I have tested each ebuild for liferea since 0.6.0, rebuilding liferea-0.6.0 after each (i.e. it's not my compiler doing it). Each version, even after a mv ~/.liferea ~/.liferea.old, segfaults upon opening a folder: read(25, "##\n## The contents of this file "..., 8192) = 7448 read(25, "", 8192) = 0 close(25) = 0 futex(0x552b6beeec, FUTEX_WAKE, 1) = 1 futex(0x552b6beee8, FUTEX_WAKE, 1) = 1 futex(0x552b6beeec, FUTEX_WAKE, 1) = 1 futex(0x552b6beee8, FUTEX_WAKE, 1) = 1 futex(0x552b6beeec, FUTEX_WAKE, 1) = 1 futex(0x552b6beee8, FUTEX_WAKE, 1) = 1 --- SIGSEGV (Segmentation fault) @ 0 (0) --- unlink("/home/bluefox/.liferea/mozilla/liferea/lock") = 0 rt_sigaction(SIGSEGV, {SIG_DFL}, NULL, 8) = 0 rt_sigprocmask(SIG_UNBLOCK, [SEGV], NULL, 8) = 0 tgkill(30612, 30612, SIGSEGV) = 0 --- SIGSEGV (Segmentation fault) @ 0 (0) --- +++ killed by SIGSEGV +++ This is on amd64. The only stable version of liferea seems to be liferea-0.6.0; and the ebuild has been removed. Please re-add it ASAP.
set hardware to amd64
Works perfectly for me. Can you attach your emerge info please?
Portage 2.0.51-r8 (hardened/amd64, gcc-3.4.3, glibc-2.3.4.20041102-r0, 2.6.9-pax x86_64) ================================================================= System uname: 2.6.9-pax x86_64 AMD Athlon(tm) 64 Processor 2800+ Gentoo Base System version 1.6.7 Python: dev-lang/python-2.3.4 [2.3.4 (#1, Oct 21 2004, 18:39:55)] ccache version 2.3 [enabled] dev-lang/python: 2.3.4 sys-devel/autoconf: 2.13, 2.59-r6 sys-devel/automake: 1.5, 1.4_p6, 1.6.3, 1.9.3, 1.7.9, 1.8.5-r2 sys-devel/binutils: 2.15.92.0.2-r1 sys-devel/libtool: 1.5.2-r7 virtual/os-headers: 2.6.8.1-r1 ACCEPT_KEYWORDS="amd64 ~amd64" AUTOCLEAN="yes" CFLAGS="-march=athlon64 -O2 -pipe -ftracer -fweb -funit-at-a-time -fomit-frame-pointer -mfpmath=387 -mno-sse -fno-stack-protector-all -fstack-protector" CHOST="x86_64-pc-linux-gnu" CONFIG_PROTECT="/etc /usr/X11R6/lib/X11/xkb /usr/kde/2/share/config /usr/kde/3/share/config /usr/lib/mozilla/defaults/pref /usr/share/config /var/qmail/control" CONFIG_PROTECT_MASK="/etc/gconf /etc/terminfo /etc/env.d" CXXFLAGS="-march=athlon64 -O2 -pipe -ftracer -fweb -funit-at-a-time -fomit-frame-pointer -mfpmath=387 -mno-sse -fno-stack-protector-all -fstack-protector" DISTDIR="/home/shared/sysdirs//distfiles" FEATURES="autoaddcvs buildpkg ccache distlocks sandbox" GENTOO_MIRRORS="http://open-systems.ufl.edu/mirrors/gentoo http://gentoo.mirrors.pair.com/ ftp://ftp.tu-clausthal.de/pub/linux/gentoo/" MAKEOPTS="-j2" PKGDIR="/home/shared/sysdirs//packages" PORTAGE_TMPDIR="/var/tmp" PORTDIR="/usr/portage" PORTDIR_OVERLAY="/usr/local/portage/olay64 /usr/local/portage/common /usr/local/portage/bmg-gnome /usr/local/portage/bmg" SYNC="rsync://rsync.us.gentoo.org/gentoo-portage" USE="3ds X aalib acl acpi alsa amd64 apm avi berkdb bitmap-fonts bmp bootsplash caps cdr composite crypt cups dlloader dmx dri dvd esd f77 fam flac gcj gdbm gif gimpprint gnome gpm gstreamer gtk gtk2 gtkhtml hal hardened imagemagick imlib ipv6 ithreads java jbig jpeg justify lcms ldap libwww lzw-tiff mad mikmod mmap mng motif mozilla moznocompose moznoirc moznomail mpeg multilib mysql ncurses nls nptl objc offensive oggvorbis openal opengl oss pam perl pic pie png ppds python quicktime readline samba sdl slang speex spell ssl tcltk tcpd theora threads tiff truetype unstable-meta usb videos wmf xchatdccserver xml xml2 xprint xv zlib" No modifications were made between compiling 0.6.0 and future versions (I actually rebuilt 0.6.0 during the downgrades to check to make sure it wasn't me).
i've disabled ccache (htf did that get on), remerged 0.6.4b, it crashes when selecting a feed (again); remerged 0.6.0 (from source), it works. :/ well, wasn't ccache at least. . .
Why don't you start by toning: CFLAGS="-march=athlon64 -O2 -pipe -ftracer -fweb -funit-at-a-time -fomit-frame-pointer -mfpmath=387 -mno-sse" down to something like CFLAGS="-march=athlon64 -O2 -pipe"
Tried. Liferea 0.6.4b still segfaults with fresh ~/.liferea
I've all but tracked this bug down. I had to execstack -c some stuff LoadPlugin: failed to initialize shared library /opt/blackdown-jdk-1.4.2.01/jre/plugin/amd64/mozilla/libjavaplugin_oji.so [/opt/blackdown-jdk-1.4.2.01/jre/plugin/amd64/mozilla/libjavaplugin_oji.so: cannot enable executable stack as shared object requires: Permission denied] LoadPlugin: failed to initialize shared library /opt/netscape/plugins/libflashplayer.so [/opt/netscape/plugins/libflashplayer.so: cannot open shared object file: No such file or directory] Segmentation fault bluefox@icebox ~ $ execstack \ /opt/blackdown-jdk-1.4.2.01/jre/plugin/amd64/mozilla/libjavaplugin_oji.so \ /opt/netscape/plugins/libflashplayer.so ? /opt/blackdown-jdk-1.4.2.01/jre/plugin/amd64/mozilla/libjavaplugin_oji.so ? /opt/netscape/plugins/libflashplayer.so bluefox@icebox ~ $ su -c "execstack -c \ /opt/blackdown-jdk-1.4.2.01/jre/plugin/amd64/mozilla/libjavaplugin_oji.so \ /opt/netscape/plugins/libflashplayer.so" After this and deleting the gconf key /apps/liferea/browser-module, the mozilla module works fine. Oddly, this doesn't KILL firefox, but it prevents it from loading the affected plug-ins. What the heck is with the ? in PT_GNU_STACK (read by execstack)? A set PT_GNU_STACK puts an X there. Hardened should probably have a look at this, so I'm putting on my helmet and CCing them. I sent a message at their dev list too, so the Liferea devs got a heads-up on this. The message hasn't hit the archive yet, but if it ever does I'll put a link here. What I wanna know is, what broke in >0.6.0 that started triggering this?
> What I wanna know is, what broke in >0.6.0 that started triggering this? --- That's a queastion upstream is much more likely to answer. As I have seen you have reported your problem at their SF bugtracker already, so maybe you'll get your answer there.
http://sourceforge.net/mailarchive/message.php?msg_id=10682111 And suddenly it crashes AGAIN, so this bug STILL STANDS. 0.9.0b crashes with either plug-in. There's an strace. Anyone who wants to tackle taking this apart can. Upstream will have to deal with this most likely; it's nobody here's job, though volunteers are always welcome (I've had this one pulled on me before, I didn't bite). Currently I keep >0.6.0 masked in /etc/portage, so eh.
Ok, I'll finally mark this bug as RESOLVED - UPSTREAM. If you have anything to share, feel free to post the information here if you want.
