Repository navigation
fix(computers): the probe tries a drag or a chord once more when a busy desk may have missed it - #671
Merged
Conversation
…sy desk may have missed it
Gate 3's P7 drag check (reference driver) and P6 chord check (cua-driver) failed now and then on busy runners, which
failed the compose and every test that needed that computer. Neither was the driver: the X11 driver moves through
each step of a drag with a pause, but a Tk desk starved through all 400 ms of it sees the motion as one event ("in 3
points": the press, one motion, the release), so its waypoint is lost; and keys sent while the click that focuses a
box still waits go wherever focus was. Each of the two now gets a second try before it fails, and a pass on its second
try says in its detail what the first found, so a driver that only sometimes manages it shows.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…he one before The drag check waited for any stroke, so on its second try the first try's stroke, still in the desk's state, answered at once and was judged again (Greptile). It now notes the desk's strokes before it drags and waits for a new one. A probe test drives a desk starved through its first drag that saves its second a second after the drag: the check passes on its second try, and failed without this. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…the drag happens test_a_drag_that_doesnt_arrive_whole_fails_p7 served a desk that held its stroke before any drag, which the check, now waiting for a stroke the drag adds, read as no drag at all (CodeBuild on #671). Its desk now records the stroke as a real one does, on the drag. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
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.
What changes
Two gate-3 probe checks failed now and then on busy runners, and a probe check that fails fails the compose, so every test that needed that computer failed with it:
website_browser_test), da06b72set_test)Neither is the driver. The X11 driver moves through each step of a drag with a pause between them (
_x11.drag_along). The Tk desk, though, records a stroke from its<B1-Motion>events, and Tk merges the motion events that queue while it's busy. A desk starved through the whole 400 ms drag sees the press, one motion and the release: "in 3 points", and the waypoint is gone. Keys sent while the click that focuses the box is still waiting go wherever focus was.The change.
Probe.checktakesattempts, and these two checks getBUSY_DESK_ATTEMPTS(2): a second try before they fail. A pass on the second try records in its detail what the first try found ("passed on try 2 of 2; before: …"), so a driver that only sometimes manages it still shows in the report. Every other check is tried once, as before.Testing
test_computer_conformance_attempts.py: a check that fails its first try and passes its second passes, naming the first try's problem; one that fails both fails with the last; any other check is tried once.test_computer_conformance.pyand the new tests: 43 passed.🤖 Generated with Claude Code


Confidence Score: 5/5
The PR appears safe to merge; no blocking issues remain.Summary
The probe retries the chord and drag checks once when a busy desk may have missed their input.
Reviews (3) · Last reviewed commit: "test: the drag check's unit test has its..." · Reviewed by Greptile