On Mon, Dec 19, 2005 at 01:12:58PM -0500, Paul Burba wrote:
> > > --- trunk/subversion/libsvn_wc/status.c (original)
> > > +++ trunk/subversion/libsvn_wc/status.c Mon Dec 19 09:00:09 2005
> > > @@ -1054,7 +1054,7 @@
> > > {
> > > struct dir_baton *b = baton;
> > > if (b->url)
> > > - statstruct->url = b->url;
> > > + statstruct->url = apr_pstrdup (pool, b->url);
> >
> > At this point, if b->url was NULL, statstruct->url will retain the same
> > value as it had from the call to svn_wc_status2(), and the same is true
> > of ood_last_cmt_author. Is this correct in all four cases (the two
> here,
> > and the two in the file_baton case), or should we be setting either of
> > the fields to NULL if the corresponding baton field is NULL?
>
> I don't think that is necessary since svn_wc_status2() calls
> assemble_status(), initializing ood_last_cmt_author and url to null.
>
Okay.. Ah, svn_wc_status2() doesn't populate any of the OOD information.
Fair enough, though it'd be nice if we documented that explicitly,
however obvious it might be.
I think you're wrong about the URL field though, since assemble_status()
will initialise that to entry->url, which, as far as I can see, should
always be populated. Perhaps b->url is always the same as
statstruct->url? (in which case this wouldn't matter)
> > [I could imagine that perhaps b->url is never expected to be NULL if
> > statstruct->url was non-NULL,
>
> Sorry, I don't follow you on that...did you mean that the other way
> around?
>
No, I meant it that way round, I'm fairly sure :-)
I was hypothesising that perhaps we were relying on an assumption that
b->url could only be NULL if statstruct->url was also NULL, and hence
the code shown above would be fine.
Regards,
Malcolm
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Mon Dec 19 21:10:36 2005