[PATCH 2 of 2] bundlerepo: fix inconsistency of parsed and internal name (issue #821)

[Peter Arrenbrecht]

Hmm. I've tried again, but I cannot reproduce the findings I mentioned
in the original bug report about repo+bundle breaking a lot of tests.
Using bundle+repo would break a test now, but that is a test I
introduced myself a while ago when fixing the `hg -R repo pull bundle`
stuff (issue 820).

Nevertheless, the current patches _are_ of the form
bundle://path/to/repo+/path/to/bundlename and pass all tests. I just
updated to hg+crew and ran them again (convert-* and no-symlinks were
skipped). They also add a dedicated test for this issue.

So I don't think I can explain other than by stipulating a confused
state of mind when I wrote the original report. :(

But I think the current patches are good. Any objections? Do you still
need me to resubmit?

-peo


On Feb 13, 2008 10:16 PM, Thomas Arendsen Hein <thomas at intevation.de> wrote:
> * Peter Arrenbrecht <peter.arrenbrecht at gmail.com> [20080210 19:29]:
>
> > On Feb 10, 2008 7:13 PM, Thomas Arendsen Hein <thomas at intevation.de> wrote:
> > > * Peter Arrenbrecht <peter.arrenbrecht at gmail.com> [20080210 08:15]:
> > > > +        if path:
> > > > +            self._url = 'bundle:' + path + '+' + bundlename
> > > > +        else:
> > > > +            self._url = 'bundle:' + bundlename
> > >
> > > Hmm, this is bundle://repo+bundle or bundle://bundle
> > >
> > > In issue821 you say "So I suspect bundle+repo is correct." and I'd
> > > say you're right there.
> > >
> > > Why did you change your mind?
> >
> > If I remember correctly, it is what tonfa told me on IRC. But that is
> > quite some time ago. Since now one else had taken an interest then,
> > I'd let it slip.
> >
> > Strangely enough, test-bundle did *not* fail with this change this
> > time. A quick look at the log shows no relevant changes. Either I had
> > tomatoes on my eyes then or now. In any case, the change does not
> > break any tests (at least those that weren't skipped on my box).
> >
> > Given this, I think the usage that people might have been using at the
> > API level (as in `hg log bundle:repo+bundle`) is a little more
> > important to keep stable than what was reported to hooks. I may be
> > mistaken there, though.
>
> As I just added to issue821, bundle://path/to/repo+/path/to/bundlename
> is correct. Please resubmit your patches with the "Strangely enough"
> things fixed or explained :)
>
>
> Thomas
>
> --
> thomas at intevation.de - http://intevation.de/~thomas/ - OpenPGP key: 0x5816791A
> Intevation GmbH, Osnabrueck - Register: Amtsgericht Osnabrueck, HR B 18998
> Geschaeftsfuehrer: Frank Koormann, Bernhard Reiter, Dr. Jan-Oliver Wagner
> _______________________________________________
> Mercurial-devel mailing list
> Mercurial-devel at selenic.com
> http://selenic.com/mailman/listinfo/mercurial-devel
>