f5f8ae410c Reviewed-by: Simon Hausmann <simon.hausmann@qt.io> Reviewed-by: Qt CMake Build Bot
* Update add_qml_module() to use INSTALL_QML_FILESLeander Beernaert2019-10-081-0/+1
| | | | | | | | | | | Update add_qml_module() to use the new INSTALL_QML_FILES argument from qt6_add_qml_module(). This patch also updates pro2cmake.py to remove the QT_QML_SOURCE_INSTALL property from qml files. Change-Id: I6623b2de76bb55bd6750e48f7d45c53ca536b391 Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
* Update add_qml_module to use DO_NOT_INSTALL_METADATALeander Beernaert2019-10-081-1/+1
| | | | | | | Update to match latest changes to QtDeclarative. Change-Id: Ie455c0418e95c288149b4b1a29b065a8876e8b7e Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
* cmake: Add IOBluetooth to list of apple system frameworksOliver Wolff2019-10-011-0/+1
| | | | | | | | The framework is needed for qtconnectivity. Change-Id: I6a502564fb5543ca94ba5ae458a544286e34564c Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io> Reviewed-by: Qt CMake Build Bot
* Enchance qt_find_package with opt-in debug capabilityAlexandru Croitor2019-10-011-3/+27
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Sometimes when writing a qt_find_package and configuring a project, you might get the following error from CMake: CMake Error at QtBuild.cmake (set_property) Attempt to promote imported target "WrapFreetype::WrapFreetype" to global scope (by setting IMPORTED_GLOBAL) which is not built in this directory. This means that another find_package call, found WrapFreetype in another directory scope other than the current one, and thus the found target cat not be made global. Sometimes that implies that the qt_find_package might not be needed if WrapFreeType will always be found via a transitive depdendency. By setting QT_DEBUG_QT_FIND_PACKAGE=1 on the command line, you can make qt_find_package skip all of its behavior if the package was already found and the provided targets were also found. Unfortunately this behavior can not be made the default, because there is no way to find out in what scope the package was found, and if it's legal to make the targets global. At least I haven't found a way to do that yet. Thus the opt-in QT_DEBUG_QT_FIND_PACKAGE is a means to help with debugging such cases. Change-Id: I04242ed0f2fd0a75bc199386d28a1a0bd92da41a Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
* Add a find_library entry for the GameController frameworkMårten Nordheim2019-10-011-0/+1
| | | | | | | | Needed for QtGamepad Change-Id: I50227949eb509c021ba8d9b4dbe2b36b6a6de87f Reviewed-by: Qt CMake Build Bot Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
* Fix qt_find_package to not show incorrect packages at features summaryAlexandru Croitor2019-09-271-0/+7
| | | | | | | | | | | | | | | | | | | | | qt_find_package usually does 2 find_package() calls, one in CONFIG mode and one in MODULE mode. If the CONFIG mode doesn't find a Config file, the package_DIR cache variable is set to NOTFOUND, and this causes the FeatureSummary at the end to show that the package was not found, even if it is found by the next MODULE mode find_package call. Make sure to unset the _DIR variable in case if the Config module call fails. This fixes XRender showing up as not found even when it's found via the FindXRender.cmake file. Change-Id: I6ce39dad9cbb11836ca71f735a3267070c75b444 Reviewed-by: Cristian Adam <cristian.adam@qt.io> Reviewed-by: Albert Astals Cid <albert.astals.cid@kdab.com> Reviewed-by: Qt CMake Build Bot
* cmake: Fix QT_TOOL_PATH_SETUP_COMMANDJoerg Bornemann2019-09-261-1/+1
| | | | | | | | | | | | The set command looked like this: set \"PATH=...\" which was setting the environment variable \"PATH. Removing the escaped double quotes makes it actually work. Change-Id: I2c1d5d01b4415220512b005b75b7b67c695e33ae Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io> Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
* Add SKIP_TYPE_REGISTRATION option to add_qml_moduleLeander Beernaert2019-09-241-0/+8
| | | | | | | | | | If the qml files are not listed in the qmldir file they do not need to be written again, since it is expected that they are registered by the c++ plugin code. Change-Id: I624aedb182583f942bf1d1a322861c47a64b5065 Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io> Reviewed-by: Qt CMake Build Bot
* macOS: Prevent warnings about object files without any symbolsTor Arne Vestbø2019-09-241-0/+12
| | | | | | Change-Id: I0860b95cd75f536ff20defde97c3df7dc78a1e18 Reviewed-by: Qt CMake Build Bot Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
* Fix qtdeclarative buildAlexandru Croitor2019-09-231-4/+5
| | | | | | | | | | | | There is an issue about versionless tool target not finding the regular targets, and that fails the configuration phase of qtdeclarative. Temporarily don't export the versionless targets, to get the qt5 build going. Change-Id: I5c7baff7f677f4a3f1f91b9e8082ba8a80f9cddd Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
* Create module tools packages (config files) automaticallyAlexandru Croitor2019-09-221-6/+1
| | | | | | | | | | | | | | | | | | | | | Previously you had to make sure to use DISABLE_TOOLS_EXPORT in an add_qt_module call if the tools are built after the module, as well as to manually call qt_export_tools after all associated tools are built. This was needlessly complex, especially for people that are porting a repo with tools for the first time. The tools package creation is now automatically done at QtPostProcess step, so there is no need to use either DISABLE_TOOLS_EXPORT or qt_export_tools() manually. DISABLE_TOOLS_EXPORT is now a no-op, and will be removed once all repos are updated not to use it. Change-Id: I965b0d3a8a0cb908afae87b047083ed7bea9f02f Reviewed-by: Qt CMake Build Bot Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
* PCH: Only enable automatically for libraries/modulesMårten Nordheim2019-09-191-1/+11
| | | | | | | | | | | Some tests are changing headers using defines inside the tests. Let's more closely mirror what qmake does to fix this. And it will also save quite a lot of space since most tests don't include all of e.g. QtCore Change-Id: I6f7e530f922418944d690bd2a1ee5f459ba755e1 Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io> Reviewed-by: Qt CMake Build Bot
* Add Qml plugin example supportLeander Beernaert2019-09-191-37/+42
| | | | | | | | | | | Extend the example writing section to handle qml plugins. Detection is based on whether the plugin config is detected an the target links against Qml. add_qt_qml_module() now uses the the public facing API underneath. Change-Id: I2ccd06548e7b5b18c9dc3338b835bc115fa97809 Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
* Check if Qt source dir is not empty to use syncqtCristián Maureira-Fredes2019-09-181-1/+1
| | | | | | | | | In an extreme case where /bin/syncqt.pl does exists, is better to check if the QtBase_SOURCE_DIR is not empty to avoid a mismatch script version that is being used. Change-Id: Ia5694eadc5517998b827eccf70bdf6f3c14ebfa3 Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
* Fix QtDeclcarative static buildsLeander Beernaert2019-09-181-6/+4
| | | | | | | | Incorrect variable name was being used to set the target in the parent scope. Change-Id: I73ea644ebf94c9b9a62b34b1ad493e488729ff2f Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
* Fix up documentation integration placeholderSimon Hausmann2019-09-181-1/+11
| | | | | | | | | | In the future need will need to continue to tie qdoc runs still to targets, just like with qmake. This change prepares us for that by ensuring that add_docs takes two parameters and that any re-generated CMakeLists.txt from now on gets it right. Change-Id: Id0256dc1e2f2f59f3b4e4ca98f0d10d025d189fb Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
* Fix for getting windows running testsMårten Nordheim2019-09-171-1/+7
| | | | | | | | | | | | | | | | | By adding the path to the DLLs early on in the path. This fixes the issue seen in CI (0xc0000135, DLL not found) and resolves local issues where you might have forgotten to add this to path yourself potentially grabbing libraries from elsewhere. The ${path} seems to be a holdover that is no longer used, so it was removed while the code was changed anyway. Also disable WIN32_EXECUTABLE for all tests so that we can actually get some output from them :) Change-Id: Iec42c809c37be4f31c7f0a7af3a30c3528022dbe Reviewed-by: Simon Hausmann <simon.hausmann@qt.io> Reviewed-by: Qt CMake Build Bot
* pro2cmake: Support QML IMPORTPATH with multiple elementsJüri Valdmann2019-09-171-4/+4
| | | | | | Change-Id: I8113d7dd4e7967d020d59a5b4104e8366d55283c Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io> Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
* Teach automoc to run moc when it sees Q_ENUM_NSAlexandru Croitor2019-09-171-1/+1
| | | | | | | | | Needed for qtcoap, otherwise AUTOMOC doesn't run moc on qcoapnamespace.h. Change-Id: I4ca43fcbbc5db6163f9f9f788b920eae86f5b174 Reviewed-by: Qt CMake Build Bot Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
* Prospective fix for QtGui private symbol versioningSimon Hausmann2019-09-171-1/+1
| | | | | | | | | | | | | In qmake, the header files used for private symbol versioning is done via private_api_headers = $$SYNCQT.PRIVATE_HEADER_FILES $$SYNCQT.QPA_HEADER_FILES So we must do the same with CMake. Change-Id: Iaebeb13592241b6c4d89f70d2e6ac3ebfb374207 Reviewed-by: Qt CMake Build Bot Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
* Fix OUTPUT_TARGET propagation in qt6_add_resource()Leander Beernaert2019-09-161-6/+8
| | | | | | | | | | | Value was not being propagated to the parent scope when set. This patch also changes OUTPUT_TARGET to OUTPUT_TARGETS since it is possible that two targets can be generated. Change-Id: If489a609ed363a319224fcd6c5a4fc878d0d8617 Reviewed-by: Simon Hausmann <simon.hausmann@qt.io> Reviewed-by: Qt CMake Build Bot
* Don't set the MODULE_HEADER property on header-only modulesMårten Nordheim2019-09-131-1/+3
| | | | | | | | | | | | | | | | They're INTERFACE-type targets and can thus only have whitelisted properties set. That fixes the cmake configure step for the UiPlugin target in qttools. This has the unfortunate side-effect that the headers will not be picked up for our pre-compiled headers. Although it is not a big issue since we don't have many header-only modules. An example is QtTools' UiPlugin. Amends 2cf0ba1fba9293b3265a186527dbc90d395dfd20 Change-Id: If722928f64727ffaf2e9d0746668c0198fa1a647 Reviewed-by: Simon Hausmann <simon.hausmann@qt.io> Reviewed-by: Qt CMake Build Bot
* Use pre-compiled headers when building Qt with cmakeMårten Nordheim2019-09-111-4/+41
| | | | | | | | | | | | | | | | | | Some modules define their own manually-maintained lists, and we can rely on the headers generated by each module to include in the pch as well e.g. QtCore/QtCore. There's also e.g. QtWidgetDepends for QtWidgets, but this only works for modules, not for tools, examples or other applications. For now we'll use the Qt<Module>/Qt<Module> headers for the modules we depend on. Building with PCH can be disabled with -DBUILD_WITH_PCH=NO, and it only works for versions of CMake newer than 3.15.20190829. Change-Id: Iae52bd69acfdfd58f4cd20d3cfa3c7f42775f732 Reviewed-by: Qt CMake Build Bot Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
* Don't propagate INSTALL_INTERFACE headers for modules with no syncqtAlexandru Croitor2019-09-101-1/+1
| | | | | | | | | | If syncqt was not executed for a module, it will not have generated headers, so we should not propagate the include/${module} header location in that case. Change-Id: I6dc0628a11ababb4d237215a9f4d3fc331383848 Reviewed-by: Liang Qi <liang.qi@qt.io> Reviewed-by: Qt CMake Build Bot
* Fix build of installed header-only modules in non-prefix buildsSimon Hausmann2019-09-051-4/+4
| | | | | | | | | | QtUiTools in qttools is a public module that has only headers, so it's an interface library in cmake. That means INTERFACE_INCLUDE_DIRECTORIES cannot contain paths to the source or build directory, unless they are prefixed with BUILD_INTERFACE, which this patch adds. Change-Id: I047700f447237dfbe5a901072eb413a159ae537d Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
* Fix qmlcachegen invocation on Windows, part 2Simon Hausmann2019-09-041-3/+1
| | | | | | | | | | | | Fix up the previous commit and use separate COMMAND parameters for add_custom_command to ensure the expansion of the PATH. The injection via -E env "/some/dir;%PATH%" does not work as the dollar expansion is not applied when the ninja command uses cmd /C. This relies now on undocumented behavior of cmake ;( Change-Id: I5b5fc88e4c13f8fb6c6bba3131204c2eb35404d6 Reviewed-by: Qt CMake Build Bot Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
* Fix qtdeclarative build on WindowsSimon Hausmann2019-09-041-0/+14
| | | | | | | | | | | | | | Our tools are typically installed into the Qt bin directory, after which they are useable right away on Windows, since the Qt dlls are located there, too. An exception to the rule are tools that are built within a module and used right away. At that point they are in the top-level bin dir of the build directory, along with "local" DLLS they might need, but with the Qt dlls out of reach. To cover this case, we need to prepend a PATH adjusting command when using these with add_custom_command. Change-Id: I7f58581f5060c8004b5d1fa1f6f17297256783de Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
* Remove dead codeSimon Hausmann2019-09-041-71/+0
| | | | | | | Change-Id: Iac7c7c77744b43772faf3402566685a674ab8ff3 Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io> Reviewed-by: Qt CMake Build Bot Reviewed-by: Leander Beernaert <leander.beernaert@qt.io>
* Don't include module syncqt header locations if syncqt was not runAlexandru Croitor2019-09-021-17/+27
| | | | | | | | | | | | When building QtQmlDevTools we don't generate syncqt headers, but the non-existent include paths were still added, which means that anything that used QtQmlDevTools would fail to configure (the standalone tests). Make sure to only add the include paths if syncqt was executed. Change-Id: I6d79ecf53e9a5d396e8df801584ce2c9f119f9be Reviewed-by: Simon Hausmann <simon.hausmann@qt.io> Reviewed-by: Qt CMake Build Bot
* Update Android build configurationLeander Beernaert2019-08-301-0/+8
| | | | | | | | | | | | | | | | | | Update Android build configuration to be compatible with the multi-arch android build patch to qmake. We can now build and generate the apk correctly. Executing on the device/simulator will only work once the latest changes from 5.14 have been merged in. We now replace target suffix with ${CMAKE_ANDROID_ARCH_ABI}.[so|a] so we don't have to deal with handling targets which might have any of the OUTPUT_NAME properties set. The dependency and deploy settings generation has also been updated to append the file contents to a string and then do a single file write at the end. Change-Id: Id3c5bd0428141ecaf962124a100390e3a4e41feb Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
* Fix qtdeclarative static buildsAlexandru Croitor2019-08-301-0/+5
| | | | | | | | qmlplugindump is not build during a Qt static build, so there's no point in creating custom targets that would run the tool. Change-Id: Id7d6d44da63312f298f46eab08a5637cb0c97c8f Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
* Fix static builds after revamp of add_qt_resourceAlexandru Croitor2019-08-301-2/+3
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | The first issue is that instead of arg_OUTPUT_TARGET we should use rcc_OUTPUT_TARGET (there was also a typo in the OUTPUT word). The second issue came with the fact that the object library targets that were created for resources had a "Qt6::" namespace prefix in the exported Targets files, and yet the link generator expression did not contain the prefix. This failed when building qtsvg in a static build, because the exported object library could not be found. The solution is to use the TARGET_NAME generator expression which takes care of adjusting the name of the target when exported, thus making the linking work both when building qtbase and when building qtsvg. This had the fallout, that all resource object libraries need to be associated with an export set and installed. This wasn't the case for plugins, because plugins have an export name of the form "FooTargets", whereas the export name for the resource object library is "Qt6FooTargets". Adjust the export names of plugins to be the same as for modules and resource library objects (aka contain the "Qt6 prefix"). Change-Id: I1ca28d20c51e4447e5783cc20571a68ec77f6513 Reviewed-by: Qt CMake Build Bot Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
* Fix "mkspec selection" when building with EmscriptenSimon Hausmann2019-08-281-0/+2
| | | | | Change-Id: Id9d466221294651c2f8dfde149d82f45b4081238 Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
* Fix CMake errors around linker flags when targeting WASMSimon Hausmann2019-08-281-1/+4
| | | | | | | | | | | * The -no-undefined linker option only makes sense for shared libraries. The WASM build is a static build though. * The -gc-sections linker option is only relevant for host tools and certainly applicable to the WASM toolchain. Change-Id: Ib8daec66a6d80f5025887448882dd891e6176268 Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
generated by cgit v1.2.3 (git 2.25.1) at 2024-06-09 01:07:46 +0000