unmerge 751569 clone 751569 -1 clone 751569 -2 reassign -1 chromium-browser reassign -2 libwebkitgtk-1.0-0 merge 751569 751358 thanks
>On Sun, Jun 15, 2014 at 11:36:09AM +0900, Mike Hommey wrote: > clone 751569 -1 > clone 751569 -2 > reassign -1 chromium-browser > reassign -2 libwebkitgtk-1.0-0 > thanks > > On Sun, Jun 15, 2014 at 09:51:49AM +0900, Mike Hommey wrote: > > On Sun, Jun 15, 2014 at 09:21:30AM +0900, Mike Hommey wrote: > > > On Sun, Jun 15, 2014 at 02:02:44AM +0200, Vincent Lefevre wrote: > > > > Control: tags -1 upstream > > > > > > > > On 2014-06-15 08:16:12 +0900, Mike Hommey wrote: > > > > > So, I was able to reproduce this crash... but not with 30.0-1. I > > > > > reproduced with both 31.0~b1-1 and 32.0~a2+20140613004003-1. > > > > > I did *not* reproduce with 31.0~b1-1~bpo70+1 from mozilla.debian.net, > > > > > or with upstream Firefox, and that may be indicating a compiler bug > > > > > \o/ > > > > > > > > I could reproduce it with upstream Firefox 30.0, still on mcl-samar.com, > > > > but not at Firefox startup, it was a bit later. Here's the crash report: > > > > > > > > https://crash-stats.mozilla.com/report/index/d429af40-44c8-4b9a-acfe-62f9f2140614 > > > > > > That's not the same crash at all. The one with the debian build is > > > because somehow there are no pointers to destructors in the vtable for > > > TSymbol, which causes a crash in TSymbolTableLevel::~TSymbolTableLevel. > > > > And the reason it does so is because TSymbol is an abstract class, and > > gcc 4.9 stops emitting destructors in the vtable of abstract classes. > > That shifts the question to why an object has that type. > > And this is an angle bug that also affects chromium and webkit. See > http://code.google.com/p/angleproject/issues/detail?id=651 > > Mike -- To UNSUBSCRIBE, email to debian-bugs-rc-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
