Bug 49395
| Summary: | EWS bots can't understand when they can't update | ||
|---|---|---|---|
| Product: | WebKit | Reporter: | Eric Seidel (no email) <eric> |
| Component: | Tools / Tests | Assignee: | Nobody <webkit-unassigned> |
| Status: | NEW | ||
| Severity: | Normal | CC: | abarth, aroben |
| Priority: | P2 | ||
| Version: | 528+ (Nightly build) | ||
| Hardware: | PC | ||
| OS: | OS X 10.5 | ||
| Bug Depends on: | 46636 | ||
| Bug Blocks: | |||
Eric Seidel (no email)
EWS bots can't understand when they can't update
It turns out that if you get your svn checkout wedged (as often happens on windows), then EWS bots can get into a state where they can't update and can't build, but don't report this.
The way we do an update is part of the _build call:
http://trac.webkit.org/browser/trunk/WebKitTools/Scripts/webkitpy/tool/commands/earlywarningsystem.py#L60
But! we can't distinguish between a failed first build and a failed update. The update will fail, causing the build to fail, and then if the checkout happens to be bad enough that a no-update build fails on it too, then we just assume the tree is hosed and do nothing.
The way I've fixed this for the win-ews for now is to add a:
svn status | grep ? | awk '{print $2}' | xargs rm
step to the wrapper script. (To emulate git clean -f)
The right fix is to separate out the update step so we can detect when it goes wrong. I think the commit-queue needs a separate update step too.
| Attachments | ||
|---|---|---|
| Add attachment proposed patch, testcase, etc. |
Eric Seidel (no email)
This is very similar to bug 46636.