Re: [PATCH] git-cherry.txt: Change symbol marks in graph.

From: Junio C Hamano <gitster@pobox.com>
Date: Dec 2, 2010, 17:25 UTC
Message-ID: <7vy688rsry.fsf@alter.siamese.dyndns.org>
In-Reply-To: <m31v604nc0.fsf@localhost.localdomain>

Jakub Narebski <jnareb@gmail.com> writes:

Show 15 quoted lines

> Jari Aalto <jari.aalto@cante.net> writes:
>
>> From: Jari Aalto <jari.aalto@cante.net>
>> 
>> The common commit were marked with a minus sign (-), but that is
>> usually interpreted as something less or substracted. Use natural
>> equal sign (=). Commits that are not in upstream were marked with plus
>> sign (+) but a question mark (?) is visually a litle more striking
>> (erect) in context where all other signs are "flat". It also helps
>> visually impared to see difference between (* ... ?) as opposed to
>> similar signs (* ... +).
>
> Reading comprehension FAIL.  The graph in question uses '+' and '-' because
> that is what 'git cherry' *output( uses ('+' included, '-' excluded).  They
> are very natural in git-cherry output (diff-like).

Well, cherry's output is not "diff like". '+' is what you need to keep, as opposed to '-' that marks a commit that needs to be dropped from your side. If it were "diff like", '+' would have been spelled with a single ' ', but then commits on the other side you do not have need be included with '+' mark. However, 'git cherry' is not trying to produce a diff between branches; it does not matter what extra stuff the other side contains.

I however do agree with your conclusion; '=' is not "natural" in the context of 'git cherry' at all. "You need to drop this" comes more natural with '-'.

And obviously, the illustration in the documentation uses the same +/- that appear in the program's output.

So any change along the line of Jari's patch would be a dis-improvement.