Created attachment 1151939 [details] f24 build root.log Description of problem: The v8 package fails to build from source on both Fedora 24 and Rawhide (as of 2016-04-28). Version-Release number of selected component (if applicable): v8-3.14.5.10-23.fc24 How reproducible: Always Steps to Reproduce: 1. koji build --scratch f24 v8-3.14.5.10-23.fc24.src.rpm Actual results: g++ -o obj/release/spaces.os -c -fno-rtti -fno-exceptions -Wall -Werror -W -Wno-unused-parameter -Woverloaded-virtual -Wnon-virtual-dtor -pedantic -m64 -O2 -g -pipe -Wall -Werror=format-security -Wp,-D_FORTIFY_SOURCE=2 -fexceptions -fstack-protector-strong --param=ssp-bu ffer-size=4 -grecord-gcc-switches -specs=/usr/lib/rpm/redhat/redhat-hardened-cc1 -m64 -mtune=generic -fPIC -fno-strict-aliasing -Wno-unused-parameter -Wno-error=strict-overflow -Wno-unused-but-set-variable -fomit-frame-pointer -fdata-sections -ffunction-sections -ansi -f PIC -DV8_TARGET_ARCH_X64 -DV8_SHARED -DBUILDING_V8_SHARED -DENABLE_DEBUGGER_SUPPORT -Isrc src/spaces.cc In file included from src/mark-compact.h:32:0, from src/incremental-marking.h:33, from src/heap.h:35, from src/elements.h:33, from src/objects-inl.h:38, from src/v8.h:60, from src/spaces.cc:28: src/spaces.h: In member function 'bool v8::internal::CodeRange::exists()': src/spaces.h:837:39: error: nonnull argument 'this' compared to NULL [-Werror=nonnull-compare] bool exists() { return this != NULL && code_range_ != NULL; } ^ src/spaces.h: In member function 'bool v8::internal::CodeRange::contains(v8::internal::Address)': src/spaces.h:839:5: error: nonnull argument 'this' compared to NULL [-Werror=nonnull-compare] if (this == NULL || code_range_ == NULL) return false; ^~ cc1plus: all warnings being treated as errors scons: *** [obj/release/spaces.os] Error 1 scons: building terminated because of errors. Expected results: Successful build Additional info: I notice the rebuild of -24 against ICU 57.1 in Rawhide failed for the same reason: http://koji.fedoraproject.org/koji/buildinfo?buildID=755094
Created attachment 1151940 [details] f24 build.log
We're carrying a very old v8 and this code no longer exists upstream, so the fix isn't a simple backport.
I am strongly motivated to simply retire v8 in Fedora. Chromium uses its bundled copy and the only consumers are: * rubygem-therubyracer * uwsgi-plugin-v8
I've got no objections, but note this was pointed out to me by someone who was hitting it trying to upgrade, so it'd be nice if it got auto-removed somehow. did we solve that problem yet?
(In reply to Tom "spot" Callaway from comment #3) > I am strongly motivated to simply retire v8 in Fedora. Chromium uses its > bundled copy and the only consumers are: > > * rubygem-therubyracer I'm using rubygem-therubyracer to precompile assets with Ruby on Rails, but it's trivial to switch to NodeJS as Rails uses rubygem-execjs to select an available runtime. I'll probably switch anyway due to bug #1331458 making the package unusable, so retiring v8 isn't a big deal if it's unmaintainable.
This is related upstream change: https://github.com/v8/v8/commit/99d9abaf82aff155a62a1ebb03c6a12920554d91
Created attachment 1164036 [details] Fix for "nonnull argument 'this' compared to NULL" This is based on https://github.com/v8/v8/commit/99d9abaf82aff155a62a1ebb03c6a12920554d91
Actually the patch header would deserve some cleanup, but I leave it to anybody else for now.
I am sick of this, so I applied "-fno-delete-null-pointer-checks" to workaround this. This should be fine for F24 and F25 has completely different version of V8 which does not suffer this issues anyway.
v8-3.14.5.10-24.fc24 has been submitted as an update to Fedora 24. https://bodhi.fedoraproject.org/updates/FEDORA-2016-67dcc8e9dd
v8-3.14.5.10-24.fc24 has been pushed to the Fedora 24 testing repository. If problems still persist, please make note of it in this bug report. See https://fedoraproject.org/wiki/QA:Updates_Testing for instructions on how to install test updates. You can provide feedback for this update here: https://bodhi.fedoraproject.org/updates/FEDORA-2016-67dcc8e9dd
> I'll probably switch anyway due to bug #1331458 making the package unusable, > so retiring v8 isn't a big deal if it's unmaintainable. After several people came out of the woodwork as consumers of v8 3.14, I've decided to make a v8-314 package so that people dependent on this specific release will have something to build against. https://bugzilla.redhat.com/show_bug.cgi?id=1344415 (This will only go into Fedora 25+)
v8-3.14.5.10-24.fc24 has been pushed to the Fedora 24 stable repository. If problems still persist, please make note of it in this bug report.