fix(firefox): fix test timeouts after #41224#41336
Conversation
limit when `Page.navigationCommitted` should be treated as a `Page.sameDocumentNavigation` to only if the `url` and `frameId` match a known active navigation without these, we end up rerouting more than intended to handle the original process-swap scenario
Test results for "MCP"2 failed 7352 passed, 1122 skipped Merge workflow run. |
Test results for "tests 1"4 flaky39599 passed, 743 skipped Merge workflow run. |
Test results for "tests others"1 fatal errors, not part of any test 1 flaky20001 passed, 647 skipped Merge workflow run. |
Test results for "tests 2"1 failed 23 flaky93612 passed, 4115 skipped Merge workflow run. |
dgozman
left a comment
There was a problem hiding this comment.
I don't like what we are doing here. What's the symptom we are trying to fix? Perhaps we should make launchPersistentContext() wait for all pages to be restored? Or make sure that our navigation issued through Juggler has a priority over anything else going in the browser? I think we are trying to fix a symptom, and not with an elegant solution.
limit when
Page.navigationCommittedshould be treated as aPage.sameDocumentNavigationto only if theurlandframeIdmatch a known active navigationwithout these, we end up rerouting more than intended to handle the original process-swap scenario
should fix #41224 (comment)