Bug 63981
| Summary: | new-run-webkit-tests does not correctly identify timeouts on Mac | ||
|---|---|---|---|
| Product: | WebKit | Reporter: | Eric Seidel (no email) <eric> |
| Component: | Tools / Tests | Assignee: | Nobody <webkit-unassigned> |
| Status: | RESOLVED FIXED | ||
| Severity: | Normal | CC: | abarth, dpranke, ojan |
| Priority: | P2 | ||
| Version: | 528+ (Nightly build) | ||
| Hardware: | Unspecified | ||
| OS: | Unspecified | ||
| Bug Depends on: | 70435 | ||
| Bug Blocks: | 64491 | ||
Eric Seidel (no email)
new-run-webkit-tests does not correctly identify timeouts on Mac
Example:
2011-07-06 01:55:41,276 43149 printing.py:470 INFO 10 slowest tests that are not marked as SLOW and did not timeout/crash:
2011-07-06 01:55:41,276 43149 printing.py:470 INFO fast/workers/storage/interrupt-database.html took 30.0 seconds
2011-07-06 01:55:41,276 43149 printing.py:470 INFO fast/workers/worker-close-more.html took 30.0 seconds
2011-07-06 01:55:41,276 43149 printing.py:470 INFO fast/workers/shared-worker-frame-lifecycle.html took 30.0 seconds
2011-07-06 01:55:41,276 43149 printing.py:470 INFO fast/workers/shared-worker-lifecycle.html took 30.0 seconds
2011-07-06 01:55:41,276 43149 printing.py:470 INFO fast/workers/worker-lifecycle.html took 30.0 seconds
What's happening is that we're not reading stderr correctly in ServerProcess, so we never see the timeout notification.
At least I think that's what's happening.
I believe this to be very related to bug 63683. Which is about listening for #CRASHED correctly on stderr.
| Attachments | ||
|---|---|---|
| Add attachment proposed patch, testcase, etc. |
Eric Seidel (no email)
I'm not sure my original theory was correct. I don't see any #TIMEOUT notifications over stderr.
However, once bug 70435 lands, this would be very easy to fix were there #TIMEOUT notifications in stderr.
Eric Seidel (no email)
I'm moving this to the polish bug, since it's clearly not stopped us from switching the mac bots. I'm not sure this is still an issue.
Dirk Pranke
I don't think this is an issue any more; closing. Please reopen (or file a new bug) if I'm mistaken.