"Michael Pfob" <m.pfob_at_3s-software.com> writes:
> I'm sorry, I don't get the time for supplying a complete patch for
> this topic. I think it will be possible that I try to write a patch
> for this bug but I think I'm not able to set up a build environment
> for subversion as this seems to take quite some time using windows.
>
> If you are interested, I could supply a patch without compiling or
> testing it. Otherwise I would be looking forward to seeing this bug
> fixed in one of the next versions even if I found the workaround for
> my special application.
>
> Please let me know if I should supply the patch even with these
> restrictions.
Well, if you can't build with the patch, it probably would be a waste of
your time. Thank you for offering, though.
It would be great if you could file an issue, pointing to this thread,
though. We don't want to forget about this, even though we (well, I)
don't have time to fix it right now.
-Karl
> -----Ursprüngliche Nachricht----- Von: Karl Fogel
> [mailto:kfogel_at_red-bean.com] Gesendet: Montag, 7. Juli 2008 18:13 An:
> Michael Pfob Cc: dev_at_subversion.tigris.org Betreff: Re: AW: Possible
> access violation in svn_client_copy
>
> "Michael Pfob" <m.pfob_at_3s-software.com> writes:
>> I did some debugging now and I think i found the reason for my
>> problems. The reason was that I did not pass a log_msg_func. The
>> code in repos_to_repos_copy in my opinion is wrong because the check
>> for SVN_CLIENT__HAS_LOG_MSG_FUNC executes code with side effects
>> like filling the table action_hash which leads to problems in my
>> code. I think such a side effect should not happen in this
>> if-branch.
>> Additionally I tried to format my original post correctly.
>
> Thank you for tracing this.
>
> I think you're right: the "if (SVN_CLIENT__HAS_LOG_MSG_FUNC(ctx))
> {...}" code branch should not enclose code that modifies action hash
> in a way that the rest of the function needs. Instead, that code
> should happen outside the 'if'.
>
> Since you're already set up to test this easily, would you like to try
> writing a patch? (see
> http://subversion.tigris.org/hacking.html#patches)
>
> If not, don't worry about it. You did provide a reproduction program,
> and someone else can take this if you don't have time. But it sounds
> like you have all the knowledge needed here, so it would be great if
> you could take it all the way to the finish line...
>
> Let us know.
>
> Thanks, -Karl
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe_at_subversion.tigris.org
For additional commands, e-mail: dev-help_at_subversion.tigris.org
Received on 2008-07-08 19:06:48 CEST