Make sure that culledInlineAbsoluteRect still does a translation of a 0,0 point to absolute coordinates so that the top left position is accurate. The equivalent bug for culledInlineAbsoluteQuads has already been fixed. https://bugs.webkit.org/show_bug.cgi?id=60040
<rdar://problem/9408862>
Created attachment 102444 [details] Patch
Comment on attachment 102444 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=102444&action=review > Source/WebCore/ChangeLog:8 > + This bug fixes a regression introduced in r83075. I think you mean “this patch” > Source/WebCore/testing/Internals.h:40 > class Document; > class Element; > class Node; > +class ClientRect; Please keep these in ASCII order. > Source/WebCore/testing/Internals.h:64 > > + PassRefPtr<ClientRect> boundingBox(Element*, ExceptionCode&); We need a newline before the private: label. > Source/WebCore/testing/Internals.idl:41 > + ClientRect boundingBox(in Element element) raises(DOMException); and maybe a newline above this > LayoutTests/fast/inline/skipped-whitespace-boundingBox.html:10 > + //var rect = document.getElementById('test').getBoundingClientRect(); No need for this comment :)
http://trac.webkit.org/changeset/92059
This patch appears to have broken fast/css/outline-auto-empty-rects.html on Chromium Mac: http://test-results.appspot.com/dashboards/flakiness_dashboard.html#showExpectations=true&tests=fast%2Fcss%2Foutline-auto-empty-rects.html
This patch broke the Windows Debug and Release build bots: [[ ... 4>Linking... 4>WebCoreTestSupport.lib(JSInternals.obj) : error LNK2019: unresolved external symbol "class JSC::JSValue __cdecl WebCore::toJS(class JSC::ExecState *,class WebCore::JSDOMGlobalObject *,class WebCore::ClientRect *)" (?toJS@WebCore@@YA?AVJSValue@JSC@@PAVExecState@3@PAVJSDOMGlobalObject@1@PAVClientRect@1@@Z) referenced in function "__int64 __fastcall WebCore::jsInternalsPrototypeFunctionBoundingBox(class JSC::ExecState *)" (?jsInternalsPrototypeFunctionBoundingBox@WebCore@@YI_JPAVExecState@JSC@@@Z) 4>WebCoreTestSupport.lib(Internals.obj) : error LNK2019: unresolved external symbol "private: __thiscall WebCore::ClientRect::ClientRect(void)" (??0ClientRect@WebCore@@AAE@XZ) referenced in function "public: static class WTF::PassRefPtr<class WebCore::ClientRect> __cdecl WebCore::ClientRect::create(void)" (?create@ClientRect@WebCore@@SA?AV?$PassRefPtr@VClientRect@WebCore@@@WTF@@XZ) 4>WebCoreTestSupport.lib(Internals.obj) : error LNK2019: unresolved external symbol "private: __thiscall WebCore::ClientRect::ClientRect(class WebCore::IntRect const &)" (??0ClientRect@WebCore@@AAE@ABVIntRect@1@@Z) referenced in function "public: static class WTF::PassRefPtr<class WebCore::ClientRect> __cdecl WebCore::ClientRect::create(class WebCore::IntRect const &)" (?create@ClientRect@WebCore@@SA?AV?$PassRefPtr@VClientRect@WebCore@@@WTF@@ABVIntRect@2@@Z) 4>WebCoreTestSupport.lib(Internals.obj) : error LNK2019: unresolved external symbol "public: class WebCore::IntRect __thiscall WebCore::RenderObject::absoluteBoundingBoxRect(bool)" (?absoluteBoundingBoxRect@RenderObject@WebCore@@QAE?AVIntRect@2@_N@Z) referenced in function "public: class WTF::PassRefPtr<class WebCore::ClientRect> __thiscall WebCore::Internals::boundingBox(class WebCore::Element *,int &)" (?boundingBox@Internals@WebCore@@QAE?AV?$PassRefPtr@VClientRect@WebCore@@@WTF@@PAVElement@2@AAH@Z) 4>WebCoreTestSupport.lib(Internals.obj) : error LNK2019: unresolved external symbol "public: void __thiscall WebCore::Document::updateLayoutIgnorePendingStylesheets(void)" (?updateLayoutIgnorePendingStylesheets@Document@WebCore@@QAEXXZ) referenced in function "public: class WTF::PassRefPtr<class WebCore::ClientRect> __thiscall WebCore::Internals::boundingBox(class WebCore::Element *,int &)" (?boundingBox@Internals@WebCore@@QAE?AV?$PassRefPtr@VClientRect@WebCore@@@WTF@@PAVElement@2@AAH@Z) ]] (http://build.webkit.org/builders/Windows%20Release%20%28Build%29/builds/19341/steps/compile-webkit/logs/stdio)
Build fix attempt in r92070.
For completeness, the GTK bots have also been failing since this patch landed; complaining about a missing cairo.h, weird. See bug #65426 for more details.
(In reply to comment #4) > http://trac.webkit.org/changeset/92059 This change also broke the WinCairo Debug build. Committed build fix attempt in <http://trac.webkit.org/changeset/92078>.
(In reply to comment #9) > (In reply to comment #4) > > http://trac.webkit.org/changeset/92059 > > This change also broke the WinCairo Debug build. > > Committed build fix attempt in <http://trac.webkit.org/changeset/92078>. Committed some more fixes in <http://trac.webkit.org/changeset/92079>. I needed to add the missing symbols to WebKit2CFLite.def to fix the build. Previously I only added them to WebKit_Cairo.def and WebKit_Cairo_debug.def, but this was insufficient. Are WebKit_Cairo.def and WebKit_Cairo_debug.def still used? I left the changes to WebKit_Cairo.def and WebKit_Cairo_debug.def made in <http://trac.webkit.org/changeset/92078> in the tree. It's pretty late and I'm tired. I'll look into this some more tomorrow. If these changes aren't needed then feel free to rollout <http://trac.webkit.org/changeset/92078>. CC'ing Brent Fulgham who may have some insight.
(In reply to comment #8) > For completeness, the GTK bots have also been failing since this patch landed; complaining about a missing cairo.h, weird. See bug #65426 for more details. Fixed in <http://trac.webkit.org/changeset/92081>.