[PATCH 1 of 1] manifest: add --local option to list current directory state (issue2140)

Benoit Boissinot bboissin at gmail.com
Fri Apr 16 04:41:04 CDT 2010


On Fri, Apr 16, 2010 at 11:13 AM, Adrian Buehlmann <adrian at cadifra.com> wrote:
> On 16.04.2010 11:01, Adrian Buehlmann wrote:
>> On 16.04.2010 08:14, Nicolas Dumazet wrote:
>>> # HG changeset patch
>>> # User Nicolas Dumazet <nicdumz.commits at gmail.com>
>>> # Date 1271394625 -32400
>>> # Node ID 98028d5691a751140ec32c641c0bf14c1703252a
>>> # Parent  342b20693f5b21fae0da8ce8198b99521ff7e868
>>> manifest: add --local option to list current directory state (issue2140)
>>>
>> Side question:
>> What is ctx.manifest() supposed to return if ctx is a workingctx?
>> Shouldn't this be 'not defined' (i.e. abort)?
>>

there a well defined semantics in the internal :)
>>
>> With this patch applied, it looks like hg manifest --local even emits
>> 'unknown' files. Which is quite surprising.

It's intentional, due to the way merge, etc. works. Unknown files are
part of the working directory.

>>
>> And using both --local and --debug emits some strange hashes for
>> these (see file AAAAAAA.txt in the example below)

it is hash+status, nullid is used for new and unknown files
>
> Huh? Just noticed that changed files seem to produce a longer hash
> (ReleaseNotes.txt was a modified file)
>
>> 9b1c34ca1f4c89975c571bb323a04c1ecf183d77 644   ReleaseProcedure.txt
>> 084c8f2733f4b9ebe9f930356e9105c5c9139301 644   TortoiseHgOverlayServer.py
>> c8a7f14ac0577df20c0d40b39e9963ce49c3ce3b 644   contrib/_hgtk
>> 708a24a64d7d919a077f4539aa9b76f422a9d4ae 644   contrib/bash_completion
>> adeabf3927c651dcd7d1ca814c50c11073dc6e1d 644   contrib/docdiff.py
>> ...
>
> I guess the manifest command isn't really prepared for this patch.

You did use --debug which exposes internals (and the hash construction
for workingctx is indeed internal details).

Cheers,

Benoit


More information about the Mercurial-devel mailing list