GCCSDK trunk status ?

Peter Naulls peter at chocky.org
Mon Jun 27 13:05:22 PDT 2005

In message <b85bb9814d.Jo at hobbes.bass-software.com>
          John Tytgat <John.Tytgat at aaug.net> wrote:

> In message <4e9b88814d.peter at chocky.org>
>           Peter Naulls <peter at chocky.org> wrote:
> > While we're at it, I'm going to remove !Clib from being packaged.  It's
> > only used by !LCC, and that takes its own copy interally.  !Clib
> > unhelpfully sets C$Path which conflicts with Acorn C/C++.   Unless
> > anyone has any objections of course.
> Maybe just a question where we are heading as stub recommendation for
> GCC module builders ? StubsG <URL:http://support.riscos.com/Support/Resources/Libraries/StubsG/index.html> ?
> Of course the Castle C:o.Stubs should work too but isn't accessible to
> everybody.

The GCC stubs theoretically contain fully module support.

> The nice thing about the StubsG release is that it also contains a
> CMunge 32-bit capable release.

I hadn't realised that.  We ought to note that.

> !Clib would be another candidate as stub with additional work but I don't
> see much advantage to create just another one (maybe apart from the
> riscosify() code but that's something which could be extracted from UnixLib
> and delivered separately).
> Or am I overlooking something ?

You're suggesting use of CLib with Norcroft in order to obtain the
riscosify stuff?  I think if you need it that badly, you can either
extra the code yourself or use GCC.

Peter Naulls - peter at chocky.org        | http://www.chocky.org/
RISC OS C Programming                  | http://www.riscos.info/c/

More information about the gcc mailing list