RE: svn commit: r1443705 - /subversion/trunk/subversion/libsvn_ra_serf/commit.c

> -----Original Message-----
> From: C. Michael Pilato [mailto:cmpilato_at_collab.net]
> Sent: vrijdag 8 februari 2013 15:22
> To: dev_at_subversion.apache.org
> Cc: Bert Huijben; commits_at_subversion.apache.org
> Subject: Re: svn commit: r1443705 -
> /subversion/trunk/subversion/libsvn_ra_serf/commit.c
>
> On 02/07/2013 05:23 PM, Bert Huijben wrote:
> > Could this be caused by the difference between redirecting to
> > /some/subdir and http://server/some/subdir ?
>
> RFC2616 states that the value of the Location header as an "absoluteURI".
> So, I *hope* that's not the difference.

As philip's mail confirmed this *was* the problem.

The RFC's defining a new header don't always tell what users do on their servers, nor what older apache httpd versions implemented.
(Recent versions fix the location header if a wrong value is supplied... older versions didn't)

See r1443906, which fixed the failure on the buildbot.

And I know that every common browser allows at least server root relative urls in this header, but I think even normal relative paths.

        Bert
Received on 2013-02-08 15:42:12 CET