error pulling from symbolic repo

[TK Soh]

--- Matt Mackall <mpm at selenic.com> wrote:
> On Thu, Aug 11, 2005 at 07:18:20AM -0700, TK Soh wrote:
> > --- TK Soh <teekaysoh at yahoo.com> wrote:
> > > --- TK Soh <teekaysoh at yahoo.com> wrote:
> > > > --- Thomas Arendsen Hein <thomas at intevation.de> wrote:
> > > > > % hg paths
> > > > > default = selected
> > > > > selected = bos
> > > > > bos = http://www.serpentine.com/hg
> > > > > tah = http://hg.intevation.org/mercurial-tah/
> > > > > hg = http://selenic.com/hg
> > > > > 
> > > > > With your patch 'hg pull' would pull from subdirectory 'selected',
> > > > > while 'hg pull selected' would pull from Bryan.
> > > > > 
> > > > > So you want recursive lookup of paths, but here one has to make sure
> > > > > that no endless loop happens.
> > > > 
> > > > Agreed. Here's a quick hack on ui.expandpath(). Please comment.
> > > 
> > > I've made a patch out of this with a slight change:
> > 
> > Thomas,
> > 
> > Did you have a chance to look at this? I think this should be fixed
> > before Matt cut the 0.6c release.
> 
> I'm not sure I like this one. I'd rather store the looked up path at
> clone time in .hgrc. I think your first change did this. The recursion
> can break in unexpected ways otherwise.

With this patch, Hg already is storing the looked up (expanded) path in .hgrc.
It also has the advantage of being 'backward compatible', so we don't need to
go fix every affected clones.

My first change, as Thomas pointed out, already has its own problems. Of
course, I am happy enough if anyone find a better fix for the bug - while I am
trying to beef up my Python skill :)


__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com