[qt/qt/qtbase]: Summary of bulk changes made
KDE Git Services - Bulk Change <[email protected]>
| Newsgroups | gmane.comp.kde.cvs |
|---|---|
| Message-ID | <[email protected]> |
Git repository change summary for qt/qt/qtbase Pushed by mirror-service into branch '6.12'. Changed from 8a383d62a212b4061b1dcc8beab420857549de5d to 6f5209567bd9df710a469a4c8ee640f5cfef4844 Acknowledgement was received that this change introduces only existing code that has been pushed to another public open source repository. This change contains the following new commits: Git commit 6d4136145a591720785e86cae493de4a9791d975 by Qt Cherry-pick Bot (on behalf of Zbigniew Chyla) on 17/07/2026 at 07:54.. ohos: resolve owning ability for sub and float windows tryGetQAbilityPeerByQWindow only matches a QWindow against each QAbilityPeer's main-window reference, so sub and float windows never resolve to their owning ability (and tryGetQAbilityByQWindow, built on it, inherits the same limitation). Fall back to the JS window registry: when no peer's main window matches, look up the JsWindowRef by its owning QWindow and resolve the peer from the ability instance id it now carries. Main windows keep matching via the existing fast path, so their result is unchanged; sub and float windows now resolve to the same ability the window proxy records. No in-tree caller passes a non-main window yet, so behavior is unchanged for now; this unblocks resolving a dialog's owning ability from any QWindow on the JS thread. Task-number: QTBUG-147194 Change-Id: I1811feed70932ae5227a3c3e4e4ca92670693c9f Reviewed-by: SanthoshKumar Selvaraj <[email protected]> (cherry picked from commit 876748257f0c6656435b1446e88068fd4d9ccfab) Reviewed-by: Qt Cherry-pick Bot <[email protected]> https://invent.kde.org/qt/qt/qtbase/-/commit/6d4136145a591720785e86cae493de4a9791d975 Git commit 8b5f2a63d25a1f40ec811ceec65eaee8a5f7bb29 by Qt Cherry-pick Bot (on behalf of Zbigniew Chyla) on 17/07/2026 at 07:54.. ohos: add QOhosJsState::tryGetJsWindowByQWindow() QOhosJsState can resolve a QWindow to its owning ability (tryGetQAbilityByQWindow) but not to its JS window.Window object. Add tryGetJsWindowByQWindow() alongside it: JsStateImpl looks the QWindow up via QOhosJsWindowRegistry::tryFindJsWindowByQWindowRef() and returns its JS window object, on the JS thread and by reference identity, for main, sub and float windows alike. This is for plugin-external code that wants to use the JS Window API but can identify a window only by its QWindow (it has no access to the plugin's JsWindowId or window registries). It has no in-tree caller yet. Task-number: QTBUG-147194 Change-Id: I7437642b41ea524c597881f928a87c686128b7e1 Reviewed-by: SanthoshKumar Selvaraj <[email protected]> (cherry picked from commit 5d70c78154cf59866af2ffd149e0b7ea720b3a92) Reviewed-by: Qt Cherry-pick Bot <[email protected]> https://invent.kde.org/qt/qt/qtbase/-/commit/8b5f2a63d25a1f40ec811ceec65eaee8a5f7bb29 Git commit f10830ca586d52a6e96fb53173a9e848af8b1174 by Qt Cherry-pick Bot (on behalf of Jie Liu) on 17/07/2026 at 08:54.. wayland: protect mSurface access in commit() with read lock Commit 0a8fe091f7 added a mSurface->commit() call in the buffer->committed() fast path, but it was placed before the QReadLocker that was introduced by commit 06906e0d11, leaving the mSurface access unprotected. If another thread (e.g. the main thread in reset()) concurrently acquires the write lock and destroys mSurface, this can lead to a null pointer dereference. Move the QReadLocker to the top of commit() so that all mSurface accesses are covered, including the buffer->committed() early return path. Change-Id: I56bd4795332d4f3d740d6bfd42f2ef3edcf0a300 Reviewed-by: David Edmundson <[email protected]> (cherry picked from commit ec3e04184443f34e3e8a29628be2bdc7b21666c0) Reviewed-by: Qt Cherry-pick Bot <[email protected]> https://invent.kde.org/qt/qt/qtbase/-/commit/f10830ca586d52a6e96fb53173a9e848af8b1174 Git commit 6f5209567bd9df710a469a4c8ee640f5cfef4844 by Qt Cherry-pick Bot (on behalf of Joerg Bornemann) on 17/07/2026 at 09:34.. CMake: Fix QML deployment when configuring through a junction _qt_internal_get_deploy_impl_dir returned the raw ${CMAKE_BINARY_DIR}/.qt, but file(GENERATE) resolves junctions and writes the per-target QML deploy scripts under the real build directory. The stored __QT_DEPLOY_IMPL_DIR kept the junction path, so qt6_deploy_qml_imports could not find them. Resolve the deploy impl dir with REALPATH so generation and lookup agree. Fixes: QTBUG-148296 Change-Id: I87ba814801d64e1d9627c12cc7c493739c3dcdf8 Reviewed-by: Alexandru Croitor <[email protected]> (cherry picked from commit 78f595689378643dc39e29b340a24039769ed08c) Reviewed-by: Qt Cherry-pick Bot <[email protected]> https://invent.kde.org/qt/qt/qtbase/-/commit/6f5209567bd9df710a469a4c8ee640f5cfef4844