Fwd: Re: Strange gdb backtraces on 64-bit Cygwin

[Markus Hoenicka]

At 2014-09-23 16:02, Ken Brown was heard to say:
> On 9/23/2014 9:23 AM, Markus Hoenicka wrote:
>> At 2014-09-23 14:31, Ken Brown was heard to say:
>>> 
>>> So I think it's pretty clear that the strange backtrace I observed
>>> with gdb-7.6.50-4 on 64-bit Cygwin was indeed due to a deficiency in
>>> gdb.
>>> 
>>> I hope that people who have been experiencing emacs crashes with
>>> "impossible" backtraces will update to gdb-7.8-2.
>>> 
>> 
>> Hi,
>> 
>> I might be dense somehow, but the two emacs crashes with said
>> "impossible" backtraces that I reported yesterday [1][2] were observed
>> with the gdb version that you recommend:
>> 
>> $ gdb -version
>> GNU gdb (GDB) 7.8
>> [...]
>> 
>> $ cygcheck -f /usr/bin/gdb
>> gdb-7.8-2
>> 
>> Are you sure this is all it takes to get sane backtraces?
> 
> The "impossible" backtraces that you were getting earlier were the
> ones that showed run_timers in Thread 2.  This problem didn't occur
> with the backtraces that you reported yesterday.  The only problem
> with those backtraces was the lack of detail in Thread 1.  Eli
> suggested one possible explanation.  Another is that the stack had
> gotten messed up by whatever caused the crash, so that gdb was unable
> to unwind it.  If we ever get any clue as to what's causing these
> crashes, we might come up with a reasonable place to set a breakpoint
> in order to get a useful backtrace.
> 
> By the way, I'm sure that one or more of your threads are created by
> Glib.  So installing the debuginfo for Glib as I suggested yesterday
> might help, at least in those cases where the crash/abort involves
> Glib.
> 

Thanks for the clarification. I did install the Glib debuginfo as you 
suggested earlier today.

regards,
Markus

-- 
Markus Hoenicka
http://www.mhoenicka.de
AQ score 38


--
Problem reports:       http://cygwin.com/problems.html
FAQ:                   http://cygwin.com/faq/
Documentation:         http://cygwin.com/docs.html
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple