[qt/qt/qtapplicationmanager]: 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/qtapplicationmanager Pushed by mirror-service into branch 'dev'. Changed from 64b00a7e3a854c526a938942ce71fa98fa1e46cc to d3a3ed947848c3f550856275f7aa3df514a84e9c 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 dcb174f2f71f11be3bd81e9484ccb2bd9929e779 by Robert Griebl on 07/08/2026 at 13:25.. Fix missing include Change-Id: Ib347abefb208da6ea7f948711ef3962975d65eef Pick-to: 6.12 Reviewed-by: Thomas Senyk <[email protected]> https://invent.kde.org/qt/qt/qtapplicationmanager/-/commit/dcb174f2f71f11be3bd81e9484ccb2bd9929e779 Git commit 0cc512216f4175ec9ed21b8eb47ba35f9d7e5216 by Robert Griebl on 07/08/2026 at 13:25.. Security review: bound the in-memory package header and footer buffers The header and footer entries of a package are accumulated in memory while extracting, with their size dictated by the package being read. A crafted package could drive the extracting process out of memory - including the appman itself, as the installer uses the same code. Refuse anything above 1MB, as manifests and config files already do. Also document that the packager's sign and verify commands unpack into TMPDIR, so operators processing third-party packages need to bound that filesystem: the extracted size itself is deliberately not capped, since no limit would suit every deployment. Change-Id: Ie487682fa38400fa214ef8aaf02428fd82e2da0d Fixes: QTBUG-148989 Pick-to: 6.12 Reviewed-by: Thomas Senyk <[email protected]> https://invent.kde.org/qt/qt/qtapplicationmanager/-/commit/0cc512216f4175ec9ed21b8eb47ba35f9d7e5216 Git commit 63a4f32e426e8a397530f5b2c42acfb13bb3399a by Robert Griebl on 07/08/2026 at 13:25.. Add recursion guards to the crash handlers std::terminate is entered again, if anything throws while we are reporting - e.g. from a what() implementation or an exception's destructor. The Windows exception filter can likewise be re-entered by a fault while parsing and re-throwing the MSVC exception internals, or by a second thread crashing into its static buffer. Change-Id: I2e569567e00a819b9b8bc890f82aaeb7168be417 Pick-to: 6.12 Reviewed-by: Thomas Senyk <[email protected]> https://invent.kde.org/qt/qt/qtapplicationmanager/-/commit/63a4f32e426e8a397530f5b2c42acfb13bb3399a Git commit 74d54256a344e3a94fcf99208b7434a562b61f33 by Robert Griebl on 07/08/2026 at 13:25.. Replace the linear scan over waiting intent requests with a QHash Requests that were gc'ed on the JS side before a reply arrived also only got cleaned out of the list by a timer-throttled sweep on the next successful reply, letting the list grow unboundedly. Keying by requestId and erasing via QObject::destroyed removes entries as soon as the request dies and makes the reply lookup O(1). Change-Id: Iec3d360424a901c42c42a9ed0d3ac71fafd6a708 Pick-to: 6.12 Reviewed-by: Thomas Senyk <[email protected]> https://invent.kde.org/qt/qt/qtapplicationmanager/-/commit/74d54256a344e3a94fcf99208b7434a562b61f33 Git commit d3a3ed947848c3f550856275f7aa3df514a84e9c by Robert Griebl on 07/08/2026 at 13:25.. Reject dangling backslash in debug wrapper specs parseSpecification silently consumed a trailing '\' via the synthetic null terminator. Reject it as malformed instead. Change-Id: I82a2b181abc1798e792b23532f30db6e269c1860 Pick-to: 6.12 Reviewed-by: Thomas Senyk <[email protected]> https://invent.kde.org/qt/qt/qtapplicationmanager/-/commit/d3a3ed947848c3f550856275f7aa3df514a84e9c