Re: [Patch] MinGW support for Glib-Object-Introspection

["Martin Schlemmer" <Martin Schlemmer nwu ac za>]

> > > On 2012/06/29 at 10:55 PM, Torsten Schoenfeld <kaffeetisch gmx de> wrote:
> The patch is looking pretty good now.  A few remaining things:
>
> On 27.06.2012 15:01, Martin Schlemmer wrote:
> > - Add OPTIMIZE to CFLAGS for the tests, fixing problems with some Glib's use
> >    of atomic intrinsics and not specifying a suitable -march depending on
> >    platform.
>
> Hrm, so every MSWin32 user would need to pass a correct OPTIMIZE flag to 
> Makefile.PL.  Is there no better way?  Why does glib's pkg-config file 
> not list the necessary cflags?

I will have to take a look - with 2.28.x it was not needed, definitely needed with
2.30.x, but unsure about 2.32.x (where I know glib's configure at least enforce it).

> > - The setting of PATH for the tests works with Dmake now at least.
>
> Do you have the possibility to test with nmake/cl?
>
> > +      # XXX: G-O-I defaults to CC=cc
> > +      $ENV{CC} = 'gcc' if (!defined $ENV{CC});
>
> On Unix, people usually symlink 'cc' to their preferred compiler, so 
> that using 'cc' in code is fine.  Does it not work like this on MSWin32?

For both MinGW32 and MinGW64 there is by default no 'cc' setup, so I think the
best way will be to go as suggested by Sergei:

$ENV{CC} = 'gcc' unless exists $ENV{CC};

Question is if this should also be applied to the rest of the 'gcc' calls?


Regards,
Martin



Vrywaringsklousule / Disclaimer:  http://www.nwu.ac.za/it/gov-man/disclaimer.html