You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@maka.apache.org by "liugddx (via GitHub)" <gi...@apache.org> on 2026/08/26 04:51:31 UTC
[I] test(desktop): make prompt-rail tail mounting deterministic [maka]
liugddx opened a new issue, #3862:
URL: https://github.com/apache/maka/issues/3862
## Summary
The prompt-rail eviction Desktop e2e can set the transcript to the bottom without causing the turn virtualizer to observe a scroll event. The test then waits for the exact tail turn (`turn-prompt-rail-120`), but the mounted window remains stale and the assertion times out.
## Evidence
An unrelated Windows-storage PR (#3789) failed on exact head `b98f3b329` after all build, release-contract, workspace, Eval, Runtime Host, and Windows recovery checks passed:
- [Desktop e2e failure](https://github.com/apache/maka/actions/runs/32930874507/job/98062726773)
- scenario: `evicting a turn-owned sibling interaction hands focus back to the transcript`
- expected `[data-virtual-turn-id="turn-prompt-rail-120"]`: 1
- observed for 10 seconds: 0
#3789 has no Desktop/UI diff against `main`.
## Cause boundary
`scrollTranscriptTo` only assigns `scroller.scrollTop`. The virtualized window is updated through the scroll observation path, but a programmatic assignment does not provide a deterministic event/paint barrier for the test. The same spec already has `notifyTranscriptScrolled` and `waitForPaintedFrames` helpers later in the eviction journey.
## Expected closure
- Make `scrollTranscriptTo` notify the virtualizer explicitly and wait for the resulting painted frames.
- Preserve the test's exact tail mount, later eviction, focus-return, and selection-collapse assertions.
- Run the focused prompt-rail spec and full hosted Desktop e2e suite.
Found while validating #3789; the fix must remain in a separate PR.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@maka.apache.org.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@maka.apache.org
For additional commands, e-mail: commits-help@maka.apache.org
Re: [I] test(desktop): make prompt-rail tail mounting deterministic [maka]
Posted by "liugddx (via GitHub)" <gi...@apache.org>.
liugddx commented on issue #3862:
URL: https://github.com/apache/maka/issues/3862#issuecomment-5420806654
Fix proposed in #3863. The shared prompt-rail scroll helper now explicitly notifies the virtualizer and waits two painted frames before callers assert the mounted window; all exact eviction/focus assertions remain unchanged.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@maka.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@maka.apache.org
For additional commands, e-mail: commits-help@maka.apache.org
Re: [I] test(desktop): make prompt-rail tail mounting deterministic [maka]
Posted by "M4n5ter (via GitHub)" <gi...@apache.org>.
M4n5ter closed issue #3862: test(desktop): make prompt-rail tail mounting deterministic
URL: https://github.com/apache/maka/issues/3862
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@maka.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@maka.apache.org
For additional commands, e-mail: commits-help@maka.apache.org
Re: [I] test(desktop): make prompt-rail tail mounting deterministic [maka]
Posted by "liugddx (via GitHub)" <gi...@apache.org>.
liugddx commented on issue #3862:
URL: https://github.com/apache/maka/issues/3862#issuecomment-5420859074
Fix #3863 is exact-head green in hosted CI, including the full Desktop e2e suite. Keeping this issue open until the PR receives maintainer review and merges.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@maka.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@maka.apache.org
For additional commands, e-mail: commits-help@maka.apache.org
Re: [I] test(desktop): make prompt-rail tail mounting deterministic [maka]
Posted by "Astro-Han (via GitHub)" <gi...@apache.org>.
Astro-Han commented on issue #3862:
URL: https://github.com/apache/maka/issues/3862#issuecomment-5456339456
The same tail-mount failure recurred after #3863.
PR #4090 failed on exact head `89f8a9d89e839afb51c8710e54e03486e7efca45`:
- run: https://github.com/apache/maka/actions/runs/33195433169/job/98931236633
- scenario: `evicting a turn-owned sibling interaction hands focus back to the transcript`
- expected `[data-virtual-turn-id="turn-prompt-rail-120"]`: 1
- observed for the full 10-second timeout: 0
- Desktop E2E result: 76 passed, 1 failed, 1 skipped
The current helper already assigns `scrollTop`, dispatches the scroll event, and waits two painted frames—the synchronization added by #3863—so that barrier is not sufficient in every hosted run.
The prompt-rail fixture contains only plain completed user/assistant turns and does not enter #4090's aborted-turn rendering branch. This therefore does not point to the aborted-status change, but it is an exact recurrence of the symptom tracked here.
The broader scroll-ownership work is now tracked in #4099 and implemented in #4105. Recording this run here so the original failure mode is not lost after this issue's closure.
_Posted with Codex assistance; @Astro-Han verified the linked run and diagnosis._
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@maka.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@maka.apache.org
For additional commands, e-mail: commits-help@maka.apache.org