What would it take to get a patch merged into gnulib for this issue?
Margaret's patch didn't gain much traction because it didn't solve the
problem for non-darwin platforms, but at the same time we also acknowledge
that there's no platform-agnostic way to do this. To me, it would be better
to accept a patch that improves the situation for one platform, with the
understanding that we'll have to keep adding special cases as more people
realize that the program_name approach doesn't work on their platform.

Or if program_name isn't used in a critical manner, we could just have an
#define that makes it "<unknown program name>" on non-glibc, non-darwin
platforms?

Thanks,
Dan

(apologies for the possibly weirdly formatted reply but I only just
subscribed to this list and don't have an email in my client to reply to)

On Tue, May 31, 2016 at 09:36:40AM +0000, Tobias Pflug wrote:
> >* Hi,*
> >
> >* quite some time ago there already was a discussion about the missing *
> >* program_name symbol on Darwin and Richard proposed a patch [1] that would *
> >* #define program_name as (((char **)*_NSGetArgv())[0]) but sadly the *
> >* discussion died down without any result.*
> >
> >* In NixOS we are currently facing this problem again and are looking for a *
> >* solution [2]. It would be great if this could be fixed inside gnulib 
> >itself. *
> >* I hope I can revive the discussion around this to move this topic forward *
> >* this time.*
> >
> >* Thanks and best regards,*
> >* Tobias*
> >
> >* [1] https://lists.gnu.org/archive/html/bug-gnulib/2015-02/msg00079.html 
> ><https://lists.gnu.org/archive/html/bug-gnulib/2015-02/msg00079.html>*
>
> I think Margaret is carrying that patch downstream for libguestfs now.
> It would be nice to get this working in gnulib.
>
> Rich.
>
> >* [2] https://github.com/NixOS/nixpkgs/issues/15849 
> ><https://github.com/NixOS/nixpkgs/issues/15849>*
>
>

Reply via email to