274330 2024-05-17 13:12:50 -0700 editing/execCommand layout tests intermittently failing on mac-wk1 (off-by-one baseline issue) 2024-05-17 15:53:21 -0700 1 1 1 Unclassified WebKit Tools / Tests WebKit Nightly Build Unspecified Unspecified RESOLVED CONFIGURATION CHANGED https://bugs.webkit.org/show_bug.cgi?id=235733 InRadar P2 Normal --- 1 ryanhaddad webkit-unassigned darin jbedard webkit-bot-watchers-bugzilla webkit-bug-importer oldest_to_newest 2036102 0 ryanhaddad 2024-05-17 13:12:50 -0700 Layout tests in the editing/execCommand are intermittently failing on mac-wk1. This appears to be due to the expected results being off by one test, causing all subsequent ones to fail. Example: https://build.webkit.org/results/Apple-Ventura-Release-AppleSilicon-WK1-Tests/278913@main%20(9468)/results.html 2036103 1 webkit-bug-importer 2024-05-17 13:13:01 -0700 <rdar://problem/128290437> 2036111 2 jbedard 2024-05-17 13:33:28 -0700 This is absolutely an off-by-1 issue. Unlike last time, this isn't about ref tests, it's about normal test diffs, the bug almost certainly lies in DumpRenderTree. I have a hunch it's also specific to WebKitLegacy, but we don't have concrete data about that. 2036121 3 ryanhaddad 2024-05-17 14:18:28 -0700 It appears that this particular case was already known and the offending tests had been skipped. They were accidentally unskipped as part of 278894@main https://github.com/WebKit/WebKit/commit/f606fe226b6f65e34820bdb8b17cc2fab81a14ec#diff-a0da6d7c099dd9e2d22a140bbdf209f01f0f0394d19609d6d3e7c60a85ad760dL2330-L2332 2036124 4 darin 2024-05-17 14:28:33 -0700 Can we immediately revert those incorrect changes? 2036150 5 ryanhaddad 2024-05-17 15:53:21 -0700 The incorrect changes were un-done in https://commits.webkit.org/278925@main