bug#54034: 29.0.50; Diff prettify broken for empty files

[Matthias Meulien]

Stefan Monnier <monnier@iro.umontreal.ca> writes:

>> Wouldn't it be safer to simply disable prettification of the "diff
>> header" when diff-buffer-type isn't equal to git?
>
> Could be, tho I think the current code ends up doing something similar.
>
> FWIW, I tend to prefer using the buffer's content rather than the value
> of a buffer-local variable (to the extend possible) when deciding
> whether a given diff header can be prettified.

The buffer-local variable is currently initialized from the buffer
content!

>
> [ After all, some diffs may be the result of running several commands,
>   some of which are Git but not all, so the "type" may be different for
>   different headers in the same buffer.  ]

Ok, I understand your point.

I can't see how a regex can distinguish diff output generated in
the context of a call to diff-buffers (where compared files are
temporary files and a prettified message should refer to the
corresponding buffer names) and a generic diff output where a user may
have used any diff option including --label for whatever reason.
-- 
Matthias