thomas schorpp schrieb:
> thomas schorpp schrieb:
>> thomas schorpp schrieb:
>>> thomas schorpp schrieb:
>>>> thomas schorpp schrieb:
>>>>> thomas schorpp schrieb:
>>>>>> Tzafrir Cohen schrieb:
>>>>>>> On Fri, Oct 24, 2008 at 11:51:27PM +0200, thomas schorpp wrote:
>>>>>>>> Tzafrir Cohen schrieb:
>>>>>>>>> On Fri, Oct 24, 2008 at 09:46:25PM +0200, thomas schorpp wrote:
>>>>>>>>>> Tzafrir Cohen schrieb:
>>>>>>>>>>> Does the callweaver binary have debug information or is it
>>>>>>>>>>> stripped?
>>>>>>>>>> dh_strip
>>>>>>>>>>
>>>>>>>>>>> If you use debs, use a -dbg package.
>>>>>>>>>> current debian/ scripts included do not build one.
>>>>>>>>>>
>>>>>>>>> dh_strip --dbg-package=callweaver-dbg
>>>>>>>>>
>>>>>>>>> and add a (binary) package called callweaver-dbg in the control .
>>>>>>>>>
>>>>>>>>> diff --git a/debian/control b/debian/control
>>>>>>>>> index 10fb21a..779b39e 100644
>>>>>>>>> --- a/debian/control
>>>>>>>>> +++ b/debian/control
>>>>>>>>> @@ -33,6 +33,17 @@ Description: Development files for Callweaver
>>>>>>>>> This package contains the include files used if you wish to
>>>>>>>>> compile a
>>>>>>>>> package which requires Callweaver source file headers.
>>>>>>>>>
>>>>>>>>> +Package: callweaver-dev
>>>>>>>>> +Section: devel
>>>>>>>>> +Priority: extra
>>>>>>>>> +Architecture: any
>>>>>>>>> +Depends: Depends: callweaver (= ${binary:Version})
>>>>>>>>> +Description: Development files for Callweaver
>>>>>>>>> + Callweaver is an Open Source PBX and telephony toolkit.
>>>>>>>>> + .
>>>>>>>>> + This package contains the debugging symbols for modules. Use
>>>>>>>>> it if you want
>>>>>>>>> + to run callweaver under a debugger.
>>>>>>>>> +
>>>>>>>>> Package: callweaver-doc
>>>>>>>>> Recommends: callweaver
>>>>>>>>> Section: doc
>>>>>>>>> diff --git a/debian/rules b/debian/rules
>>>>>>>>> index 5c1fde4..1c5d6c2 100755
>>>>>>>>> --- a/debian/rules
>>>>>>>>> +++ b/debian/rules
>>>>>>>>> @@ -169,7 +169,7 @@ binary-arch: install
>>>>>>>>> # dh_installinfo -a
>>>>>>>>> # dh_installman -a
>>>>>>>>> dh_link -a
>>>>>>>>> - dh_strip -a
>>>>>>>>> + dh_strip -a --dbg-package=callweaver-dbg
>>>>>>>>> dh_compress -a
>>>>>>>>> dh_fixperms -a
>>>>>>>>> # dh_perl -a
>>>>>>>>>
>>>>>>>> does not work
>>>>>>>> +Depends: Depends: callweaver (= ${binary:Version})
>>>>>>>> gives some unresolved deps error
>>>>>>>
>>>>>>> $ cat debian/compat
>>>>>>> 4
>>>>>>
>>>>>> check'd.
>>>>>>
>>>>>>>
>>>>>>> dh_stript(1) reads:
>>>>>>>
>>>>>>> --dbg-package=package
>>>>>>>
>>>>>> ACK but:
>>>>>> SYNOPSIS
>>>>>> dh_strip [debhelper options] ...
>>>>>> which implies "-a" since other dh_* use "-a", too.
>>>>>>
>>>>>> but well, if theres nothing to strip so no output from dh_strip?
>>>>>>
>>>>>> will configure with --enable-debug then we will see symbols in the
>>>>>> *-dbg debs, i hope.
>>>>>>
>>>>>> y
>>>>>> tom
>>>>>>
>>>>>
>>>>> dh_strip -v --dbg-package=...
>>>>>
>>>>> says only strip'd but written to nowhere, giving up dbg deb.
>>>>>
>>>>> done it like below, runs fine in home env with < 10% cpu on a
>>>>> P2-266, it should segfault over the day, will give full bt.
>>>>>
>>>>> y
>>>>> tom
>>>>>
>>>>
>>>> ts. segfault gone with debug builds of CW and spandsp:
>>>>
>>>> tom3:~# grep core /var/log/syslog
>>>> tom3:~# date
>>>> Sat Oct 25 19:01:30 UTC 2008
>>>>
>>>> 12h now.
>>>>
>>>> waiting.
>>>>
>>>> y
>>>> tom
>>>>
>>>>
>>>>
>>>
>>> caught. segfault on reregister from ekiga v.3, looks like
>>> double-free, mailing full bt to steve only, may contain account info.
>>>
>>> old ekiga snapshot, trying new ekiga 3.0.1 next:
>>>
>>> (gdb) bt
>>> #0 0xb666667f in __sip_destroy (p=0x8158590, lockowner=1) at
>>> chan_sip.c:3075
>>> #1 0xb6666bbd in sip_destroy (p=0x8158590) at chan_sip.c:3262
>>> #2 0xb668c401 in handle_request_subscribe (p=0x8119c40,
>>> req=0xb66420d4, debug=0, ignore=0, sin=0xb6640e94, seqno=37,
>>> e=0xb66422f6 "sip:[email protected]") at chan_sip.c:14103
>>> #3 0xb668d52b in handle_request (p=0x8119c40, req=0xb66420d4,
>>> sin=0xb6640e94, recount=0xb6640e74, nounlock=0xb6640e78)
>>> at chan_sip.c:14418
>>> #4 0xb668e0a4 in sipsock_read (id=0x80aecb8, fd=8, events=1,
>>> ignore=0x0) at chan_sip.c:14607
>>> #5 0xb7ef787e in cw_io_wait (ioc=0x8076c90, howlong=1000) at io.c:287
>>> #6 0xb668e9c9 in do_monitor (data=0x0) at chan_sip.c:14805
>>> #7 0xb7e9f4fb in start_thread () from /lib/i686/cmov/libpthread.so.0
>>> #8 0xb7c5860e in clone () from /lib/i686/cmov/libc.so.6
>>
>> no more reproducible since ekiga upgrade to 3.0.1:
>>
>> tom3:~# zcat /var/log/syslog.1.gz |grep core
>> Oct 26 14:53:00 tom3 logger: callweaver: Found and moved core dumps to
>> /var/lib/callweaver/core/2008-10-26-14:53:00
>> tom3:~# tom3:~# grep core /var/log/syslog.0
>> tom3:~# grep core /var/log/syslog
>> tom3:~#
>> y
>> tom
>>
>
> tom3:~# ls -ltr /var/lib/callweaver/core
> total 16
> drwxr-xr-x 2 callweaver callweaver 4096 2008-10-24 15:45
> 2008-10-24-15:45:55
> drwxr-xr-x 2 callweaver callweaver 4096 2008-10-26 14:53
> 2008-10-26-14:53:00
> drwxr-xr-x 2 root root 4096 2008-11-04 12:38
> 2008-11-04-12:38:07
> drwxr-xr-x 2 root root 4096 2008-11-08 04:02
> 2008-11-08-04:02:19
>
> #0 0xb66399d7 in __sip_destroy (p=0x80b37e0, lockowner=1) at
> chan_sip.c:3127
> 3127 p->history = p->history->next;
> (gdb) bt
> #0 0xb66399d7 in __sip_destroy (p=0x80b37e0, lockowner=1) at
> chan_sip.c:3127
> #1 0xb6639bbd in sip_destroy (p=0x80b37e0) at chan_sip.c:3262
> #2 0xb665f401 in handle_request_subscribe (p=0x80c86b8, req=0xb66140d4,
> debug=0, ignore=0, sin=0xb6612e94, seqno=2, e=0xb66142f6
> "sip:xxx...@xxxxxx") at chan_sip.c:14103
> #3 0xb666052b in handle_request (p=0x80c86b8, req=0xb66140d4,
> sin=0xb6612e94, recount=0xb6612e74, nounlock=0xb6612e78)
> at chan_sip.c:14418
> #4 0xb66610a4 in sipsock_read (id=0x80acec0, fd=8, events=1,
> ignore=0x0) at chan_sip.c:14607
> #5 0xb7ecb8ce in cw_io_wait (ioc=0x8076c90, howlong=1000) at io.c:296
> #6 0xb66619c9 in do_monitor (data=0x0) at chan_sip.c:14805
> #7 0xb7e724fb in start_thread () from /lib/i686/cmov/libpthread.so.0
> #8 0xb7c2b60e in iswgraph () from /lib/i686/cmov/libc.so.6
> Backtrace stopped: previous frame inner to this frame (corrupt stack?)
> (gdb)
> full bt to Steve.
>
> y
> tom
>
fixed, thx :)
y
tom
_______________________________________________
Callweaver-dev mailing list
[email protected]
http://lists.callweaver.org/mailman/listinfo/callweaver-dev
