Rainer M Krug <Rainer@krugs.de> writes:

> Eli Zaretskii <eliz@gnu.org> writes:
>
>>> From: Rainer M Krug <Rainer@krugs.de>
>>> Cc: 21428@debbugs.gnu.org,  mituharu+bug-gnu-emacs-mac@math.s.chiba-u.ac.jp
>>> Date: Thu, 01 Oct 2015 10:21:04 +0200
>>> 
>>> OK - it seems not to continue after some time after the breakpoint at
>>> line 4199.
>>> 
>>> Attached are the log and the script I used.
>>> 
>>> I'll leave the session open.
>>> 
>>> This seems to be triggered by something else - I manually entered "c"
>>> but it just gave me this output (also at the end of the log):
>>> 
>>> ,----
>>> | (gdb) c
>>> | Continuing.
>>> | 
>>> | Program received signal EXC_BAD_ACCESS, Could not access memory.
>>> | Reason: KERN_INVALID_ADDRESS at address: 0x00000000016519e0
>>> | free_realized_faces (c=0x1016519c0) at xfaces.c:4199
>>> | 4199	      c->used = 0;
>>> | (gdb) p c->used
>>> | $502 = 26
>>> | (gdb)
>>> `----
>>
>> This is strange: the problematic address, 0x00000000016519e0, is the
>> address of the cache (0x1016519c0) with the 33rd bit reset.  Are you
>> sure you don't have some hardware problem on that machine?
>>
>> Or maybe it's a GDB bug?  Or even some kernel bug?
>
> I somehow have the feeling that it could be a kernel bug, as I have other
> strange things happening as well which also could be caused by a kernel
> bug. As I am planing to upgrade to El Capitan today or tomorrow, I will
> wait with further test until then and let you know.
>
> Another pointer towards the kernel: the error only occurred after a
> holiday, which resulted in the upgrade of the OS, emacs, packages, ...
>
> So I could not pinpoint what caused it.
>
>>
>> I don't really understand what this means.  I can only suggest to run
>> another session with the same commands, and see if it fails in the
>> same way.
>
> This was already the third session, as it also did not make sense to
> me. Also: continued "continue" resulted in exactly the same error
> message and no continuation.
>
> OK - I will then upgrade and report back in a few days time to see if
> the upgrade solved the problem.
>
> If it was kernel, I am really sorry for your wasted time,
>
> I will report back then,

OK - I upgraded the OS and emacs is still crashing (although I think
less frequent?).

I attach the log and I think we are going somewhere.

I am still on

,----
| commit 9e4a6ad1c77c62e4a7d5b708567e5a1e23b9cdfd
`----

and use the same compilation as before.

I have left the session open so just let me know if you need further
info.

Thanks,

Rainer