Re: [PATCH 3/3] hook-list.h: add a generated list of hooks, like config-list.h

["Ævar Arnfjörð Bjarmason" <avarab@gmail.com>]

On Fri, Jun 18 2021, Eric Sunshine wrote:

> On Fri, Jun 18, 2021 at 1:06 PM SZEDER Gábor <szeder.dev@gmail.com> wrote:
>> On Thu, Jun 17, 2021 at 12:09:36PM +0200, Ævar Arnfjörð Bjarmason wrote:
>> > diff --git a/generate-hooklist.sh b/generate-hooklist.sh
>> > @@ -0,0 +1,24 @@
>> > +#!/bin/sh
>> > +
>> > +echo "/* Automatically generated by generate-hooklist.sh */"
>> > +
>> > +print_hook_list () {
>> > +     cat <<EOF
>> > +static const char *hook_name_list[] = {
>> > +EOF
>> > +     perl -ne '
>>
>> Why Perl?
>>
>> At the moment I can run 'make' and get a functioning git even when
>> Perl is not installed.  With this patch that wouldn't work anymore.
>>
>> Both 'generate-cmdlist.sh' and 'generate-configlist.sh' can process
>> the documentation into a header file with a long list in it without
>> resorting to Perl; I'm sure that hooks could be processed without Perl
>> as well.
>
> That's a good point and not an idle question. It wasn't all that long
> ago that I rewrote `generate-cmdlist` in Perl and got a complaint that
> the Git project was no longer buildable on FreeBSD[1], with the result
> that I ended up re-implementing it (again) in shell[2].
>
> [1]: https://lore.kernel.org/git/loom.20150814T171757-901@post.gmane.org/
> [2]: https://lore.kernel.org/git/1440365469-9928-1-git-send-email-sunshine@sunshineco.com/

I'm sympathetic to changing this from Perl, as can be seen from the "v1"
(as part of another series) I intially used an unportable grep flag for
it. I'll test and probably convert it to René's few-liner.

The [1] link is not it not being buildable on FreeBSD though, it's
someone understandably not understanding what NO_PERL=Y means, because
we really should call it PERL_ONLY_FOR_BUILDING_AND_TESTING or
something.

Aside from this series it would be useful to clarify that point, what do
we really think it means? I was under the impression that it meant that,
and I really shouldn't be giving it a second thought if we introduced a
perl dependency during the build or testing.