Repository navigation
docs: the iPad's WDA stalls are testmanagerd's, whichever way WDA starts - #136
Merged
Merged
Conversation
An A/B on an iPad mini 5 (iPadOS 26.7.3), same build, alternating the background launch and an XCTest-held WDA: 0 stalls in 66 min against 2 in 67 min (with the earlier runs, 2 in ~91 min against 2 in ~82 min). During a stall testmanagerd stops logging until WDA is restarted. D18 and regression case #103 record it. 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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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 and why
D18 (the daemon always launches WDA in the background, #125) left one question open: are the iPad's WDA stalls (#103) caused by the background launch? An A/B on the iPad answers it, so D18 and regression case #103 now say so. Documentation only.
Setup: iPad mini 5 (iPad11,1), iPadOS 26.7.3, Debian Linux. One build (
main8d3e69d plus a temporary local switch, not pushed, that starts WDA with an XCTest session instead). The two launch methods alternate in four 30-minute blocks, each running the same regression script (≈140 tool calls per pass, 4 passes per block). Stalls are counted from the daemon log ("WDA 卡死"). The phone's syslog for WebDriverAgent/testmanagerd was captured at the same time.At each stall testmanagerd stops logging (normally ~2,200 lines per 10 s, up to ~5,800) until the daemon restarts WDA. So the stall is in testmanagerd, not in how WDA is started; D18's decision stands.
Tests
1. Unit tests
2. Real-device regression
main8d3e69d: 131/132, the one failure a test-script check).3. Regression cases
4. Real-device proof
Checklist
🤖 Generated with Claude Code