"bzr status" in a lightweight checkout

[Alexander Belchenko]

John Arbash Meinel пишет:
> Martin Pool wrote:
>> On 10 August 2010 03:10, Eli Zaretskii <eliz at gnu.org> wrote:
>>> In a lightweight checkout, is "bzr status" supposed to access the
>>> remote repository?  Or do lightweight checkouts have enough
>>> information locally to provide the status?
>>>
>>> I'm asking because someone complained on the Emacs development list
>>> that their modem LED indicates traffic when they bootstrap Emacs.
>>> Bootstrapping should not (AFAIK) do any bzr ops except "bzr status"
>>> when they visit or save files in a versioned directory.
>>>
>>> If it turns out that "bzr status" needs to access the repository in
>>> this case, we should probably fix Emacs to use some other technique to
>>> report status of versioned files, or maybe not at all, because
>>> accessing the network every time the user does an entirely local file
>>> operation is not nice, to say the least.
>> It probably does need to access the repository if you have pending
>> (uncommitted) merges, to show their revision data.  Otherwise, I don't
>> think it should.  If it does, please get a traceback (eg by blocking
>> access to the repository) and file a bug tagged {emacs performance}.
> 
> 
> It does because accessing a lightweight checkout dereferences is branch
> pointer, which requires opening up the Branch which is remote (which
> also opens up its Repository).

Does `bzr status` check for `tree revision is not up-to-date with
branch revision` condition?

> 
> I mentioned this a while back. ISTR that Robert didn't like lazy Branch
> objects because then they might be pointing to something invalid and you
> wouldn't know about it.
> 
> John
> =:->