[PATCH] Windows testsuite fixes [1/4]

[Aaron Bentley]

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

Vincent Ladeuil wrote:
>>>>>>"aaron" == Aaron Bentley <aaron.bentley at utoronto.ca> writes:
>     aaron> Requests are marked superseded if a new request
>     aaron> includes the head revision from an old request in its
>     aaron> bundle.
> 
> Hmm, so I lost here. I rebuild the bundle by uncommit/shelve/take
> the unwanted part out/branch a fresh bzr/unshelve.

Yeah, but that's what the manual controls are for, so I marked it manually.

>     aaron> I think carrying the already-gathered +1 isn't
>     aaron> necessarily appropriate in the general case,
>     aaron> because the changes may have introduced a new problem
>     aaron> that negates the original +1.
> 
> Fully agreed.

On the other hand, it would be good to encourage the original voters to
re-examine it, somehow.

> In that particular case, I was afraid that, reverting only one
> change on top of just adding it, was asking for trouble for
> future merges.

Well, you would revert those changes, commit, un-revert (i.e. revert -r
- -2), commit, continue work.  That wouldn't have serious side effects,
unless the original branch had been merged and then altered by others.

It's not a big deal either way.  With uncommitting, there's a chance
that someone else has picked up the original (i.e. uncommitted) version.
When then submit it, it will slip in as part of their merge request.
But the reality is no one does that for other good reasons.

> Thanks for explaining bundle buggy's behavior,

No problem.  I'll make sure that's covered properly in the help section.

Aaron
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org

iD8DBQFFdIKC0F+nu1YWqI0RAu71AJwJ3VKw2iV43r/XxWhTjQHDLOhD6gCfRBbM
mluODTWZLJg9wr5pyEQHUfk=
=nphh
-----END PGP SIGNATURE-----