Repository navigation
Conversation
Record the reviewer, the review standard, and the hotspot findings for this tree.
Keep the 2026-10-22 clauses in place and name the shared standard used for later reviews.
Up to standards ✅🟢 Issues
|
…erifiers Since the X11 resources are deferred, platform_wrapper.mouse is a proxy object with no __name__, so the FreeBSD and X11 verifiers failed on a backend they had in fact selected. They now ask the proxy's module which backend it chose.
The informational classification step still caught main's PortalConsentNotGranted, which the development line's oeffis does not define, and the AttributeError ended the whole portal job.
PySide6 6.11.2 types loadFromData's format as bytes, 6.12.0 also accepts str, and the runtime takes only str: no spelling satisfied both stubs in the typing-extras job. The JPEG header identifies the image anyway.
The container image has PySide6 but no libEGL, so importing QtWidgets
fails; importorskip("PySide6") let these four modules error at
collection instead of skipping like the rest of the GUI tests.
- Fixture threads raise or record a mismatch instead of asserting inside the handler that forwards every failure to the parent test. - The Android helper declares cleartext traffic off and shrinks its release build with R8, keeping the instrumentation it is started by. - The workspace example's private value is no longer called a password. - The loopback server base, the CI tooling's own path arguments, its 0600 temp file and the seeded fixture generator are marked as reviewed, each with its reason.
A request is capped at 0.25 s, and on the Quality workflow's Ubuntu runners the observer sometimes had not polled within it, so the helper's first acquire raised 'deadline exhausted' and the test failed before reaching what it checks. A timed-out request allocates no lease, so the helper now retries it like a busy source, within five seconds.
SonarCloud flags the helper build because its dependency versions are not pinned by a lock file. Every configuration is now locked; while no gradle.lockfile exists Gradle resolves as before. A temporary step in the instrumentation job writes the lock files with --write-locks and keeps them as an artifact, so the real files can be committed and the step removed.
The checks resolved the wrapper module from the proxy's __module__ with importlib, which static analysis reads as importing arbitrary code. The wrapper is always _platform_linux, so it is imported directly and the proxy is checked to be its _Mouse.
…-checks Fix the checks the 1.1.0 release PR failed
Written by the android-helper workflow with --write-locks on the dev push of b10aece, from the configurations the build actually resolves. The root project only declares plugins and resolves nothing, so its lock file lists no dependency. The temporary step that wrote them is removed, and the build file says how to refresh them.
…per-gradle-locks Commit the Android helper's dependency lock files
…sent
The container image has PySide6 but no libEGL, so QtWidgets fails to
import. These 13 modules guarded with importorskip("PySide6"), which
succeeds there, and their tests failed with ImportError instead of
skipping like the rest of the GUI suite.
With a 10 ms budget, the poll's wait can wake just before the deadline on Windows' coarse timer and read the task status once more before the timeout fires; the test required exactly one read and failed on the release PR's Windows 3.13 runner. For the timeout case it now checks what the waiter promises: submit first, one or more status reads, a pause last.
The viewer panel connects off the GUI thread, and its clipboard authority is adopted when the connect answers. The reconnect case enqueued the new viewer's clipboard after a single processEvents(), so on a slow runner (the release PR's Windows 3.10) the connect had not answered yet and the authorised clipboard was dropped. It now settles the connect first, as the test's own _tab() helper does.
The test gave the task a 2 s wall budget and expects it to run out while the helper's fourth observation sleeps. On the release PR's Windows 3.10 runner the helper start and the first observe/act alone took longer, so the budget ran out during setup instead. The budget is now 6 s under a 20 s action limit: enough for setup, and still spent during the 30 s sleep.
…skips Skip the agent GUI tests when Qt cannot load
…out-test Fix two timing-dependent tests the release PR hit on Windows
The server drops the accepted connection and then retires its source. The test checked for the retirement as soon as the client saw the close, and on the release PR's Windows 3.13 runner it had not landed yet. It now waits up to five seconds for both the source close and the failed status.
…er-retire-wait Wait for the presence server to retire its source after a refused client
|
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Ships the development line as 1.1.0.
pyproject.tomldeclares 1.1.0, and that version has no tag yet, soscripts/stable_release.pypublishes it as written.mainback intodev. In each subsystem both lines re-implemented while beta, dev's version is kept. Everything elsemainadded since 1.0 is kept too.CHANGELOG.md(1.1.0) lists every removed facade name, command, MCP tool and environment variable, and the behaviour that now follows the development line. The beta rule is indocs/API_LIFECYCLE.md.je_auto_control.apikeeps every 1.0 name.Checks on #537: Codacy, SonarCloud, and the headless suite on Ubuntu 3.14, macOS 3.10, Windows 3.10 and Windows 3.14 all passed. This PR also runs the Quality workflow, which only runs against
main: lint, bandit, both typing-contract jobs, the free-threaded scope job and the Sphinx docs build.