From: Nguyen Thai Ngoc Duy <pclouds@gmail.com>
Date: Sat, 31 Jul 2010 03:12:59 GMT
Subject: Re: [PATCH 2/3] unpack-trees.c: Do not check ce_stage in will_have_skip_worktree()
Message-ID: <AANLkTi=K1Q+BkMvTPxKCo_VFgWY-ZLRznFcBEN2qQ3Yk@mail.gmail.com>
In-Reply-To: <20100731021144.GA906@burratino>

2010/7/31 Jonathan Nieder <jrnieder@gmail.com>:
> Nguyễn Thái Ngọc Duy wrote:
>
>> The idea of sparse checkout is conflicted entries should always stay
>> in worktree, regardless $GIT_DIR/info/sparse-checkout. Therefore,
>> ce_stage(ce) usually means no CE_SKIP_WORKTREE. This is true when all
>> entries have been merged into the index, and identical staged entries
>> collapsed.
>>
>> However, will_have_skip_worktree() since f1f523e (unpack-trees():
>> ignore worktree check outside checkout area) is also used earlier in
>> verify_* functions, when entries have not been merged to index yet
>> and ce_stage() may not be zero. Checking ce_stage() then can make
>> unnecessary verification on entries outside checkout area and error out.
>
> s/make/provoke/?
>
> So: conflicts from unmerged index entries are supposed to be kept in
> the work tree; but unpack_trees() is checking too early and seeing
> conflicts where there are none.  Do I understand correctly?

Correct.
-- 
Duy