Re: [PATCH] bugreport: add 'seconds' to default outfile name

From: Junio C Hamano <gitster@pobox.com>
Date: Oct 14, 2023, 16:27 UTC
Message-ID: <xmqq4jitw4nk.fsf@gitster.g>
In-Reply-To: <20231014040101.8333-1-jacob@initialcommit.io>

Jacob Stopak <jacob@initialcommit.io> writes:

> Currently, git bugreport postfixes the default bugreport filename (and
> diagnostics zip filename if --diagnose is supplied) with the current
> calendar hour and minute values, assuming the -s flag is absent.

Is "postfix" a verb that is commonly understood? I would say "append" would be understood by more readers. Also, is "calendar" hour different from other kinds of hours, perhaps stopwatch hours and microwave-oven hours?

> If a user runs the bugreport command more than once within a calendar
> minute, a filename conflict with an existing file occurs and the program
> errors, since the new output filename was already used for the previous
> file.

This is totally expected and you made an excellent observation.

I personally do not think it is a problem, simply because a quality bug report that would capture information necessary to diagnose any issue concisely in a readable fashion would take at least 90 seconds or more to produce, though.

Instead of lengthening the filename for all files by 2 digits, the command can retry by adding say "+1", "+2", etc. after the failed filename to find a unique suffix within the same minute. It would mean that after writing git-bugreport-2023-10-14-0920.txt and you start another one without spending enough time, the new one may become git-bugreport-2023-10-14-0920+1.txt or something unique. It would be really unlikely that you would run out after failing to find a vacant single digit suffix nine times, i.e. trying "+9". It would also help preserve existing user's workflow, e.g. they may have written automation that assumes the down-to-minute format and it would keep working on their bug reports without breaking.