--- Begin Message ---
Subject: |
current HEAD has been crashing on startup |
Date: |
Sat, 17 Oct 2020 11:06:23 -0600 (MDT) |
User-agent: |
Alpine 2.20 (DEB 67 2015-01-07) |
I waited a few days before reporting this because I'm not sure how much
useful information I can provide in this report, but it hasn't been
mentioned by anyone else, so here goes.
Running ./emacs in src/ within gdb (invoking it with -Q), I'm getting:
Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1".
Program received signal SIGSEGV, Segmentation fault.
build_frame_matrix_from_leaf_window (w=0x555555b8c140,
frame_matrix=<optimized out>) at dispnew.c:2531
2531 fill_up_frame_row_with_spaces (frame_row,
window_matrix->matrix_x);
This is on a Linux 5.7 kernel, Debian unstable, x86-64, but I can
replicate it on a raspi running 5.4 and stable.
It appears to be related to -nw somehow; using that triggers the crash on
the Pi (otherwise it starts up normally).
On the first machine, configure options are "--with-x-toolkit=no
--without-x --with-mailutils --enable-link-time-optimization".
I'm sorry I can't be more detailed, but it's kind of hard to run M-x
report-emacs-bug under the circumstances. I can, of course, test any
patches.
--
Liam Stitt
stittl@cuug.ab.ca
--- End Message ---
--- Begin Message ---
Subject: |
Re: bug#44047: current HEAD has been crashing on startup |
Date: |
Sun, 01 Nov 2020 17:34:43 +0200 |
> Date: Sun, 1 Nov 2020 06:48:45 -0700 (MST)
> From: Liam Stitt <stittl@cuug.ab.ca>
> cc: 44047@debbugs.gnu.org
>
> > Ping! Any news on this?
>
> Belatedly, yes: the problem spontaneously went away(!), after I
> accidentally full-screened my terminal emulator then restored it to
> normal size. While I am by no means certain of why that should have
> such an effect, what digging I have been able to do suggests the
> explanation might involve (1) a SIGWINCH that operation generated which
> (2) caused that particular screen-sessions's bash to set $COLUMNS and
> $LINES, which were previously nonextant.
>
> I haven't been able to replicate the crash since. I can only suggest
> closing this bug report for now, and I'll reopen it if I manage to
> figure out how to trigger the crash again.
OK, thanks. Done.
--- End Message ---