Re: *more* relocatable perl

"Smylers  " via perl5-porters <[email protected]> Tue, 21 Jul 2026 12:06:13 +0200 (CEST)
Newsgroups gmane.comp.lang.perl.perl5.porters
Message-ID <[email protected]>
Gianni Ceccarelli writes:

> I'm asking =C2=ABdoes it make sense to do for binary paths what we're
> already doing for library paths, in the core perl distribution?

Yes. It makes sense.

There have been many responses, and none of them have even hinted at a
reason it doesn't make sense.

And your proposal is a configuration option anyway, so it doesn't break
anything for people not using it.

My only question is: Should this actually be a new build option, or
should the new behaviour be triggered by the existing =E2=80=98make reloca=
table=E2=80=99
option.

While it's good to preserve exact backwards-compatibility, it's also
suboptimal to end up in a future where we have build options which are
effectively =E2=80=98make mostly relocatable=E2=80=99 and =E2=80=98additio=
nally make actually
relocatable=E2=80=99.

Is there a situation in which somebody wants the current relocatable
behaviour but would be disadvantaged by your proposed additional
relocatableness?

Smylers