Re: [PATCH] notes: Allow treeish expressions as notes ref

From: Mike Hommey <mh@glandium.org>
Date: Oct 8, 2015, 02:50 UTC
Message-ID: <20151008025018.GA29722@glandium.org>
In-Reply-To: <xmqqzj30yq03.fsf@gitster.dls.corp.google.com>

On Mon, Jul 13, 2015 at 09:35:40AM -0700, Junio C Hamano wrote:

Show 14 quoted lines

> What is interesting is to think what should happen when amlog@{1}
> is given.  The version with your patch gives the same output as we
> saw earlier, because amlog@{1} tree exists, but does not know about
> your patch (yet), so does not add "Notes" section, which makes sense
> by itself.
> 
> But we cannot tell if amlog@{1} was somehow malformed or it was OK
> and there was no notes on the commit.
> 
> We probably should do a few more things:
> 
>  - Make sure that we show "there is no such tree-ish, no way to look
>    up any note to any commit from there" and "I understood the tree
>    you gave me, but there is no note for that commit" differently.

After refreshing the patch against current "next", it appears that there is such a distinction:

$ ./git-log --notes=fdsfgsfdg HEAD^! --pretty=short
warning: notes ref refs/notes/fdsfgsfdg is invalid
commit e5b68b2e879608d881c2e3600ce84962fcdefc88
Author: Mike Hommey <mh@glandium.org>

    notes: allow treeish expressions as notes ref

$ ./git-log --notes=foo HEAD^! --pretty=short
commit e5b68b2e879608d881c2e3600ce84962fcdefc88
Author: Mike Hommey <mh@glandium.org>

    notes: allow treeish expressions as notes ref

(I do have a refs/notes/foo ref)

>  - Decide if we want to "fail" the operation when the notes tree
>    given by the user is not even a tree-ish or just "warn" and keep
>    going.  And do so consistently.

Currently, it warns and doesn't fail.

Cheers,

Mike