bug#31547: 25.3; Core dump: g_main_context_prepare() called recursively from within a source's check() or prepare() member.

[Marc Wilhelm Küster]

Hello,

in order to get a more meaningful backtrace I've compiled emacs 26.1 directly 
from its source (source directly downloaded from 
http://mirror.switch.ch/ftp/mirror/gnu/emacs/emacs-26.1.tar.xz)

The crash remains fully reproducible. I attach you the backtraces. Of yourse, 
I'm at your disposal for any additional information you may need

Best regards,

Marc

backtrace.orig
Description: Binary data

backtrace.addr2line
Description: Binary data

On Tue, 29 May 2018 07:12:09 +0200,
Marc Wilhelm Küster wrote:
> 
> [1  <text/plain; US-ASCII (7bit)>]
> Dear Eli,
> 
> tonight the emacs delivered with Archlinux has updated itself to GNU Emacs 
> 26.1 (build 1, x86_64-pc-linux-gnu, GTK+ Version 3.22.30) of 2018-05-29. You 
> might be interested that the crash persists (scenario to provoke it 
> unchanged).
> 
> I attach you the backtrace (before and after processing with sed and 
> addr2line) and will look into options to get an emacs version with debugging 
> symbols (might take a moment, unfortunately).
> 
> Best regards,
> 
> Marc
> 
> [2 bt.txt <text/plain; UTF-8 (base64)>]
> (emacs:14478): GLib-WARNING **: 06:47:47.100: g_main_context_prepare() called 
> recursively from within a source's check() or prepare() member.
> 
> (emacs:14478): GLib-WARNING **: 06:47:47.101: g_main_context_check() called 
> recursively from within a source's check() or prepare() member.
> 
> Backtrace:
> emacs[0x50f76f]
> emacs[0x4f4aa2]
> emacs[0x50f824]
> emacs[0x452e6b]
> emacs[0x4555f4]
> emacs[0x5af552]
> emacs[0x5b7585]
> emacs[0x4f4848]
> emacs[0x4c090a]
> emacs[0x4c484b]
> emacs[0x4c48db]
> /usr/lib/libX11.so.6(_XError+0x11d)[0x7f329f7076ed]
> /usr/lib/libX11.so.6(+0x41617)[0x7f329f704617]
> /usr/lib/libX11.so.6(+0x416d5)[0x7f329f7046d5]
> /usr/lib/libX11.so.6(_XEventsQueued+0x55)[0x7f329f704fd5]
> /usr/lib/libX11.so.6(XPending+0x57)[0x7f329f6f6ca7]
> /usr/lib/libgdk-3.so.0(+0x66f6f)[0x7f32a0f25f6f]
> /usr/lib/libglib-2.0.so.0(g_main_context_prepare+0x1c9)[0x7f329fe71a49]
> /usr/lib/libglib-2.0.so.0(+0x4c44e)[0x7f329fe7244e]
> /usr/lib/libglib-2.0.so.0(g_main_context_pending+0x29)[0x7f329fe725e9]
> /usr/lib/libgtk-3.so.0(gtk_events_pending+0xf)[0x7f32a13e3e3f]
> emacs[0x4c1339]
> emacs[0x4fbeb9]
> emacs[0x4fc585]
> emacs[0x5dd884]
> emacs[0x5861e4]
> emacs[0x5e03db]
> emacs[0x5e0663]
> emacs[0x5e097c]
> emacs[0x443ce4]
> emacs[0x440441]
> emacs[0x44583a]
> emacs[0x445a6d]
> emacs[0x470b81]
> emacs[0x56cd1d]
> emacs[0x5a8ff1]
> emacs[0x56f684]
> emacs[0x56cc9b]
> emacs[0x56cd39]
> emacs[0x56c2fd]
> emacs[0x56c430]
> ...
> 
> 
> [3 bt2.txt <text/plain; UTF-8 (base64)>]
> XSetWMNormalHints at ??:?
> XSetWMNormalHints at ??:?
> XSetWMNormalHints at ??:?
> ?? ??:0
> ?? ??:0
> error at ??:?
> error at ??:?
> XSetWMNormalHints at ??:?
> ?? ??:0
> ?? ??:0
> ?? ??:0
> ?? ??:0
> ?? ??:0
> ?? ??:0
> ?? ??:0
> ?? ??:0
> ?? ??:0
> ?? ??:0
> ?? ??:0
> ?? ??:0
> ?? ??:0
> ?? ??:0
> XSetWMNormalHints at ??:?
> XSetWMNormalHints at ??:?
> error at ??:?
> error at ??:?
> error at ??:?
> error at ??:?
> error at ??:?
> ?? ??:0
> ?? ??:0
> ?? ??:0
> ?? ??:0
> ?? ??:0
> re_compile_pattern at ??:?
> error at ??:?
> error at ??:?
> re_compile_pattern at ??:?
> re_compile_pattern at ??:?
> re_compile_pattern at ??:?
> re_compile_pattern at ??:?
> [4  <text/plain; ISO-8859-1 (quoted-printable)>]
> 
> 
> On Wed, 23 May 2018 17:18:47 +0200,
> Eli Zaretskii wrote:
> > 
> > > Date: Wed, 23 May 2018 07:10:00 +0200
> > > From: Marc Wilhelm Küster <marc@budabe.eu>
> > > Cc: Marc Wilhelm Küster <marc@budabe.eu>,
> > >   31547@debbugs.gnu.org
> > > 
> > > > That's because your Emacs binary is stripped of debugging symbols.  Is
> > > > there perhaps an auxiliary distribution with those symbols as separate
> > > > files?  If so, could you please install that and repeat what you've
> > > > done?
> > > > 
> > > Not to my knowledge out of the box, I'm afraid. I could perhaps configure 
> > > and compile my own version of the emacs package with debuging included, 
> > > but I'd need some time to study how that might work under Archlinux.
> > 
> > If you do decide to build your own Emacs, I suggest to build version
> > 26.1, because it's quite possible this problem is fixed there (and if
> > it isn't, it becomes much more urgent).
> > 
> > > Are there alternatives?
> > 
> > Not that I know of.