cmpilato@collab.net writes:
> Michael Price <mprice@atl.lmco.com> writes:
>
> > cmpilato@collab.net wrote:
> > > Michael Price <mprice@atl.lmco.com> writes:
> > >
> > >>svn: Trying to use an unsupported feature
> > >>svn: Bad encoding option: prop's value isn't stored as UTF8.
> > >>
> > >>Those two lines above are popping up repeatedly in my tests. On
> > >>different platforms they pop up in different tests.
> > >>
> > >>I'm assuming this is a library issue of some sort as if it was an svn
> > >>code problem it would have been noticed/fixed before this.
> > >>
> > >>Where should I start looking?
> > > Do you have the log-encoding option set in your .subversion?
> >
> > I did on my Solaris box (which explained those failures) but do not
> > have it set on my FreeBSD box (which is still failing commit test 25
> > for the same reason.
>
> Actually, I'm not convinced that the code is accurate even when that
> option is set in your config file. The client binary sets the
> opt_state->encoding from the config option without regard to what
> operation is happening. So a command that wouldn't normally fail
> (like "svn pset name value target") is failing because the
> configuration option is making the code behave as if you'd done "svn
> pset --encoding some_encoding name value target", which is cause for
> complaint. That said, I dunno why we're complaining about the
> encoding anyway -- if the property operation doesn't require the
> encoding, then why not just ignore it?
Now I'm convinced, the current code is wrong. Am fixing now.
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Wed May 7 18:16:34 2003