Re: Crash when moving full-width glyphs across lines

From	Chet Ramey <chet.ramey@case.edu>
Newsgroups	gnu.bash.bug
Subject	Re: Crash when moving full-width glyphs across lines
Date	2019-12-20 15:43 -0500
Organization	ITS, Case Western Reserve University
Message-ID	<mailman.1201.1576874605.1979.bug-bash@gnu.org> (permalink)
References	<d9e015f7-aa5c-4c5e-90dc-8cf0a341ab0a@www.fastmail.com> <20191219190032.m2p7vsuowclcp355@RedPanda.localdomain> <8a669137-3205-e8a0-5e0f-84d87ca7a126@case.edu>

Show all headers | View raw

On 12/19/19 2:00 PM, Kirill Chibisov wrote:
> I was able to capture the stack trace under gdb. However you should
> probably use the following command in it to repro the issue.
> 
> `set disable-randomization off`
> 
> The issue happens quite randomly, however the important thing is exit code
> from bash (139). So there's something with memory access, I guess.
> 
> (gdb) info stack
> #0  0x00007fe30e6955ba in _rl_col_width (flags=<optimized out>, end=1735628519, start=<optimized out>, str=0x557e988c691a "/こんにち")
>      at /usr/src/debug/sys-libs/readline-8.0_p1-r1/readline-8.0/display.c:3206
> #1  _rl_col_width (str=0x557e988c691a "/こんにち", start=<optimized out>, end=1735628519, flags=<optimized out>)

Well, the values for `end' are obviously suspicious, but you can't really
tell too much from this traceback, especially since other values of
interest are optimized out. The value of `end' is computed in the caller,
from `oe - old'. Since those are pointers into the same string, the value
shouldn't be that far out of whack. The values of `old' and `new' are
helpful, though. I'll see if I can find it.

Chet

-- 
``The lyf so short, the craft so long to lerne.'' - Chaucer
		 ``Ars longa, vita brevis'' - Hippocrates
Chet Ramey, UTech, CWRU    chet@case.edu    http://tiswww.cwru.edu/~chet/

Back to gnu.bash.bug | Previous | Next | Find similar

Thread

Re: Crash when moving full-width glyphs across lines Chet Ramey <chet.ramey@case.edu> - 2019-12-20 15:43 -0500

csiph-web