Garrett Rooney wrote:
>
> In svn_ra_dav__get_locks, there's a comment that says if the path
> doesn't exist in HEAD we should error out, but that's the only place
> it's mentioned. The description of the function in svn_ra.h doesn't
> mention it, and neither the ra_local or ra_svn versions have the same
> comment or the same behavior.
>
> I've got a patch that "corrects" the ra_dav behavior by looking for
> the 404 error and simply returning no locks in that case. It seems to
> work, but I just wanted to make sure I'm not papering over any
> potential problems by catching that error. Anyway, the patch is
> attached, and if anyone could give me a "yeah, that seems sane" I'd
> appreciate it.
1) It's unusual to a function that returns information on a path not to return
an error if the path doesn't exist, so if that's how we want it to behave its
doc string should explicitly mention it.
2) I thought it WAS possible to have locks on non-existent paths (called
"reserving" a path). Even if our client doesn't allow creating such a lock,
others might, so we should still look for them.
- Julian
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Sun Nov 13 23:44:52 2005