205757 2020-01-03 15:57:53 -0800 REGRESSION: [ Mac ] webgl/1.0.3/conformance/textures/texture-upload-size.html is a flaky timeout 2020-04-28 23:26:39 -0700 1 1 1 Unclassified WebKit WebGL WebKit Nightly Build Unspecified Unspecified RESOLVED CONFIGURATION CHANGED InRadar P2 Normal --- 207006 1 tsavell webkit-unassigned dino justin_fan webkit-bot-watchers-bugzilla webkit-bug-importer oldest_to_newest 1602628 0 tsavell 2020-01-03 15:57:53 -0800 webgl/1.0.3/conformance/textures/texture-upload-size.html Description: This test is very infrequent timeout across Mac. It most frequently times out on Mac wk1 History: https://results.webkit.org/?suite=layout-tests&test=webgl%2F1.0.3%2Fconformance%2Ftextures%2Ftexture-upload-size.html 1602629 1 webkit-bug-importer 2020-01-03 15:58:11 -0800 <rdar://problem/58309347> 1602631 2 tsavell 2020-01-03 16:14:36 -0800 marked this test as timeout in: https://trac.webkit.org/changeset/254020/webkit 1602637 3 ap 2020-01-03 16:26:07 -0800 When did it regress, at least roughly? 1602648 4 tsavell 2020-01-03 16:30:20 -0800 This test has been timing out sense it was introduced in https://trac.webkit.org/changeset/252121/webkit 1613552 5 justin_fan 2020-01-31 13:22:21 -0800 Logging reveals getError() is the last method being called; it is making it down to the GraphicsContextGL call, which means the issue likely lies in the driver. Switching to ANGLE may actually fix this (since right now it does not call the ANGLE version of glGetError). Even if it is not getError, it doesn’t appear anything is hanging at the WebKit level so it is worth revisiting this after the ANGLE switch. 1646868 6 ryanhaddad 2020-04-28 23:26:39 -0700 Removed flaky expectation in r260879 since this test has been consistently passing for the past few weeks.