Hi, On Mon, Jun 18, 2012 at 1:56 AM, Lluís Batlle i Rossell <[email protected]> wrote: > On Mon, Jun 18, 2012 at 12:31:36AM +0200, Lluís Batlle i Rossell wrote: >> On Sun, Jun 17, 2012 at 11:22:34PM +0200, Lluís Batlle i Rossell wrote: >> > On Sat, Jun 09, 2012 at 08:54:23PM +0400, Kirill Elagin wrote: >> > > Ooops, I should have looked through the commit log… Eelco mentioned this >> > > in >> > > commit#f677edf57f71cf0c8c9e0906d26b4304e4213ef8<https://github.com/nbp/nixpkgs-2/commit/f677edf57f71cf0c8c9e0906d26b4304e4213ef8> >> > > . >> > > Still would be nice is someone using Firefox managed to find some spare >> > > time to resolve this… >> > >> > I thought it could be a problem in our old binutils, and I tried disabling >> > ICF >> > (--disable-icf at xulrunner), but the result is the same. So, I wonder, >> > maybe >> > it's a problem in glibc ld.so? Maybe a problem of patchelf (it is >> > patchelfed, >> > xulrunner-stub not). >> >> patchelfing 'xulrunner-stub' did not make it fail. But although I've not >> tested >> it, I think it may be patchelf breaking the binaries. >> >> There patchelf patches two executables: xpcshell and plugin-container. >> xpcshell >> was broken already in firefox12 (it suicides before the ld.so much), and >> also is >> in 13. I don't know what is xpcshell for, and what can be used for, if it >> suicides that way. Anyone knows about xpcshell? > > I confirm it's patchelf. An unpatched plugin-container, with a > LD_LIBRARY_PATH, > works fine. But xpcshell gets killed with and without patchelf.
Have you tried to configure firefox with --disable-elf-hack and potentially --disable-elf-dynstr-gc? -- Nicolas Pierron http://www.linkedin.com/in/nicolasbpierron - http://nbp.name/ _______________________________________________ nix-dev mailing list [email protected] http://lists.science.uu.nl/mailman/listinfo/nix-dev
