Bug 271716
| Summary: | Port.default_child_processes is totally ignored for second-and-later device_types, even when zero | ||
|---|---|---|---|
| Product: | WebKit | Reporter: | Sam Sneddon [:gsnedders] <gsnedders> |
| Component: | Tools / Tests | Assignee: | Sam Sneddon [:gsnedders] <gsnedders> |
| Status: | RESOLVED FIXED | ||
| Severity: | Normal | CC: | jbedard, ryanhaddad, webkit-bug-importer |
| Priority: | P2 | Keywords: | InRadar |
| Version: | WebKit Nightly Build | ||
| Hardware: | Unspecified | ||
| OS: | Unspecified | ||
| See Also: | https://bugs.webkit.org/show_bug.cgi?id=261175 | ||
Sam Sneddon [:gsnedders]
After bug 261175, 267745@main, due to explicitly setting `child_processes_option_value` the minimum value we set self._options.child_processes to is now zero, which means the below conditional can only ever evaluate to True on the first iteration. This is almost certainly unintended.
| Attachments | ||
|---|---|---|
| Add attachment proposed patch, testcase, etc. |
Sam Sneddon [:gsnedders]
I guess this only actually has broken anything when Port.default_child_processes() is zero and Port.max_child_processes() is not, which I think never actually happens for any existing Port.
Regardless, it adds very much a confusing edge-case in our code flow, and the comment is misleading: it has no impact on retries (we already hardcoded num_workers=1 in _run_test_subset).
Sam Sneddon [:gsnedders]
Pull request: https://github.com/WebKit/WebKit/pull/26475
Radar WebKit Bug Importer
<rdar://problem/125793341>
EWS
Committed 277755@main (0de43f2ad98a): <https://commits.webkit.org/277755@main>
Reviewed commits have been landed. Closing PR #26475 and removing active labels.