do not stack on unmirrored branches

[Robert Collins]

Launchpad should not tell bzr to stack on a branch that doesn't have a
local mirror.

This affects the control file generation, and also limitting
type-changes for branches that are stackedupon.

 affects launchpad-bazaar

I'm tempted to say this is critical...

11:16 < johnf> when checking out lp:~bzr/bzr-svn/beta-ppa-gutsy
11:16 < johnf> bzr: ERROR: Not a branch: "bzr
+ssh://bazaar.launchpad.net/~jelmer/bzr-svn/0.5/"
11:17 < johnf> launchpad says it's stacked on
lp:~jelmer/bzr-svn/0.5-mirrored
11:18 < jelmer> hmm, that's because lp:~jelmer/bzr-svn/0.5/ is now a
"remote branch"
11:18 < jelmer> it used to be a "mirror branch", mirrorring the URL
that's now mirrorred by lp:~jelmer/bzr-svn/0.5-mirrored

[Jonathan Lange (jml)]

I think this bug is mis-filed.

Did the beta-ppa-gutsy branch get pushed up after jelmer made the branch 'remote'? If so, this bug is filed correctly. If not, the actual bug is that we allow stacked-on branches to be turned into remote branches.

[Robert Collins (lifeless)]

On Sat, 2009-03-21 at 09:21 +0000, Jonathan Lange wrote:
> I think this bug is mis-filed.
>
> Did the beta-ppa-gutsy branch get pushed up after jelmer made the branch
> 'remote'?

The state in lp of the involved branches is no longer what it was when I
filed it - I asked jelmer to change things so the branch worked again :P

> If so, this bug is filed correctly. If not, the actual bug is
> that we allow stacked-on branches to be turned into remote branches.

There are two issues which I thought I gathered in the report: turning
stacked-on branches into remote ones; and secondly (not possible now,
but in 1.14 frame) having a remote branch as the default stacking
branch.

-Rob

[Jonathan Lange (jml)]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Sat, Mar 21, 2009 at 8:43 PM, Robert Collins wrote:
> On Sat, 2009-03-21 at 09:21 +0000, Jonathan Lange wrote:
>> I think this bug is mis-filed.
>>
>> Did the beta-ppa-gutsy branch get pushed up after jelmer made the branch
>> 'remote'?
>
> The state in lp of the involved branches is no longer what it was when I
> filed it - I asked jelmer to change things so the branch worked again :P
>

This doesn't answer my question.

>> If so, this bug is filed correctly. If not, the actual bug is
>> that we allow stacked-on branches to be turned into remote branches.
>
> There are two issues which I thought I gathered in the report: turning
> stacked-on branches into remote ones; and secondly (not possible now,
> but in 1.14 frame) having a remote branch as the default stacking
> branch.
>

That brings the count of issues in this report up to three: the two
you mention here, and having the default stacked-on branch set to a
remote branch.

jml
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (GNU/Linux)
Comment: Use GnuPG with Firefox : http://getfiregpg.org (Version: 0.7.5)

iEYEARECAAYFAknEul0ACgkQX+0xPtMTonyiTQCfSKm3kTyf2epJ4yE1mtHOuBuU
0HAAnj44mq6IoOcxTYWOFQLacvp3F7y9
=Xenn
-----END PGP SIGNATURE-----

[Robert Collins (lifeless)]

On Sat, 2009-03-21 at 09:58 +0000, Jonathan Lange wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> On Sat, Mar 21, 2009 at 8:43 PM, Robert Collins wrote:
> > On Sat, 2009-03-21 at 09:21 +0000, Jonathan Lange wrote:
> >> I think this bug is mis-filed.
> >>
> >> Did the beta-ppa-gutsy branch get pushed up after jelmer made the branch
> >> 'remote'?
> >
> > The state in lp of the involved branches is no longer what it was when I
> > filed it - I asked jelmer to change things so the branch worked again :P
> >
>
> This doesn't answer my question.

'No'.

> >> If so, this bug is filed correctly. If not, the actual bug is
> >> that we allow stacked-on branches to be turned into remote branches.
> >
> > There are two issues which I thought I gathered in the report: turning
> > stacked-on branches into remote ones; and secondly (not possible now,
> > but in 1.14 frame) having a remote branch as the default stacking
> > branch.
> >
>
> That brings the count of issues in this report up to three: the two
> you mention here, and having the default stacked-on branch set to a
> remote branch.

isn't "having the default stacked-on branch set to a remote branch." ==
"having a remote branch as the default stacking branch."

-Rob

[Jonathan Lange (jml)]

Possibly fixed by lp:~jml/launchpad/package-stacking

[Jonathan Lange (jml)]

The above mentioned branch fixes the bug -- we don't let remote branches be recommended as default stacked-on branches. We never recommended that unmirrored branches be stacked on, so the bug title is incorrect. Bug 348732 is filed to track the issue of changing an existing stacked-on branch to be remote.