[qt/qtqa/gerrit]: 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/qtqa/gerrit
Pushed by mirror-service into branch 'upstream/master'.
Changed from 82ad298d5df25952d78534ce57d3281a2b7a8456 to e84c5628a547f80610d4495536387c534fb30c49
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 2226895bad7d7ecdbdab82af857b09a17416bc6c by Sarah Silva on 13/08/2026 at 22:27..
Support line range selection in diff view
Add support for navigating to and highlighting multi-line ranges in the
diff view.
- Parse line range hashes (e.g. #50-156 or #b50-156) in gr-router.
- Store endLineNum in diffView state and pass it to gr-diff-cursor.
- Render a continuous multi-line focus outline box around the selected
section using target-range-row styles rather than synthetic DOM text
selection.
Release-Notes: Support line range selection in diff view
Google-Bug-Id: b/539995068
TAG=agy
CONV=82ab189a-2e5d-4915-94c2-824425982475
Change-Id: Ibb77971a266e7f243d1698f81ddb19a9c5be0f59
https://invent.kde.org/qt/qtqa/gerrit/-/commit/2226895bad7d7ecdbdab82af857b09a17416bc6c
Git commit 6f4ca22988af67d402fdb3e67f16c51390b912f5 by Milutin Kristofic on 14/08/2026 at 16:29..
Cancel endpoint properties init timeout on disconnect
When navigating between pages (e.g. from Change view to Settings),
GrEndpointDecorator elements get detached while waiting for properties.
Because the 10s initialization timeout wasn't cleared on disconnect,
it triggered false-alarm timeout errors.
Track timeout IDs and clear them in disconnectedCallback, and avoid
attaching custom components if the decorator was disconnected before
initialization completed.
Release-Notes: skip
Change-Id: Ic155dd5e8ccedd519beccd8e6cd766d5940b6e87
https://invent.kde.org/qt/qtqa/gerrit/-/commit/6f4ca22988af67d402fdb3e67f16c51390b912f5
Git commit 24753f688a0dfc1e751de360137d0c87153a6b02 by Milutin Kristofic on 14/08/2026 at 17:58..
Fix visual alignment in gr-change-list-reviewer-flow
The reviewer flow dialog had two alignment issues: 1. The confirmation
dialogs for large groups (confirm-reviewer and
confirm-cc) were rendered directly inside the CSS grid (.grid)
alongside the reviewer and CC rows. Since .grid is configured with
two columns (min-content 1fr), these dialog elements became direct
children of the grid, shifting the CC label and input into incorrect
grid cells.
2. In the CSS grid (.grid), align-items defaulted to stretch/normal,
which placed label text (<span>Reviewers</span> and <span>CC</span>)
at the top of the grid track, causing vertical misalignment with
the centered input text fields and account chips.
Fix both issues by:
- Moving confirmation dialogs outside of the CSS grid container.
- Adding align-items: center to .grid so labels vertically align with
input chips and autocomplete placeholders.
Also adds visual regression screenshot tests for
gr-change-list-reviewer-flow under both light and dark themes with
updated golden baselines.
Google-Bug-Id: b/544715082
Release-Notes: skip
Change-Id: Ieec19902d44b5c830cc221f5a3cda59e96138752
https://invent.kde.org/qt/qtqa/gerrit/-/commit/24753f688a0dfc1e751de360137d0c87153a6b02
Git commit f4e6515c3dbfef67c96e28036c928fc1e762f160 by Gerrit Code Review (on behalf of Milutin Kristofic) on 14/08/2026 at 18:42..
Merge "Cancel endpoint properties init timeout on disconnect"
https://invent.kde.org/qt/qtqa/gerrit/-/commit/f4e6515c3dbfef67c96e28036c928fc1e762f160
Git commit 963e5f39fcc54f088fc16a75ea5dea4e10a01c64 by Gerrit Code Review (on behalf of Sarah Silva) on 14/08/2026 at 19:18..
Merge "Support line range selection in diff view"
https://invent.kde.org/qt/qtqa/gerrit/-/commit/963e5f39fcc54f088fc16a75ea5dea4e10a01c64
Git commit e84c5628a547f80610d4495536387c534fb30c49 by Gerrit Code Review (on behalf of Kaushik Lingarkar) on 14/08/2026 at 20:44..
Update git submodules
* Update plugins/replication from branch 'master'
to 5c3414d99c36c293e2119191ae9bd0aca5b46380
- ProjectRepairer: Add a --copy-loose-objects repair action
Add a COPY_LOOSE_OBJECTS action, exposed as '--copy-loose-objects',
rsyncing the two hex character fanout directories under objects/. The
include patterns keep rsync out of pack/, info/ and their siblings.
The action is declared ahead of COPY_PACKS, so when both are selected
loose objects are copied first. '--full' now covers both.
Change-Id: I9c3d12182bdc5fa9da662269d8681db174057f21
- ProjectRepairer: Refactor to select actions through an enum
repair() takes a 'copyPacks' boolean, which can only describe the one
thing it knows how to do. A second kind of repair would mean a second
boolean at every call site, with the order the two run in implicit in
how repair() tests them.
Take a collection of Action values instead. Individual flags are kept in
flag order. '--full' and the no-flag default use Action.all() so actions
run in enum declaration order. Action has only COPY_PACKS, so behaviour
is unchanged.
Change-Id: Ie7c0d9af40cab893a01a754bc984b8593ff6a5bc
- ProjectRepairer: Refactor to extract the repository lookup
copyPackTo() validates the destination URI and opens the repository to
locate it on disk before copying anything, neither of which has to do
with packs.
Move both into objectsDir(), which reports the repository's objects
directory, or nothing if the project cannot be repaired at that
destination. What is left is copyPacksTo(), taking the directory to
copy from. The now trivial copyInOrder() folds into it.
Change-Id: I2d1ba0a23671d6db115f18a35420f4c04a24c0ec
- ProjectRepairer: Refactor copy() to take a destination dir
buildCopyDestination() appends objects/pack/ to the remote path
itself, so copy() can only ever populate a destination's pack
directory.
Pass the remote subdirectory in instead, with the pack callers naming
PACK_DIR. The rsync commands produced are unchanged.
Change-Id: I7bf05966dd95b7827864f2ae5a0a39fe9ca050eb
- ProjectRepairer: Report an interrupted copy as InterruptedIOException
copy() declares InterruptedException, but the only one that escapes
comes from the halt() call in its own catch block, leaving copyInOrder()
to catch an interrupt it takes no part in. The interrupt copy() does
take part in becomes a -1 exit code, and as throwing
InterruptedException clears the interrupt status, nothing above can tell
an interrupted transfer from a failed one, or that the repair was
cancelled at all.
Absorb the halt() interrupt where it happens, and report the copy's own
interrupt as InterruptedIOException, which is what java.io raises when
an interrupt terminates a transfer. It unwinds the repair, so the copies
and destinations left after a cancel are skipped instead of running to
completion.
The SSH command lets it escape, which BaseCommand already treats as the
client having dropped off, exiting 127 without logging. Auto-repair logs
the interrupt and skips its follow-up replication.
Change-Id: Ie61e0329aaa3131c8ac1f8b4932e57e48c5e5e0b
https://invent.kde.org/qt/qtqa/gerrit/-/commit/e84c5628a547f80610d4495536387c534fb30c49