Re: Two update notes

On Wed, Sep 05, 2001 at 11:13:01AM -0400, Kevin Pilch-Bisson wrote:
> On Tue, Sep 04, 2001 at 02:52:01PM -0500, cmpilato@collab.net wrote:
> > Kevin Pilch-Bisson <kevin@pilch-bisson.net> writes:
> >
> > > > Believe it or not, this is by design. I seem to remember Karl wanting
> > > > to break people of the habit of using `svn up' to revert changes to a
> > > > file (since we have `svn revert').
> > >
> > > Ahhh, that makes sense. Perhaps we should print a warning message to this
> > > effect? At least temporarily.
> >
> > Oops. Ben tells me we never actually implemented this policy. You
> > *should* be getting those files back (and in fact, I just watched Ben
> > do this using ra_local with no problem). Not sure what the issue is
> > here, then.
>
> Alright, the source of this seems to be in
> libsvn_wc/adm_crawler.c:svn_wc_crawl_revisions. We have:
>
> if (entry_kind == svn_node_dir)
> report_revisions
> else if ((entry->kind == svn_node_file) && (entry->revision != base_rev))
> reporter->set_path
>
> within report revisions itself we check the on disk state of the file, but
> we don't within svn_wc_crawl_revisions. I'll check it out more.

Here is a quick follow-up. With the following patch, we generate the same
report. However, things are still messed, because
libsvn_client/update.c:svn_client_update assume that PATH is either the
current directory, or one of its immediate children.

We need to fix it so that it can be a dir or file, anywhere.

-- 
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Kevin Pilch-Bisson                    http://www.pilch-bisson.net
     "Historically speaking, the presences of wheels in Unix
     has never precluded their reinvention." - Larry Wall
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~