From: Richard Hansen <hansenr@google.com>
Date: Wed, 11 Jan 2017 18:36:41 GMT
Subject: Re: [PATCH 2/2] diff: document the pattern format for diff.orderFile
Message-ID: <2fb4296d-f084-4a76-44f3-7dc7d7cca7b1@google.com>
In-Reply-To: <xmqqwpe1o43k.fsf@gitster.mtv.corp.google.com>

On 2017-01-11 13:15, Junio C Hamano wrote:
> Richard Hansen <hansenr@google.com> writes:
>
>> On 2017-01-10 21:46, Junio C Hamano wrote:
>>> Richard Hansen <hansenr@google.com> writes:
>>>
>>>> I was looking at the code to see how the two file formats differed and
>>>> noticed that match_order() doesn't set the WM_PATHNAME flag when it
>>>> calls wildmatch().  That's unintentional (a bug), right?
>>>
>>> It has been that way from day one IIRC even before we introduced
>>> wildmatch()---IOW it may be intentional that the current code that
>>> uses wildmatch() does not use WM_PATHNAME.
>>
>> You are the original author (af5323e027 2005-05-30).  Do you remember
>> what your intention was?
>
> Yes.
>
> Back then we didn't even have wildmatch(), and used fnmatch()
> instead, so forcing FNM_PATHNAME would have meant that people
> wouldn't be able to say "foo*bar" to match "foo/other/bar"; with
> wildmatch, "foo**bar" lets you defeat WM_PATHNAME so having
> WM_PATHNAME always in effect is less of an issue, but with
> fnmatch(), having FNM_PATHNAME always in effect has a lot of
> downside.

Ah, that makes sense.

>
> I'd expect that orderfile people have today will be broken and
> require tweaking if you switched WM_PATHNAME on.

OK, so we don't want to turn on WM_PATHNAME unless we do it for a new 
major version.

I'll do another re-roll and document the non-WM_PATHNAME behavior. 
Perhaps I'll encourage users to prefer ** over * if they want to match 
slash (even though they are equivalent) so that migration is easier if 
we ever do turn on WM_PATHNAME.

-Richard