[svn.haxx.se] · SVN Dev · SVN Users · SVN Org · TSVN Dev · TSVN Users · Subclipse Dev · Subclipse Users · this month's index

Re: Subversion test results.

From: Gavin Baumanis <gavinb_at_thespidernet.com>
Date: Sat, 14 Nov 2009 11:09:44 +1100

Firstly - a big thanks Philip and Stefan for your help.

I can certainly run any command that you request and send you back the results - but please realise that I genuinely have no idea what you're talking about - so you might need to be a little more thorough in your explanations.

Anyway,
using (gdb) up I got to the line : #4 0x00000001001e94ae in representation_string - and ran the command you requested Stefan. The output is below;

(gdb) up
#1 0x0000000100519594 in apr_vformatter ()
(gdb) up
#2 0x000000010052343c in apr_pvsprintf ()
(gdb) up
#3 0x00000001005236db in apr_psprintf ()
(gdb) up
#4 0x00000001001e94ae in representation_string (rep=0x1008858b8, format=4, mutable_rep_truncated=1, pool=0x100883e28) at subversion/libsvn_fs_fs/fs_fs.c:2163
2163 return apr_psprintf(pool, "%ld %" APR_OFF_T_FMT " %" SVN_FILESIZE_T_FMT
(gdb) p *rep
$1 = {
  md5_checksum = 0x1008858f8,
  sha1_checksum = 0x0,
  revision = 0,
  offset = 0,
  size = 4,
  expanded_size = 4,
  txn_id = 0x0,
  uniquifier = 0x0
}

On 13/11/2009, at 22:08 , Stefan Sperling wrote:

> On Fri, Nov 13, 2009 at 09:26:43AM +0000, Philip Martin wrote:
>> Gavin Baumanis <gavinb_at_thespidernet.com> writes:
>>
>>> #0 0x00007fff86704cc0 in strlen ()
>>> #1 0x0000000100519594 in apr_vformatter ()
>>> #2 0x000000010052343c in apr_pvsprintf ()
>>> #3 0x00000001005236db in apr_psprintf ()
>>> #4 0x00000001001e94ae in representation_string (rep=0x1008858b8, format=4, mutable_rep_truncated=1, pool=0x100883e28) at subversion/libsvn_fs_fs/fs_fs.c:2163
>>> #5 0x00000001001e9763 in svn_fs_fs__write_noderev (outfile=0x100888e48, noderev=0x100885800, format=4, include_mergeinfo=1, pool=0x100883e28) at subversion/libsvn_fs_fs/fs_fs.c:2206
>>
>> In gdb you move up or down the stack using 'up' or 'down', and you
>> print variables using 'p', so
>> (gdb) up
>> (gdb) up
>> (gdb) up
>> (gdb) up
>> (gdb) p rep->md5_checksum[0]
>>
>> I'm guessing that something has gone wrong with the checksum
>> calculation and svn_checksum_to_cstring_display is returning NULL.
>
> This same segfault was recently reported in #svn, great that someone
> with debug symbols can reproduce it.
>
> I'd be interested in:
> (gdb) p *rep
> within the 'representation_string' function frame.
>
> Thanks,
> Stefan

------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=2417802
Received on 2009-11-14 01:10:02 CET

This is an archived mail posted to the Subversion Dev mailing list.

This site is subject to the Apache Privacy Policy and the Apache Public Forum Archive Policy.