Hi!

On Mon, 18 Jun 2012 12:23:47 +0200, Arnaud Charlet <char...@adacore.com> wrote:
> > Here is a patch that adds a basic GCC Ada/GNAT configuration for x86
> > GNU/Hurd.  This has originally been created by Svante Signell for use
> > with Debian's GCC 4.6 package, and now I have ported it to trunk, and
> > also applied some further changes.
> > 
> > 2012-06-16  Svante Signell  <svante.sign...@telia.com>
> >         Thomas Schwinge  <tho...@codesourcery.com>
> > 
> > gcc/ada/
> >     * gcc-interface/Makefile.in: Add configuration for x86 GNU/Hurd.
> >     * gsocket.h [__GNU__] (Has_Sockaddr_Len): #define to 1.
> >     * s-oscons-tmplt.c: Add checks for __GNU__ next to several __linux__
> >     and linux conditionals.
> >     * terminals.c [__GNU__]: Don't #include <termio.h>.  #define USE_GETPT.
> >     * s-osinte-gnu.ads: New file.
> >     * s-taprop-gnu.adb: Likewise.
> 
> Well, the main question is: who is volunteering to maintain this code
> and fix it when things break?

I'm available; I also maintain other GNU/Hurd toolchain bits.

> Also, is s-taprop-gnu.adb really necessary? Assuming GNU Hurd provides
> a posix pthread interface, then you should be using s-taprop-posix.adb,
> NOT s-taprop-gnu.adb

Thanks for the suggestion, I'll look into that.

> Another comment:
> 
> > --- a/gcc/ada/s-oscons-tmplt.c
> > +++ b/gcc/ada/s-oscons-tmplt.c
> > @@ -80,7 +80,7 @@ pragma Style_Checks ("M32766");
> >  
> >  /* Feature macro definitions */
> >  
> > -#if defined (__linux__) && !defined (_XOPEN_SOURCE)
> > +#if (defined (__linux__) || defined (__GNU__)) && !defined
> > (_XOPEN_SOURCE)
> >  /** For Linux _XOPEN_SOURCE must be defined, otherwise IOV_MAX is not
> >  defined
> >   **/
> >  #define _XOPEN_SOURCE 500
> 
> You need to update the comment here, since the section so far only
> applied to GNU/Linux and not GNU/Hurd.

In fact, should that perhaps (unverified) simply say »For glibc,
_XOPEN_SOURCE must be defined [...]« -- or is this code meant to be
usable on GNU/Linux with a C library different from glibc?

> So assuming the above comment update, my main objection is to introduce
> a s-taprop-gnu.adb which will cause maintenance troubles.
> 
> If you use s-taprop-posix.adb instead and you or someone else volunteers
> for maintaining these files, then this could work.

Ack, thanks.


Grüße,
 Thomas

Attachment: pgp2Vwx2JeoZB.pgp
Description: PGP signature

Reply via email to