[svn.haxx.se] · SVN Dev · SVN Users · SVN Org · TSVN Dev · TSVN Users · Subclipse Dev · Subclipse Users · this month's index

Re: [PATCH] fix for issue #2556: support working copies on the root of a (virtual) drive

From: Erik Huelsmann <ehuels_at_gmail.com>
Date: 2006-08-17 20:38:50 CEST

On 8/17/06, Erik Huelsmann <ehuels@gmail.com> wrote:
> > > In that case, could you provide 2 patches? It's best to
> > > separate conceptually different changes into separate commits.
> >
> > Ok, attached you'll find two patches.
>
> Thanks.
>
> > > > 1. Introduce 'X:/' as a syntax for a root folder, on Windows.
> >
> > This change is contained in issue-2556-wc-on-root.patch.txt.
>
> Regarding this patch: It's basically fine, but you skip testing for
> the first character of "X:\\"<nul>, but that means you may be checking
> past the end of the string. (this is in svn_path_is_root and further
> down in the patch [the last hunk]; in is_canonical, you do it
> correctly.)
> Isn't it so that the X should be in the A-Za-z set of characters?
> Should we explicitly test for that?

Ok, but I did find a problem: svn_path_is_root returns an int in your
patch, but it should return svn_boolean_t, since that's what it is (a
boolean).

I have no working copy here, so I'll ask you this question instead:
Does this change make it so that the parent of the root is also the
root? Or was that the defined behaviour already? (Your next patch is
based on that behaviour, that's why I ask...)

> > > > 2. Fix an issue where Subversion thinks a working copy on a root path
> > (both '/' and 'X:/') is switched.
> >
> > This change is contained in issue-2556-wc-switched.patch.txt.
>
> I haven't had time to look into this patch yet; I'll come back to you later.

I've looked at it now. It's a nice and localized change. But what
would need to happen so that the parent of the root was something that
returns 'unversioned' (parent_entry == NULL), so that this change
isn't even required?

What I mean is: currently, if a parent of an entry isn't versioned, we
already have code which determines that the current entry can't be
switched. My question is: can't we make some change somewhere so that
we can benefit from the fact that this behaviour is correctly coded
throughout the system? (Instead of introducing yet another edge
case...)

bye,

Erik.

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Thu Aug 17 20:39:52 2006

This is an archived mail posted to the Subversion Dev mailing list.

This site is subject to the Apache Privacy Policy and the Apache Public Forum Archive Policy.