[rfc] 'bzr missing' for checkout could use master branch for comparing histories

[Alexander Belchenko]

Erik Bågfors пишет:
> On 7/31/06, Alexander Belchenko <bialix at ukr.net> wrote:
>> Because I heavily use checkouts I wonder why 'bzr missing' works by
>> default with pull location only? I think there is sense to automatically
>> choose branch to compare histories:
>>
>> - if current branch is standalone or lightweight checkout then it will
>> use stored pull location (as and before);
>> - if current branch is heavyweight checkout then it will use master
>> branch location.
>>
>> This change will allows to user easily inspect what difference between
>> local and master branch. Especially when someone do local commits.
>>
>> What you think about?
> 
> I don't like it, because I often do this
> bzr branch $public_trunk project.trunk
> bzr branch project.trunk project.erik
> # hack in project.erik, commit etc
> # time to publish project.erik
> bzr push sftp://....../
> bzr bind sftp://....../
> 
> # get feedback... see if I need to merge trunk again
> bzr missing
> # Yes, it's needed
> bzr merge
> bzr ci -m 'merge trunk'
> 
> And it will automatically publish my changes.
> 
> So yes, I use heavy checkouts primarily to mirror my branches, but it
> works great!
> 
> For me it would be confusing if missing was going to a different
> location for heavy checkouts than for anything else.
> 
> I think that a heavy checkout should work like a lightweight checkout.
> I think we should be able to do "bzr update -n" or something like that
> for a dry-run update to print the missing revisions for both heavy and
> lightweight checkouts.

IIRC, in CVS 'update -n' show list of files that will be updated. Not 
revisions log.

What about syntax: 'bzr missing --master' ?
It's will be clear enough to specify that user want to inspect 
difference with master branch, not with pull location.

How about this?

--
Alexander