Re: [PATCH 4/4] diff-lib: parallelize run_diff_files for submodules

From: Junio C Hamano <gitster@pobox.com>
Date: Sep 24, 2022, 20:17 UTC
Message-ID: <xmqqsfkgr0ht.fsf@gitster.g>
In-Reply-To: <220923.86o7v64ik0.gmgdl@evledraar.gmail.com>

Ævar Arnfjörð Bjarmason <avarab@gmail.com> writes:

Show 10 quoted lines

> On Thu, Sep 22 2022, Calvin Wan wrote:
>
>> +status.parallelSubmodules::
>> +	When linkgit:git-status[1] is run in a superproject with
>> +	submodules, a status subprocess is spawned for every submodule.
>> +	This option specifies the number of submodule status subprocesses
>> +	to run in parallel. If unset, it defaults to 1.
>
> Why do we default to 1, instead of e.g. grep.threads defaulting to the
> "cores available"?

I would imagine we would want to be able to say:

 - I do not trust the parallel mode yet, just use the single process
   method that we have always been using.

 - I do not know how many cores I have, just use a reasonable
   default parallelism.

 - I want to use N processes because I know better than auto-scaling
   based on num_cpus.

And the value of 1 would be a reasonable way to express the first one.