[ruby-core:115407] Re: Ruby 3.2.2 - rbconfig.rb's MAKEFILE_CONFIG

Hello,

Where do you see such code,
https://github.com/ruby/ruby/blob/master/lib/mkmf.rb doesn't seem to have
it.

In any case, it's not a good idea to use `CONFIG` in `extconf.rb`, use
`RbConfig::CONFIG` instead (like everywhere else), that's always expanded.

Benoit

On Fri, Nov 17, 2023 at 2:48=E2=80=AFAM Jay Mav via ruby-core <
[e-mail]> wrote:

> Hello Ruby Dev Team,
> With Ruby 3.2.2, gem install fails because in rbconfig.rb,
> MAKEFILE_CONFIG's $(target_os) is not expanded (whereas CONFIG's is
> expanded correctly).
>
> The inline doc in mkconfig.rb states that the caller can expand
> MAKEFILE_CONFIG's values by calling RbConfig::expand().
>
> However, ruby-shadow gem's Makefile generator -- extconf.rb -- does not
> call expand and directly uses mkmf.rb's CONFIG instance (which is
> RbConfig::MAKEFILE_CONFIG), because of which $(target_os) does not get
> expanded.
> This causes ruby-shadow gem install to fail with Ruby 3.2.2.
>
> Would the Ruby Dev Team accept the following proposal to fix mkconfig.rb,
> so that the generated rbconfig.rb instantiates MAKEFILE_CONFIG AFTER
> CONFIG's values have been expanded?
>
>   # Existing code...
>   CONFIG.each_value do |val|
>     RbConfig::expand(val)
>   end
>   # Move instantiation of MAKEFILE_CONFIG after expanding CONFIG's values=
.
>   MAKEFILE_CONFIG =3D {}
>   CONFIG.each{|k,v| MAKEFILE_CONFIG[k] =3D v.dup}
>
> If we fix it in mkconfig.rb, we can avoid callers having to explicitly
> call expand for MAKEFILE_CONFIG's values.
>
> Thanks.
>  ______________________________________________
>  ruby-core mailing list -- [e-mail]
>  To unsubscribe send an email to [e-mail]
>  ruby-core info --
> https://ml.ruby-lang.org/mailman3/postorius/lists/ruby-core.ml.ruby-lang.=
org/
 ______________________________________________
 ruby-core mailing list -- [e-mail]
 To unsubscribe send an email to [e-mail]
 ruby-core info -- https://ml.ruby-lang.org/mailman3/postorius/lists/ruby-core.ml.ruby-lang.org/