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

Re: Problem with svn status -u via http

From: Daniel Rall <dlr_at_collab.net>
Date: 2005-12-01 22:43:04 CET

On Thu, 01 Dec 2005, Daniel Rall wrote:

> On Thu, 01 Dec 2005, Daniel Rall wrote:
>
> > On Thu, 01 Dec 2005, Mark Phippard wrote:
> > ...
> > > > On 12/1/05, Daniel Rall <dlr@collab.net> wrote:
> > > > > > I debugged a command line session to rule out problems with JavaHL
> > > > > > and saw that status.c's change_file_prop receives this odd looking
> > > > > > name value pair:
> > > > > >
> > > > > > + name 0x00d5fac8 "svn:entry:committed-rev"
> > > > > > - value 0x0012f510
> > > > > > + data 0x00d5fa88 "/svn13X/13TR/!svn/ver/9/Dir%20A/Doc%20A.txt9"
> > > > > > len 0x0000002c
> ...
> > > > > "data" looks like it's pointing to a mod_dav_svn URI which has been
> > > > > URL-encoded, but the "9" at the end of the string looks especially
> > > > > suspicious. Memory management problem?
> ...
> > We've got some leftover cruft before the 9. We need to
> > determine whether the server is responding with that (or something
> > else invalid), or whether the client is mishandling the server's
> > response.
>
> The latter is seeming more and more likely, unless this area is
> already significantly different between trunk and 1.3.

And indeed, that was the problem (also reproducible in trunk). I've
committed the following fix to libsvn_ra_dav:

--- subversion/libsvn_ra_dav/fetch.c (revision 17588)
+++ subversion/libsvn_ra_dav/fetch.c (revision 17589)
@@ -2747,14 +2747,15 @@
       break;

     case ELEM_href:
+ if (rb->fetch_content)
+ /* record the href that we just found */
+ svn_ra_dav__copy_href(rb->href, rb->cdata_accum->data);
+ svn_stringbuf_setempty(rb->cdata_accum);
+
       /* do nothing if we aren't fetching content. */
       if (!rb->fetch_content)
         break;

- /* record the href that we just found */
- svn_ra_dav__copy_href(rb->href, rb->cdata_accum->data);
- svn_stringbuf_setempty(rb->cdata_accum);
-
       /* if we're within a <resource> tag, then just call the generic
          RA set_wcprop_callback directly; no need to use the
          update-editor. */

I've proposed this non-destabilizing change for backport to 1.3. (It
will not require a restart of the soak.)

-- 
Daniel Rall
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Thu Dec 1 22:53:02 2005

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