Summary: | DumpRenderTree should clear a WebHTMLView static variable between tests | ||||||||
---|---|---|---|---|---|---|---|---|---|
Product: | WebKit | Reporter: | Simon Fraser (smfr) <simon.fraser> | ||||||
Component: | Tools / Tests | Assignee: | Simon Fraser (smfr) <simon.fraser> | ||||||
Status: | NEW --- | ||||||||
Severity: | Normal | CC: | ap, jbedard, simon.fraser, webkit-bug-importer | ||||||
Priority: | P2 | Keywords: | InRadar | ||||||
Version: | WebKit Nightly Build | ||||||||
Hardware: | Unspecified | ||||||||
OS: | Unspecified | ||||||||
Attachments: |
|
Description
Simon Fraser (smfr)
2021-09-16 09:55:34 PDT
Created attachment 438358 [details]
Patch
I did see behavior differences when debugging the failures in 230295 related to this static, but I don't know if this fixes any WK1 flakiness. Created attachment 438360 [details]
Patch
Comment on attachment 438360 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=438360&action=review > Source/WebKitLegacy/mac/ChangeLog:8 > + Add testing-only SPI to clear the 'lastHitView' static in WebHTMLView. We navigate to about:blank between tests; should this not clear lastHitView? It does not. Should it though? That seems like it may be a better fix than a test-only SPI. I don't want to change WebKitLegacy behavior; that's all risk and little reward. I do not object to landing this, but don't feel excited enough to r+ given that isn't known to improve anything, and fixes the issue in a "wrong" way. |