Summary: | =net-libs/webkit-gtk-1.8.1-r301 - Source/WebCore/xml/XPathParser.cpp:478:39: error: too many arguments to function «int WebCore::XPath::xpathyyparse()» with sys-devel/bison-2.6 | ||
---|---|---|---|
Product: | Gentoo Linux | Reporter: | Sergey Isachenko <vortexius> |
Component: | [OLD] GNOME | Assignee: | Gentoo Linux Gnome Desktop Team <gnome> |
Status: | RESOLVED FIXED | ||
Severity: | normal | CC: | dschridde+gentoobugs, floppym, halcy0n, hwoarang, phabcak, remi, thev00d00, vortexius |
Priority: | Normal | ||
Version: | unspecified | ||
Hardware: | All | ||
OS: | Linux | ||
URL: | https://bugs.webkit.org/show_bug.cgi?id=91943 | ||
See Also: | https://bugs.gentoo.org/show_bug.cgi?id=427438 | ||
Whiteboard: | |||
Package list: | Runtime testing required: | --- | |
Attachments: |
emerge info
build log (-j1) |
Description
Sergey Isachenko
2012-07-25 10:37:22 UTC
Created attachment 319204 [details]
emerge info
Created attachment 319206 [details]
build log (-j1)
Same problem as in bug 427438. No version of webkit-gtk currently in portage will compile with bison-2.6. As a temporary work around, I'm trying the following: emerge -1 =sys-devel/bison-2.5.1 && emerge webkit-gtk && emerge bison The problem appears to be related to chromium's recent requirement for bison-2.6 {and subsequent bump thereof}. On my systems, portage currently wants to downgrade bison but won't because of the build dependency in my existing builds of chromium. I believe the choices right now until this is fixed is either try the same work around I'm trying or downgrade to the last version of chromium which works with an earlier version of bison and then downgrade bison as well. I don't yet know if this possible work around will succeed. I'm trying it now. As this is my 5th attempt at building webkit-gtk .. on an Atom based netbook, it'll be awhile before I can report results. (In reply to comment #4) > No version of webkit-gtk currently in portage will compile with bison-2.6. > > As a temporary work around, I'm trying the following: > > emerge -1 =sys-devel/bison-2.5.1 && emerge webkit-gtk && emerge bison > > > I don't yet know if this possible work around will succeed. I'm trying it > now. As this is my 5th attempt at building webkit-gtk .. on an Atom based > netbook, it'll be awhile before I can report results. I'm sorry to say I cannot report a positive result. Bison-2.5.1 emerged fine. The subsequent attempt to emerge webkit-gtk apparently ran into some kind of race condition. The drive light was on solid and I ended up killing the machine through a hard power off after the load average reached 15.0. You read that right. An Atom {gen 2} with no other substantive activity {other than normal background processes in a razor-qt environment with no other processes {i.e. no browsers, no other terminal windows, no external sessions - nada} reached an ave load value of 15.0. The system was totally unresponsive to user input. Even ssh timed out before I could enter a user id. Oh well. Guess I'll have to live without the latest calibre on this machine for now. Confirming on my relatively standard amd64 box. (In reply to comment #6) > Confirming on my relatively standard amd64 box. Also confirming that the build works as intended after downgrading bison to 2.5.1. This just closed bug (with patch) from upstream seems relevant: https://bugs.webkit.org/show_bug.cgi?id=92264 "Build fix with newer bison 2.6." Apparently this is from bison 2.6: "The release notes for bison 2.6 mention that the YYPARSE_PARAM macro has be deprecated." Yeah, that webkit bug has the fix. I've just committed the patch for chromium; it should be work for webkit-gtk as well. http://trac.webkit.org/changeset/124099 (In reply to comment #9) > Yeah, that webkit bug has the fix. I've just committed the patch for > chromium; it should be work for webkit-gtk as well. > > http://trac.webkit.org/changeset/124099 Could someone at least fix the bison (<2.6) deps in webkit-gtk ebuild so we can workaround this problem until a proper patch is applied in the ebuild? (In reply to comment #10) > (In reply to comment #9) > > Yeah, that webkit bug has the fix. I've just committed the patch for > > chromium; it should be work for webkit-gtk as well. > > > > http://trac.webkit.org/changeset/124099 > > Could someone at least fix the bison (<2.6) deps in webkit-gtk ebuild so we > can workaround this problem until a proper patch is applied in the ebuild? I added <sys-devel/bison-2.6 as DEPEND in both ebuilds (1.8.1-r201 and 1.8.1-r301). Feel free to revert this change once you commit the upstream fix. I leave this bug open in order to avoid duplicates Fixed in 1.8.2-r{200,300}. Guy, thanks for figuring out the cause of the problem and finding the patch. >*webkit-gtk-1.8.2-r300 (09 Aug 2012) >*webkit-gtk-1.8.2-r200 (09 Aug 2012) > > 09 Aug 2012; Alexandre Rostovtsev <tetromino@gentoo.org> > -webkit-gtk-1.8.0-r200.ebuild, -webkit-gtk-1.8.0-r300.ebuild, > +webkit-gtk-1.8.2-r200.ebuild, +webkit-gtk-1.8.2-r300.ebuild, > +files/webkit-gtk-1.8.2-bison-2.6.patch: > Version bump with security and crash fixes. Fix building with bison-2.6 (bug > #428012, thanks to Zabuldon, Guy, Mike Gilbert et al.). Drop old. |