| b5602e7 | Brad King | 20 March 2024, 14:03:43 UTC | CMake 3.28.4 | 20 March 2024, 14:03:43 UTC |
| 5969b4f | Brad King | 20 March 2024, 13:54:50 UTC | Merge topic 'ninja-fortran-include' into release-3.28 36dc8d6d50 Ninja: Fix Fortran module deps in files INCLUDEd by preprocessed sources Acked-by: Kitware Robot <kwrobot@kitware.com> Merge-request: !9351 | 20 March 2024, 13:54:50 UTC |
| 36dc8d6 | Brad King | 18 March 2024, 23:28:09 UTC | Ninja: Fix Fortran module deps in files INCLUDEd by preprocessed sources If a preprocessed source also uses the Fortran `INCLUDE` directive, search the original source file's directory. Fixes: #25792 | 19 March 2024, 11:56:13 UTC |
| 8ebdee9 | Brad King | 14 March 2024, 15:46:08 UTC | Merge topic 'autogen-makefile-depfile' into release-3.28 d2d1763f88 cmQtAutoGenInitializer: De-duplicate autogen/timestamp target depend logic 6193d15556 Autogen: Restore target-ordering dependencies in Makefiles with DEPFILE Acked-by: Kitware Robot <kwrobot@kitware.com> Merge-request: !9340 | 14 March 2024, 15:46:08 UTC |
| d2d1763 | Brad King | 13 March 2024, 16:39:32 UTC | cmQtAutoGenInitializer: De-duplicate autogen/timestamp target depend logic We place the same target ordering dependencies on either the `_autogen_timestamp_deps` target or the `_autogen` target. Refactor the logic to avoid duplicating that code. | 14 March 2024, 13:18:56 UTC |
| 6193d15 | Brad King | 13 March 2024, 15:39:02 UTC | Autogen: Restore target-ordering dependencies in Makefiles with DEPFILE In commit aebfbcaa46 (AutoGen: Use depfiles for the XXX_autogen ninja targets, 2020-01-14, v3.17.0-rc1~58^2) the `_autogen_timestamp_deps` target was given target ordering dependencies through its custom command rather than direct target dependencies as on the `_autogen` target. Then commit 895fa3433f (cmQtAutoGenInitializer: support IMPLIB-only imported targets, 2021-09-23, v3.22.0-rc1~80^2) converted some target-level dependencies into file-level dependencies on the custom command. This only works with a monolithic build graph like Ninja. Since commit ebc9e448b3 (Autogen: Add depfile support for Makefiles, 2023-09-07, v3.28.0-rc1~101^2~1) we use the `_autogen_timestamp_deps` target in Makefile generators too. This exposed the missing target ordering dependency. Fixes: #25766 | 14 March 2024, 13:05:06 UTC |
| 7a9fc8c | Brad King | 13 March 2024, 14:52:00 UTC | Merge topic 'backport-3.28-ci-xcode-15.3' into release-3.28 8a6d05aa9d gitlab-ci: update macOS jobs to use Xcode 15.3 7bf893bc30 Tests: Update RunCMake.CheckCompilerFlag for Xcode 15.3 7f5f0ab4b3 Tests: Update RunCMake.CompileFeatures for Xcode 15.3 Acked-by: Kitware Robot <kwrobot@kitware.com> Merge-request: !9333 | 13 March 2024, 14:52:15 UTC |
| 8a6d05a | Brad King | 11 March 2024, 23:10:03 UTC | gitlab-ci: update macOS jobs to use Xcode 15.3 | 12 March 2024, 13:29:45 UTC |
| 7bf893b | Brad King | 12 March 2024, 13:27:34 UTC | Tests: Update RunCMake.CheckCompilerFlag for Xcode 15.3 Xcode passes a new `-use-frontend-parseable-output` flag to Swift that conflicts with our `-parseable-output` flag. Use a different flag for the test case. | 12 March 2024, 13:29:45 UTC |
| 7f5f0ab | Brad King | 12 March 2024, 00:46:27 UTC | Tests: Update RunCMake.CompileFeatures for Xcode 15.3 Xcode no longer puts an extra backslash in its `-std=` flags. | 12 March 2024, 13:29:29 UTC |
| 9bfb972 | Brad King | 08 March 2024, 14:32:01 UTC | Merge topic 'cxxmodules-target-objects' into release-3.28 4ddf0453d5 Tests/CXXModules: add a test using `TARGET_OBJECTS` from modules Acked-by: Kitware Robot <kwrobot@kitware.com> Merge-request: !9316 | 08 March 2024, 14:32:01 UTC |
| 4ddf045 | Ben Boeckel | 06 March 2024, 06:13:01 UTC | Tests/CXXModules: add a test using `TARGET_OBJECTS` from modules See: #25732 | 06 March 2024, 17:34:46 UTC |
| 944f10d | Brad King | 04 March 2024, 15:19:34 UTC | Merge topic 'cxxmodules-invalid-iterator' into release-3.28 844f054fe0 cmImportedCxxModuleInfo: remove unused pointer field Acked-by: Kitware Robot <kwrobot@kitware.com> Merge-request: !9303 | 04 March 2024, 15:19:34 UTC |
| 844f054 | Ben Boeckel | 01 March 2024, 16:18:18 UTC | cmImportedCxxModuleInfo: remove unused pointer field Also avoids usage of an invalid `importit` when module names are not available. | 01 March 2024, 16:23:44 UTC |
| d16ac6f | Brad King | 01 March 2024, 13:59:04 UTC | Merge topic 'FindTIFF-cmake-package' into release-3.28 b6b2f73f36 FindTIFF: Fix name of upstream TIFF cmake package Acked-by: Kitware Robot <kwrobot@kitware.com> Merge-request: !9300 | 01 March 2024, 13:59:04 UTC |
| b6b2f73 | Brad King | 29 February 2024, 13:12:26 UTC | FindTIFF: Fix name of upstream TIFF cmake package Since commit 6a3059e66f (FindTIFF: bridge `tiff-config` into FindTIFF-compatible interface, 2023-09-14, v3.28.0-rc1~87^2) we try to find the upstream TIFF cmake package. However, it is called `TiffConfig.cmake`, not `tiff-config.cmake`, so we need to match the capitalization of the package name. | 29 February 2024, 13:47:53 UTC |
| 81a8cf2 | Brad King | 29 February 2024, 13:01:22 UTC | Merge topic 'cxxmodules-no-export-basedirs-as-include-paths' into release-3.28 028f3134e5 cmExportFileGenerator: only export include paths for HEADERS file sets Acked-by: Kitware Robot <kwrobot@kitware.com> Merge-request: !9295 | 29 February 2024, 13:01:22 UTC |
| 6fa3cf4 | Brad King | 28 February 2024, 20:01:06 UTC | Merge topic 'test-cxxmodules-vs' into release-3.28 6925978f3e Tests/RunCMake/CXXModules: Update VS circular error message Acked-by: Kitware Robot <kwrobot@kitware.com> Merge-request: !9299 | 28 February 2024, 20:01:06 UTC |
| 6925978 | Brad King | 28 February 2024, 19:14:25 UTC | Tests/RunCMake/CXXModules: Update VS circular error message Match the message expected from MSBuild as of VS 17.9. | 28 February 2024, 19:16:20 UTC |
| 028f313 | Ben Boeckel | 28 February 2024, 00:15:23 UTC | cmExportFileGenerator: only export include paths for HEADERS file sets | 28 February 2024, 10:26:10 UTC |
| febe479 | Brad King | 27 February 2024, 14:02:54 UTC | Merge topic 'cxxmodules-fileset-cache-fullpath' into release-3.28 0a18f9baad cmGeneratorTarget: collapse paths before querying the fileset cache Acked-by: Kitware Robot <kwrobot@kitware.com> Merge-request: !9286 | 27 February 2024, 14:02:54 UTC |
| 5feddb6 | Brad King | 27 February 2024, 14:02:27 UTC | Merge topic 'clang-use-arg1' into release-3.28 7008fc234b CMakeDetermineCompilerId: use `_ARG1` to query the clang resource dir Acked-by: Kitware Robot <kwrobot@kitware.com> Merge-request: !9285 | 27 February 2024, 14:02:27 UTC |
| 7008fc2 | Ben Boeckel | 24 February 2024, 21:54:09 UTC | CMakeDetermineCompilerId: use `_ARG1` to query the clang resource dir Other core queries do this, so do it here too. | 26 February 2024, 16:49:17 UTC |
| 0a18f9b | Ben Boeckel | 24 February 2024, 22:17:22 UTC | cmGeneratorTarget: collapse paths before querying the fileset cache Observed with paths with `../` when outside of the source tree. I attempted to reproduce a test suite case for it but could not make it happen. | 24 February 2024, 22:17:39 UTC |
| e98cbbf | Brad King | 22 February 2024, 13:23:46 UTC | Merge topic 'restore-link-interface-objlib-with-unity' into release-3.28 5b8e9e068f Restore support for TARGET_OBJECTS in link interfaces with unity builds 1313c78a9c Tests: Update RunCMake.TargetObjects cmake_minimum_required version Acked-by: Kitware Robot <kwrobot@kitware.com> Merge-request: !9279 | 22 February 2024, 13:23:46 UTC |
| 1371a28 | Brad King | 22 February 2024, 13:22:49 UTC | Merge topic 'cxxmodules-clang-resource-dir' into release-3.28 a3ada1a241 Clang: detect -resource-dir for clang-scan-deps Acked-by: Kitware Robot <kwrobot@kitware.com> Merge-request: !9274 | 22 February 2024, 13:22:49 UTC |
| 5b8e9e0 | Brad King | 21 February 2024, 21:00:42 UTC | Restore support for TARGET_OBJECTS in link interfaces with unity builds This was broken by commit df08c37a42 (cmGlobalGenerator: Add unity/pch sources after computing compile features, 2024-02-02, v3.28.3~1^2~1^2), and 3.28.2's commit 76b5383123 (cmGlobalGenerator: add unity sources after computing target compile features, 2024-01-01, v3.28.2~17^2~1). The problem is very similar to that fixed by commit 4e8f24e977 (PCH: Clear link interface cache when adding PCH object to it, 2022-01-24, v3.23.0-rc1~44^2~9). Generalize that fix. Fixes: #25696 | 21 February 2024, 22:08:57 UTC |
| 1313c78 | Brad King | 21 February 2024, 21:22:07 UTC | Tests: Update RunCMake.TargetObjects cmake_minimum_required version This was missed in commit 1edf138506 (Tests/RunCMake: Update cmake_minimum_required versions, 2023-02-06, v3.27.0-rc1~508^2~1). | 21 February 2024, 21:22:35 UTC |
| 0d3f35f | Brad King | 21 February 2024, 18:29:36 UTC | Merge topic 'cxxmodules-transitive-std' into release-3.28 f0df692d21 Tests/CXXModules: test `cxx_std_20` coming from a dependency Acked-by: Kitware Robot <kwrobot@kitware.com> Merge-request: !9276 | 21 February 2024, 18:29:36 UTC |
| edfdec2 | Brad King | 21 February 2024, 18:28:52 UTC | Merge topic 'cxxmodules-transitive-modules-usage' into release-3.28 ec348ee4b9 cmGeneratorTarget: discover synthetic targets recursively 67466ddf65 cmExportFileGenerator: export link libraries as-is a2e3e61a31 Tests/CXXModules: test transitive modules usage Acked-by: Kitware Robot <kwrobot@kitware.com> Merge-request: !9273 | 21 February 2024, 18:28:52 UTC |
| a3ada1a | Ben Boeckel | 20 February 2024, 13:35:22 UTC | Clang: detect -resource-dir for clang-scan-deps When cross-compiling, `clang-scan-deps` needs help to find the correct location of core headers such as `stddef.h`. Always determine this path and pass it when available. Fixes: #25590 | 21 February 2024, 13:51:42 UTC |
| f0df692 | Ben Boeckel | 20 February 2024, 17:31:37 UTC | Tests/CXXModules: test `cxx_std_20` coming from a dependency Closes: #25569 | 20 February 2024, 17:32:48 UTC |
| ec348ee | Ben Boeckel | 20 February 2024, 13:00:41 UTC | cmGeneratorTarget: discover synthetic targets recursively Fixes: #25288 | 20 February 2024, 15:53:31 UTC |
| 67466dd | Ben Boeckel | 20 February 2024, 12:38:58 UTC | cmExportFileGenerator: export link libraries as-is This allows for transitive modules to work because `$<COMPILE_ONLY>`-wrapped dependencies do not end up in the `linked-target-dirs` collator property. Test suite exported property tests updated to account for the change. | 20 February 2024, 15:53:31 UTC |
| a2e3e61 | Ben Boeckel | 16 February 2024, 22:28:55 UTC | Tests/CXXModules: test transitive modules usage | 20 February 2024, 15:53:31 UTC |
| 583641a | Brad King | 20 February 2024, 15:29:18 UTC | Merge topic 'cxxmodules-transitive-target-usage' into release-3.28 8b6fc81fc3 cmTarget: copy link libraries from the right properties d4a517f82a Tests/CXXModules: add a test with transitive targets Acked-by: Kitware Robot <kwrobot@kitware.com> Merge-request: !9265 | 20 February 2024, 15:29:18 UTC |
| 8b6fc81 | Ben Boeckel | 16 February 2024, 21:33:52 UTC | cmTarget: copy link libraries from the right properties This fixes transitive dependencies' usage requirements not appearing when compiling synthetic targets. See: https://discourse.cmake.org/t/9819 | 19 February 2024, 15:25:29 UTC |
| d4a517f | Ben Boeckel | 16 February 2024, 18:42:42 UTC | Tests/CXXModules: add a test with transitive targets | 19 February 2024, 15:25:27 UTC |
| 00d908a | Brad King | 19 February 2024, 14:23:58 UTC | Merge topic 'cxxmodules-nmc-duplicate-synthetic-targets' into release-3.28 5261af9424 cmGeneratorTarget: store synthetic targets in its cache e0633a9517 Tests/CXXModules: add a test importing from a `Ninja` install 150d7dbd68 Tests/CXXModules: support building a project with `Ninja` e48e5e5506 Tests/CXXModules: document `CMake_TEST_MODULE_COMPILATION` items Acked-by: Kitware Robot <kwrobot@kitware.com> Merge-request: !9263 | 19 February 2024, 14:23:58 UTC |
| 5261af9 | Ben Boeckel | 16 February 2024, 13:42:57 UTC | cmGeneratorTarget: store synthetic targets in its cache The synthetic target cache was never actually updated, so record them in the cache so that reuses can actually be discovered. Fixes: #25568 | 16 February 2024, 16:09:56 UTC |
| e0633a9 | Ben Boeckel | 16 February 2024, 13:42:14 UTC | Tests/CXXModules: add a test importing from a `Ninja` install This ensures that consuming from a single configuration generator works in all generators. See: #25568 | 16 February 2024, 16:09:56 UTC |
| 150d7db | Ben Boeckel | 16 February 2024, 13:41:39 UTC | Tests/CXXModules: support building a project with `Ninja` This allows testing mixed generator setups. | 16 February 2024, 16:09:54 UTC |
| e48e5e5 | Ben Boeckel | 16 February 2024, 12:10:24 UTC | Tests/CXXModules: document `CMake_TEST_MODULE_COMPILATION` items | 16 February 2024, 15:28:10 UTC |
| fb0988a | Brad King | 16 February 2024, 14:40:07 UTC | Merge topic 'cxxmodules-dyndep-error-on-private-usage' into release-3.28 3f8a59a05c cxxmodules: return failure from the collator when private usage is found Acked-by: Kitware Robot <kwrobot@kitware.com> Merge-request: !9257 | 16 February 2024, 14:40:07 UTC |
| d713896 | Brad King | 16 February 2024, 14:30:50 UTC | Merge topic 'cxxmodules-collation-restat' into release-3.28 a7424b636b Ninja: make the collator rule use `restat = 1` e24eecfc33 Tests/CXXModules: add a test to ensure that `restat` works for collation Acked-by: Kitware Robot <kwrobot@kitware.com> Merge-request: !9259 | 16 February 2024, 14:30:50 UTC |
| d8fd8c2 | Brad King | 16 February 2024, 14:22:32 UTC | Merge topic 'cxxmodules-visibility-change-rebuild' into release-3.28 60a8736378 cmNinjaTargetGenerator: scanning depends on the module metadata 52036ce090 Tests/CXXModules: test that objects depend on dependent modules json files 9a45c9fbd5 cmNinjaTargetGenerator: use `emplace_back` for scanning deps Acked-by: Kitware Robot <kwrobot@kitware.com> Merge-request: !9247 | 16 February 2024, 14:22:32 UTC |
| 2808c49 | Brad King | 16 February 2024, 14:19:47 UTC | Merge topic 'cxxmodules-depend-on-modmap-file' into release-3.28 363300ace5 cxxmodules: depend on the modmap contents 940628d48d Tests/CXXModules: test that objects depend on their modmap files 681ee92030 Tests/CXXModules: support testing rebuild conditions Acked-by: Kitware Robot <kwrobot@kitware.com> Merge-request: !9248 | 16 February 2024, 14:20:09 UTC |
| 148af75 | Brad King | 16 February 2024, 14:16:54 UTC | Merge topic 'ci-test-CommandLength' into release-3.28 edb833d69e Tests/CommandLength: Refactor test commands to reduce test run time Acked-by: Kitware Robot <kwrobot@kitware.com> Merge-request: !9256 | 16 February 2024, 14:17:14 UTC |
| 403af46 | Brad King | 15 February 2024, 19:33:29 UTC | Merge branch 'backport-3.28-visionOS' into release-3.28 dbfc53a2bf Help: Add 3.28.4 release note draft for visionOS updates Acked-by: Kitware Robot <kwrobot@kitware.com> Merge-request: !9262 | 15 February 2024, 19:33:48 UTC |
| dbfc53a | Brad King | 15 February 2024, 18:13:08 UTC | Help: Add 3.28.4 release note draft for visionOS updates Hide it in a Sphinx comment pending the actual 3.28.4 release. Issue: #25266 | 15 February 2024, 18:16:15 UTC |
| c8da7b0 | Brad King | 15 February 2024, 17:54:55 UTC | Merge topic 'cxxmodules-collapse-source-paths' into release-3.28 0f0d853de7 cmDyndepCollation: collapse full path before looking up Acked-by: Kitware Robot <kwrobot@kitware.com> Merge-request: !9249 | 15 February 2024, 17:55:28 UTC |
| edb833d | Brad King | 15 February 2024, 14:36:49 UTC | Tests/CommandLength: Refactor test commands to reduce test run time Since 2024-02-09 or so this test frequently times out on a few of our Windows CI runners. Use fewer but longer command lines to achieve a similar total command length. By running fewer commands, the test takes much less time. | 15 February 2024, 14:40:13 UTC |
| a7424b6 | Ben Boeckel | 14 February 2024, 18:42:18 UTC | Ninja: make the collator rule use `restat = 1` This avoids rebuilds when the module metadata does not change. | 15 February 2024, 04:01:25 UTC |
| e24eecf | Ben Boeckel | 15 February 2024, 03:42:09 UTC | Tests/CXXModules: add a test to ensure that `restat` works for collation Also update the `modules-json-file` test to account for the fact that `CXXModules.json` is not updated for timestamp-only rebuild triggers. See: #25511 | 15 February 2024, 04:01:25 UTC |
| 60a8736 | Ben Boeckel | 13 February 2024, 17:17:15 UTC | cmNinjaTargetGenerator: scanning depends on the module metadata Fixes: #25531 | 15 February 2024, 03:36:37 UTC |
| 52036ce | Ben Boeckel | 14 February 2024, 22:19:42 UTC | Tests/CXXModules: test that objects depend on dependent modules json files | 15 February 2024, 03:36:37 UTC |
| 9a45c9f | Ben Boeckel | 13 February 2024, 17:16:59 UTC | cmNinjaTargetGenerator: use `emplace_back` for scanning deps | 15 February 2024, 01:38:06 UTC |
| 363300a | Ben Boeckel | 13 February 2024, 19:32:34 UTC | cxxmodules: depend on the modmap contents If the modmap changes, we need to recompile. It is not just a file that needs to exist to compile. Fixes: #25511 | 15 February 2024, 01:36:47 UTC |
| 940628d | Ben Boeckel | 14 February 2024, 19:20:57 UTC | Tests/CXXModules: test that objects depend on their modmap files | 15 February 2024, 01:36:47 UTC |
| 3f8a59a | Ben Boeckel | 15 February 2024, 01:25:53 UTC | cxxmodules: return failure from the collator when private usage is found | 15 February 2024, 01:26:59 UTC |
| 681ee92 | Ben Boeckel | 14 February 2024, 19:20:42 UTC | Tests/CXXModules: support testing rebuild conditions | 14 February 2024, 19:34:07 UTC |
| 0f0d853 | Ben Boeckel | 13 February 2024, 19:52:00 UTC | cmDyndepCollation: collapse full path before looking up `cmSourceFile::GetFullPath()` performs a `CollapseFullPath` before storing the path. Match this behavior when looking up paths from the source file set constructions. Fixes: #25614 | 14 February 2024, 16:51:05 UTC |
| 52b1bb0 | Brad King | 12 February 2024, 15:46:13 UTC | Merge topic 'backport-3.28-visionOS' into release-3.28 22eb1b6d67 Tests: Enable visionOS tests with Xcode 15.2 f6fa34839c Tests: Update RunCMake.XcodeProject-Device for visionOS in Xcode 15.1 beta 40be88a92e visionOS: Update deployment-target flag for Xcode 15.1 beta 3ece4553c2 cmLocalGenerator: Avoid unnecessary lookup without a deployment target Acked-by: Kitware Robot <kwrobot@kitware.com> Merge-request: !9239 | 12 February 2024, 15:46:45 UTC |
| 22eb1b6 | Brad King | 08 February 2024, 16:39:57 UTC | Tests: Enable visionOS tests with Xcode 15.2 Xcode 15.2 officially includes the visionOS SDKs. Fixes: #25266 | 09 February 2024, 16:32:33 UTC |
| f6fa348 | Brad King | 09 February 2024, 16:20:28 UTC | Tests: Update RunCMake.XcodeProject-Device for visionOS in Xcode 15.1 beta The target identification macros changed since the 15.0 beta. | 09 February 2024, 16:32:23 UTC |
| 40be88a | Brad King | 09 February 2024, 00:50:41 UTC | visionOS: Update deployment-target flag for Xcode 15.1 beta Use the same flags Xcode adds for `XROS_DEPLOYMENT_TARGET`. They are `-target arm64-apple-xros1.0` and `-target arm64-apple-xros1.0-simulator`, where `1.0` is the deployment target version. Fixes: #25188 | 09 February 2024, 16:32:13 UTC |
| 3ece455 | Brad King | 09 February 2024, 14:54:00 UTC | cmLocalGenerator: Avoid unnecessary lookup without a deployment target | 09 February 2024, 16:32:13 UTC |
| 9e80593 | Brad King | 09 February 2024, 14:01:18 UTC | Merge topic 'ci-xcode-15.2' into release-3.28 68157fc57d gitlab-ci: update macOS jobs to use Xcode 15.2 Acked-by: Kitware Robot <kwrobot@kitware.com> Merge-request: !9237 | 09 February 2024, 14:01:18 UTC |
| 1c05251 | Brad King | 09 February 2024, 14:00:25 UTC | Merge topic 'FindCUDAToolkit-cudart-deps' into release-3.28 4a3cafec4f FindCUDAToolkit: Fix cudart dependencies Acked-by: Kitware Robot <kwrobot@kitware.com> Merge-request: !9236 | 09 February 2024, 14:00:46 UTC |
| 938a286 | Brad King | 09 February 2024, 13:58:41 UTC | Merge topic 'libuv-1.48' into release-3.28 077a1d5769 libuv: win/spawn: optionally run executable paths with no file extension f02ac51150 libuv: Revert "win/spawn: run executables with no file extension" 58a271e60b cmUVProcessChain: Implement no-extension-on-Windows support with libuv 1.48 Acked-by: Kitware Robot <kwrobot@kitware.com> Merge-request: !9238 | 09 February 2024, 13:59:04 UTC |
| 077a1d5 | Brad King | 08 February 2024, 21:49:18 UTC | libuv: win/spawn: optionally run executable paths with no file extension Backport libuv commit `3f7191e5` (win/spawn: optionally run executable paths with no file extension, 2024-02-05, v1.48.0~8) to add the `UV_PROCESS_WINDOWS_FILE_PATH_EXACT_NAME` option we now use. Issue: #25450 | 08 February 2024, 22:00:47 UTC |
| f02ac51 | Brad King | 08 February 2024, 21:47:06 UTC | libuv: Revert "win/spawn: run executables with no file extension" This reverts commit fc6231bee5 (libuv: win/spawn: run executables with no file extension, 2023-12-04, v3.28.0~9^2) to make room for the corresponding fix from upstream libuv 1.48. Record this as a merge from the last-imported upstream libuv snapshot branch so that future `git blame` points to the upstream for the original code instead of this commit. Issue: #25450 | 08 February 2024, 21:47:06 UTC |
| 58a271e | Brad King | 08 February 2024, 21:35:54 UTC | cmUVProcessChain: Implement no-extension-on-Windows support with libuv 1.48 Use the new `UV_PROCESS_WINDOWS_FILE_PATH_EXACT_NAME` option when building against libuv 1.48 or higher. | 08 February 2024, 21:46:53 UTC |
| 68157fc | Brad King | 08 February 2024, 16:58:21 UTC | gitlab-ci: update macOS jobs to use Xcode 15.2 | 08 February 2024, 21:28:16 UTC |
| 4a3cafe | Robert Maynard | 08 February 2024, 17:39:27 UTC | FindCUDAToolkit: Fix cudart dependencies Fixes: #25665 | 08 February 2024, 17:57:13 UTC |
| f4b5152 | Brad King | 08 February 2024, 14:39:32 UTC | Merge topic 'findmatlab-registry-lookup-without-versionxml' into release-3.28 e22c4dfea0 FindMatlab: Revert undocumented and now-unnecessary full version lookup Acked-by: Kitware Robot <kwrobot@kitware.com> Merge-request: !9226 | 08 February 2024, 14:40:05 UTC |
| e22c4df | Hermann von Kleist | 24 January 2024, 07:41:06 UTC | FindMatlab: Revert undocumented and now-unnecessary full version lookup Partially revert commit dc9d9589e4 (FindMatlab:WIN32: return full Matlab version when found via registry, 2023-09-14, v3.28.0-rc1~82^2~2). It changed `matlab_extract_all_installed_versions_from_registry` behavior to query `VersionInfo.xml` for version information. However, the function documentation states that it only uses the registry. It's also not needed since commit c608adc236 (FindMatlab: Fix major.minor version lookups in Windows Registry, 2024-01-11, v3.28.2~13^2~1), which added the same lookup to `matlab_get_all_valid_matlab_roots_from_registry`. | 07 February 2024, 17:12:12 UTC |
| a98f363 | Brad King | 07 February 2024, 15:34:51 UTC | Merge topic 'clang-cl-showIncludes' into release-3.28 0d933fc0d9 Ninja: Update showIncludes prefix detection for clang-cl 18 a88acb0a41 Tests/RunCMake/Ninja: Rename showIncludes case for clang-cl <= 17 Acked-by: Kitware Robot <kwrobot@kitware.com> Merge-request: !9229 | 07 February 2024, 15:35:32 UTC |
| 0d933fc | Martin Storsjö | 06 February 2024, 21:58:44 UTC | Ninja: Update showIncludes prefix detection for clang-cl 18 Since commit LLVM/Clang commit `5523fefb01c2` ([clang][lex] Use preferred path separator in includer-relative lookup, 2023-09-08), part of the upcoming 18.x release, the output format of the showIncludes flag has changed, where it now prints paths with double backslashes: Note: including file: .\\foo.h Previously, we expected to see the path name in the form "./foo.h". Extend the regex to match a path name starting with `.\`, in addition to the existing matched patterns. | 07 February 2024, 14:38:14 UTC |
| a88acb0 | Brad King | 07 February 2024, 14:29:22 UTC | Tests/RunCMake/Ninja: Rename showIncludes case for clang-cl <= 17 Make room for a new case for clang-cl 18. | 07 February 2024, 14:31:42 UTC |
| 5e984bb | Brad King | 05 February 2024, 15:28:25 UTC | CMake 3.28.3 | 05 February 2024, 15:28:25 UTC |
| 292d808 | Brad King | 05 February 2024, 14:56:43 UTC | Merge topic 'restore-pch-with-unity' into release-3.28 30829da506 Unity: Clarify source comments on unity build transformation 87bf1c6c33 Merge branch 'unity-after-compile-features' into restore-pch-with-unity df08c37a42 cmGlobalGenerator: Add unity/pch sources after computing compile features 004c3c3986 Tests: Add case covering PCH in a unity build Acked-by: Kitware Robot <kwrobot@kitware.com> Tested-by: buildbot <buildbot@kitware.com> Acked-by: Ben Boeckel <ben.boeckel@kitware.com> Merge-request: !9215 | 05 February 2024, 14:57:17 UTC |
| 30829da | Brad King | 02 February 2024, 12:59:21 UTC | Unity: Clarify source comments on unity build transformation | 02 February 2024, 14:43:27 UTC |
| 87bf1c6 | Brad King | 02 February 2024, 13:23:31 UTC | Merge branch 'unity-after-compile-features' into restore-pch-with-unity Replace commit 76b5383123 (cmGlobalGenerator: add unity sources after computing target compile features, 2024-01-01, v3.28.2~17^2~1) with an alternative change. This merge commit resolves conflicts such that future `git blame` calls will follow the second parent for the relevant lines and pretend the replaced commit never existed. Fixes: #25650 | 02 February 2024, 14:43:15 UTC |
| df08c37 | Brad King | 02 February 2024, 12:50:49 UTC | cmGlobalGenerator: Add unity/pch sources after computing compile features Sources that will be scanned for C++ module dependencies need to be excluded from unity builds. We need to compute compile features in order to know which sources will be scanned. Unity build and PCH sources can be added afterward without changing the compile features. This re-implements commit 76b5383123 (cmGlobalGenerator: add unity sources after computing target compile features, 2024-01-01, v3.28.2~17^2~1) using a simpler approach that also preserves support for PCH with Unity builds. Issue: #25650 Co-authored-by: Ben Boeckel <ben.boeckel@kitware.com> | 02 February 2024, 14:39:13 UTC |
| 004c3c3 | Brad King | 02 February 2024, 13:13:22 UTC | Tests: Add case covering PCH in a unity build Issue: #25650 | 02 February 2024, 13:14:21 UTC |
| 817e309 | Brad King | 31 January 2024, 13:59:46 UTC | Merge topic 'update-zlib' into release-3.28 22a52526e2 Merge branch 'upstream-zlib' into update-zlib 6c2bed54f6 zlib 2022-10-12 (04f42cec) e868b19ed1 zlib: Fix 1.2.13 version and date notice Acked-by: Kitware Robot <kwrobot@kitware.com> Merge-request: !9204 | 31 January 2024, 14:00:01 UTC |
| 22a5252 | Brad King | 30 January 2024, 15:15:31 UTC | Merge branch 'upstream-zlib' into update-zlib # By zlib upstream * upstream-zlib: zlib 2022-10-12 (04f42cec) | 30 January 2024, 15:15:31 UTC |
| 6c2bed5 | zlib upstream | 13 October 2022, 05:06:55 UTC | zlib 2022-10-12 (04f42cec) Code extracted from: https://github.com/madler/zlib.git at commit 04f42ceca40f73e2978b50e93806c2a18c1281fc (v1.2.13). | 30 January 2024, 15:15:30 UTC |
| e868b19 | Brad King | 30 January 2024, 15:13:56 UTC | zlib: Fix 1.2.13 version and date notice Issue: #25576 | 30 January 2024, 15:14:46 UTC |
| 1f25aa1 | Brad King | 29 January 2024, 20:01:32 UTC | CMake 3.28.2 | 29 January 2024, 20:01:32 UTC |
| a8b6d36 | Brad King | 29 January 2024, 15:43:29 UTC | Merge topic 'autogen-makefile-depfile' into release-3.28 d0bedb2170 Autogen: Forward dependencies when both Makefile and DEPFILE are used 4deb9c41b8 cmQtAutoGenInitializer: Improve Const-correctness Acked-by: Kitware Robot <kwrobot@kitware.com> Merge-request: !9190 | 29 January 2024, 15:43:43 UTC |
| d0bedb2 | Orkun Tokdemir | 26 January 2024, 11:17:14 UTC | Autogen: Forward dependencies when both Makefile and DEPFILE are used Since commit ebc9e448b3 (Autogen: Add depfile support for Makefiles, 2023-09-07, v3.28.0-rc1~101^2~1) CMake does not generate the correct dependency graph when both `Makefile` and `DEPFILE` are used. The build of `<target_name>_autogen_timestamp_deps` fails due to missing dependencies. To tackle that problem, forward target dependencies to both `<target_name>_autogen_timestamp_deps` and `<target_name>_autogen` instead of just `<target_name>_autogen`. Fixes: #25600 | 28 January 2024, 14:40:47 UTC |
| 6c3ab39 | Brad King | 26 January 2024, 15:12:05 UTC | Merge topic 'revert-FindFreetype-use-config' into release-3.28 b5725ae619 FindFreetype: Revert use of upstream freetype cmake package Acked-by: Kitware Robot <kwrobot@kitware.com> Reviewed-by: Ben Boeckel <ben.boeckel@kitware.com> Merge-request: !9187 | 26 January 2024, 15:12:11 UTC |
| f6c9da9 | Brad King | 26 January 2024, 15:11:01 UTC | Merge topic 'find-matlab-docs-mcr-version-mapping' into release-3.28 29fece7ef3 FindMatlab: Document that version mappings do not apply to MCR Acked-by: Kitware Robot <kwrobot@kitware.com> Merge-request: !9184 | 26 January 2024, 15:11:08 UTC |
| 0e39f34 | Brad King | 26 January 2024, 15:09:42 UTC | Merge topic 'find-matlab-fix-nonzero-patch-version' into release-3.28 b1e27b1b9d FindMatlab: Some versions use major.minor.patch in the registry Acked-by: Kitware Robot <kwrobot@kitware.com> Merge-request: !9183 | 26 January 2024, 15:10:01 UTC |
| 4deb9c4 | Orkun Tokdemir | 16 January 2024, 14:53:26 UTC | cmQtAutoGenInitializer: Improve Const-correctness | 26 January 2024, 14:50:35 UTC |
| 0bce22e | Brad King | 25 January 2024, 19:59:16 UTC | Merge topic 'tests-c99' into release-3.28 a10d2ad397 Tests: Improve C compatibility of LoadCommand tests Acked-by: Kitware Robot <kwrobot@kitware.com> Acked-by: buildbot <buildbot@kitware.com> Merge-request: !9186 | 25 January 2024, 19:59:26 UTC |
| b5725ae | Brad King | 25 January 2024, 19:45:49 UTC | FindFreetype: Revert use of upstream freetype cmake package The upstream freetype cmake package uses generator expressions and imported target to express its dependencies. That's fine on its own. However, our use of it since commit d83d925045 (FindFreetype: use `freetype-config.cmake` if available, 2023-09-06, v3.28.0-rc1~130^2) causes the `FREETYPE_LIBRARIES` result variable to reference those imported targets and create a package-level dependency for clients that did not exist previously. Revert that change for now, along with its follow-up fixes. Further investigation will be needed to solve the motivating use case another way. Fixes: #25635 | 25 January 2024, 19:45:49 UTC |
| a10d2ad | Florian Weimer | 23 November 2023, 07:45:05 UTC | Tests: Improve C compatibility of LoadCommand tests The previous code used `char **` and `const char **`` types as if they were the same. But they are distinct types in C, so when passing these pointers as function arguments, their types have to match. Future C compilers will treat this as an error, similar to what C++ compilers do today. | 25 January 2024, 19:35:01 UTC |
