You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@subversion.apache.org by Eric Gillespie <ep...@pretzelnet.org> on 2004/01/16 14:55:46 UTC
Problem with r7950
r7950 says:
Introduce a convenient way to wrap errors from APR functions.
This introduces an infinite loop in the libsvn_subr/utf.c
functions convert_to_stringbuf and get_xlate_handle. Both these
functions call svn_error_wrap_apr. svn_error_wrap_apr then calls
svn_utf_cstring_to_utf8, which calls get_ntou_xlate_handle, which
calls get_xlate_handle. Round and round we go.
What to do about it though? Before r7950 svn_error_createf was
used here, which doesn't have the problem. But is that the right
answer?
Ben Collins-Sussman asks me for a reproduce recipe. I don't
actually have one, though the bug is quite obvious without it.
This was reported by a user of the pkgsrc package on NetBSD
-current. On that system, it is reproducible by not having
/usr/share/i18n. This results in ENOENT from apr_xlate_open, so
he suggested handling ENOENT with the other specially handled
errors. Unfortunately, that still leaves some error conditions
that lead to the infinite loop.
--
Eric Gillespie <*> epg@pretzelnet.org
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: Problem with r7950
Posted by Eric Gillespie <ep...@pretzelnet.org>.
Greg Hudson <gh...@MIT.EDU> writes:
> Probably, yeah. It's a little dicey, in that we have to just
> know what functions are in the path of svn_error_wrap_apr, but
> I can't think of a better answer.
OK, i'll fix it. Test suite running now...
--
Eric Gillespie <*> epg@pretzelnet.org
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: Problem with r7950
Posted by Greg Hudson <gh...@MIT.EDU>.
On Fri, 2004-01-16 at 09:55, Eric Gillespie wrote:
> What to do about it though? Before r7950 svn_error_createf was
> used here, which doesn't have the problem. But is that the right
> answer?
Probably, yeah. It's a little dicey, in that we have to just know what
functions are in the path of svn_error_wrap_apr, but I can't think of a
better answer.
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org