> On Dec 30, 2015, at 11:38 AM, Trevor Woerner <[email protected]> wrote: > > Hi Khem, > > On 12/30/15 13:04, Khem Raj wrote: >> can you backport and report back >> >> https://chromium.googlesource.com/v8/v8/+/e28183b5977a2d2732f7f8b9a2f26637c9566585%5E%21/#F0 > > That diff includes 4 patches: > Index: src/bignum.cc > Index: src/compiler/ia32/code-generator-ia32.cc > Index: src/compiler/x64/code-generator-x64.cc > Index: test/cctest/compiler/test-js-typed-lowering.cc > > All of these patches only apply to the code in the "v8" folder. > > The 4th one doesn't apply because I can't find a > "test-js-typed-lowering.cc" file anywhere in our version 40 sources. > > Although the files "v8/src/compiler/ia32/code-generator-ia32.cc" and > "v8/src/compiler/x64/code-generator-x64.cc" exist, these two patches > don't apply since the code in those files doesn't look anything like > what these patches are trying to patch. > > The first patch does apply, with the file path change (add v8 to the > front), and a 2 line offset. > > For some reason this bignum.cc source is found in both v8/src and > third_party/WebKit/Source/wtf/dtoa and, apart from some extra namespace > scope, is very similar (if not the same). So I added a second patch to > this diff to also correctly patch the code in third_party/... . > > Replacing Max's solution for this one does cause this one file to > compile fine, but other -Werror=strict-overflow problems are simply > found later (see attached). > > Combining both Max's patch and this one causes this part of the build to > succeed, but still causes the -Werror=uninitialized problem to happen as > I reported earlier:
Suppressing a given Werror is a workaround. So I am more interested in code changes instead if they are new so be it we can fix and send them over to chromium folks. > > | > ../../third_party/WebKit/Source/core/inspector/InspectorTimelineAgent.cpp:234:8: > error: '<anonymous>.blink::TimelineThreadState::inKnownLayerTask' is > used uninitialized in this function [-Werror=uninitialized] > This is easily fixable by helping compiler with defining this to 0 or something where its declared. > Best regards, > Trevor > <strict-overflow.txt>
signature.asc
Description: Message signed with OpenPGP using GPGMail
-- _______________________________________________ Openembedded-devel mailing list [email protected] http://lists.openembedded.org/mailman/listinfo/openembedded-devel
