Summary: | [css-writing-modes] New failures on WPT ref-tests after updating WPT import | ||||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Product: | WebKit | Reporter: | Carlos Alberto Lopez Perez <clopez> | ||||||||||||
Component: | CSS | Assignee: | Nobody <webkit-unassigned> | ||||||||||||
Status: | RESOLVED CONFIGURATION CHANGED | ||||||||||||||
Severity: | Normal | CC: | ahmad.saleem792, bfulgham, cdumez, changseok, cmarcelo, dino, esprehn+autocc, ews-watchlist, fmalita, glenn, gyuyoung.kim, kangil.han, koivisto, kondapallykalyan, lmoura, macpherson, menard, pdr, rbuis, sabouhallawa, schenney, sergio, 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=209080 https://bugs.webkit.org/show_bug.cgi?id=214278 https://bugs.webkit.org/show_bug.cgi?id=236481 |
||||||||||||||
Attachments: |
|
Description
Carlos Alberto Lopez Perez
2020-07-13 19:40:00 PDT
On top of that we have: * imported/w3c/web-platform-tests/css/css-writing-modes/normal-flow-overconstrained-vlr-005.xht was failing before for all platforms, now only seems to fail for iOS * imported/w3c/web-platform-tests/css/css-writing-modes/sizing-orthog-htb-in-vrl-013.xht was failing before for all platforms, now only seems to fail for Mac-wk1 And for platform glib (GTK/WPE): imported/w3c/web-platform-tests/css/css-writing-modes/available-size-004.html [ Pass ] imported/w3c/web-platform-tests/css/css-writing-modes/available-size-006.html [ Pass ] imported/w3c/web-platform-tests/css/css-writing-modes/available-size-008.html [ Pass ] imported/w3c/web-platform-tests/css/css-writing-modes/available-size-009.html [ Pass ] imported/w3c/web-platform-tests/css/css-writing-modes/available-size-013.html [ Pass ] imported/w3c/web-platform-tests/css/css-writing-modes/available-size-015.html [ Pass ] imported/w3c/web-platform-tests/css/css-writing-modes/available-size-016.html [ Pass ] imported/w3c/web-platform-tests/css/css-writing-modes/available-size-018.html [ Pass ] imported/w3c/web-platform-tests/css/css-writing-modes/available-size-019.html [ Pass ] imported/w3c/web-platform-tests/css/css-writing-modes/baseline-with-orthogonal-flow-001.html [ Pass ] imported/w3c/web-platform-tests/css/css-writing-modes/inline-box-border-vlr-001.html [ Pass ] For WPE: imported/w3c/web-platform-tests/css/css-writing-modes/normal-flow-overconstrained-vrl-004.xht [ Pass ] (In reply to Carlos Alberto Lopez Perez from comment #1) snip. > imported/w3c/web-platform-tests/css/css-writing-modes/available-size-013. > html [ Pass ] Started image failing somewhere between r265018 (last good, WPE) and r265024 (first bad, both ports). Maybe related to r265020? Created attachment 427508 [details]
Patch
Created attachment 427522 [details]
Patch
Created attachment 427524 [details]
Patch
Created attachment 427527 [details]
Patch
Created attachment 427532 [details]
Patch
Comment on attachment 427532 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=427532&action=review > Source/WebCore/dom/Element.cpp:3502 > + const RenderStyle* style = nullptr; > + if (allowRenderStyle == AllowRenderStyle::No && document().documentElement() == this) { > + if (hasRareData()) > + style = elementRareData()->computedStyle(); > + if (!style) > + setNodeFlag(NodeFlag::IsComputedStyleInvalidFlag); > + } else > + style = existingComputedStyle(); It is deeply wrong to insert this sort of hack into our core code to paper over a minor test failure. Pretty sure the actual issue here is not properly understood. Safari Technology Preview 170 is passing all three tests, which this patch was progressing, do we need this anymore? https://wpt.fyi/results/css/css-writing-modes?label=master&label=experimental&aligned=&q=css-writing-modes%2F Closing based on Ahmad's testing. |