Re: [PATCH v4 3/3] diff --no-index: support limiting by pathspec

From: Jacob Keller <jacob.e.keller@intel.com>
Date: Jun 4, 2025, 17:22 UTC
Message-ID: <a5bc1851-e80b-4795-8aaf-46984b0cdfb8@intel.com>
In-Reply-To: <374BC043-8FB8-4052-BDE7-6BAE7F182994@gmail.com>

On 6/3/2025 7:37 PM, Ben Knoble wrote:

Show 18 quoted lines

> Actually, one comment :)
> 
>> Le 21 mai 2025 à 19:29, Jacob Keller <jacob.e.keller@intel.com> a écrit :
>>
>> From: Jacob Keller <jacob.keller@gmail.com>
>>
>> The --no-index option of git-diff enables using the diff machinery from
>> git while operating outside of a repository. This mode of git diff is
>> able to compare directories and produce a diff of their contents.
>>
>> When operating git diff in a repository, git has the notion of
>> "pathspecs" which can specify which files to compare. In particular,
>> when using git to diff two trees, you might invoke:
>>
>> $ git diff-tree -r <treeish1> <treeish2>.
> 
> I do find it slightly confusing that this series and in particular this patch is all about git-diff(1), but the only example is about git-diff-tree(1). It’s not the best example to me, esp. since it doesn’t actually use the pathspec machinery (deferring that to prose only). But I get the gist, so not really an issue.
>

Fair point. I think my brain wires got crossed as I was thinking of git diff-tree as an example for why pathspecs always come at the end.

> Rereading a bit, it seems this message goes to lengths to teach readers about pathspecs for git-diff here; perhaps we can simplify those parts and assume the reader is familiar enough with the details to understand the implications of « no-index mode doesn’t support pathspecs to limit comparison »?
>

Yea, we could probably simplify this. I usually try to be verbose with my "this is the reasoning for why to do this" and give background.. but that can sometimes end up being too much.

Regardless, the example probably should use pathspecs if we keep it..

> Nit: Should the diff-tree command end with a period?
>

It should not.