Re: [PATCH] messages: avoid SHA-1 in end-user facing messages

From: Junio C Hamano <gitster@pobox.com>
Date: Aug 14, 2020, 04:45 UTC
Message-ID: <xmqqlfihzvhu.fsf@gitster.c.googlers.com>
In-Reply-To: <ee61b773-91c0-4b30-1d5b-e941bba03867@gmail.com>

Derrick Stolee <stolee@gmail.com> writes:

Show 6 quoted lines

>> N_("Show blank SHA-1 for boundary commits (Default: off)")),
>> N_("Blank object names of boundary commits (Default: off)")),
>
> Is there a reason you dropped "show" here? Perhaps your
> intention is to use "blank" as a verb, but it read a bit
> awkwardly to me.

You guessed my intention correctly. By using the word "blank" as a verb (i.e. to fill the space, which should ordinarily contain something meaningful, with whitespace instead), I can shorten "Show blank" to compensate the lengthening of the message caused by replacing "SHA-1" with "object name".

It also is more correct. There is no "blank SHA-1"; blank cannot be a valid SHA-1 hash value. We are showing blank instead of SHA-1, and I found that "blank" as a verb exactly has that meaning, which was convenient.

The above is *NOT* to defend the choice of the exact phasing I used, but to explain the criteria we need to use to come up with a better alternative (in other words, why "show blank object name for..." is a better replacement).

A better phrasing to replace it is of course welcome.

Thanks.