Re: textconv not invoked when viewing merge commit

From: Junio C Hamano <gitster@pobox.com>
Date: Apr 14, 2011, 20:06 UTC
Message-ID: <7v8vvcha2s.fsf@alter.siamese.dyndns.org>
In-Reply-To: <7vd3kohb5n.fsf@alter.siamese.dyndns.org>

Junio C Hamano <gitster@pobox.com> writes:

Show 15 quoted lines

> Junio C Hamano <gitster@pobox.com> writes:
>
>> Jeff King <peff@peff.net> writes:
>>
>>> We just dump the binary goo all over the terminal. So I think the whole
>>> combined-diff code path needs to learn how to handle binaries properly.
>>
>> How would you show multi-way diffs for binary files?
>>
>> It would probably be sufficient to say "binary files differ" at the
>> beginning of the patch-combining codepath of the combined diff, which
>> would at least keep the --raw -c/--cc output working.
>
> In other words, I suspect that the only places you need to touch in the
> existing codepath would be these places.

In the "here are the places" patch, I changed combine_diff() to return "is this binary?" and made show_patch_diff() to give just a single "path is binary", but I suspect that it would be simpler not try to be too nice about binary like that.

Instead, I think we should just use "Binary blob $SHA-1\n" as if that is the textconv of a binary file without textconv filter. That would certainly make the code much simpler, and more importantly, the output would become more pleasant. We would show something like:

    - Binary blob bc3c57058faba66f6a7a947e1e9642f47053b5bb
     -Binary blob 536e55524db72bd2acf175208aef4f3dfc148d42
    ++Binary blob 67cfeb2016b24df1cb406c18145efd399f6a1792

if we did so.

When showing the working-tree version, we obviously do not have the blob object name yet, so in such a case, we can say "Binary blob", or just "Binary".