Re: final fixup for 3.3.5-1

[Ben Collins]

On Sun, Sep 19, 1999 at 12:52:03AM -0400, Branden Robinson wrote:
> On Thu, Sep 16, 1999 at 09:24:17AM +0200, Sven LUTHER wrote:
> > --- xc/programs/Xserver/hw/xfree86/accel/mach64/mach64im.c.orig Wed Sep 
> > 15 17:01:53 1999
> > +++ xc/programs/Xserver/hw/xfree86/accel/mach64/mach64im.c  Wed Sep 15 
> > 17:03:55 1999
> > @@ -76,11 +76,12 @@
> >  unsigned int integer;
> >  unsigned char bytes[4];
> >  };
> > -
> > +/* Also defined in regmach64.h ...
> >  static __inline__ void regwbe(volatile unsigned long regindex, unsigned 
> > long regdata)
> >  {
> >  *(unsigned long *)(mach64MemReg + regindex) = regdata;
> >  }
> > +*/
> >  
> >  static __inline__ unsigned int bit_reverse32(unsigned int bits)
> >  {
> 
> I don't see the point of the above, regwbe is declared static anyway.

Get rid of the one in mach64im.c, but make the one in regmach64.h non-static 
and also
usable when __sparc__ is defined.

> I will apply this one as well.
> 
> Can someone tell me the whys and wherefores of "inline" versus
> "__inline__"?  The New Testament is silent on the subject.  Is it a GNU
> thing?

On SPARC, only __inline__ seems to be understood. I'm assuming the "inline" 
alone
is probably supposed to be defined somewhere to what the compiler needs:

#define inline __inline__

Not sure why it works in some places, but not others.

Ben

final fixup for 3.3.5-1

[Branden Robinson]

On Thu, Sep 16, 1999 at 09:24:17AM +0200, Sven LUTHER wrote:
> --- xc/programs/Xserver/hw/xfree86/accel/mach64/mach64im.c.orig   Wed Sep 
> 15 17:01:53 1999
> +++ xc/programs/Xserver/hw/xfree86/accel/mach64/mach64im.cWed Sep 15 
> 17:03:55 1999
> @@ -76,11 +76,12 @@
>  unsigned int integer;
>  unsigned char bytes[4];
>  };
> -
> +/* Also defined in regmach64.h ...
>  static __inline__ void regwbe(volatile unsigned long regindex, unsigned long 
> regdata)
>  {
>  *(unsigned long *)(mach64MemReg + regindex) = regdata;
>  }
> +*/
>  
>  static __inline__ unsigned int bit_reverse32(unsigned int bits)
>  {

I don't see the point of the above, regwbe is declared static anyway.

> --- xc/programs/Xserver/hw/xfree68/mach64/Imakefile.orig  Wed Sep 15 
> 18:24:42 1999
> +++ xc/programs/Xserver/hw/xfree68/mach64/Imakefile   Wed Sep 15 18:25:04 1999
> @@ -49,7 +49,7 @@
>  LinkSourceFile(mach64pntwn.c,../../xfree86/accel/mach64)
>  LinkSourceFile(mach64seg.c,../../xfree86/accel/mach64)
>  LinkSourceFile(mach64text.c,../../xfree86/accel/mach64)
> -LinkSourceFile(mach64util.c,../../xfree86/accel/mach64)
> +LinkSourceFile(mach64util.h,../../xfree86/accel/mach64)
>  LinkSourceFile(mach64win.c,../../xfree86/accel/mach64)
>  LinkSourceFile(regmach64.h,../../xfree86/accel/mach64)
>  

Yes, that obviously should be done and I am applying it.

> --- 
> /home/luther/Xfree/3.3/debian/xfree86-1-3.3.4/build-tree/xc/programs/Xserver/hw/xfree68/mach64/mach64util.h.orig
>   Thu Sep 16 07:34:58 1999
> +++ 
> /home/luther/Xfree/3.3/debian/xfree86-1-3.3.4/build-tree/xc/programs/Xserver/hw/xfree68/mach64/mach64util.h
>Thu Sep 16 07:38:08 1999
> @@ -143,6 +143,7 @@
>  
>  #endif /* __sparc__ */
>  
> +#ifndef __powerpc__
>  extern __inline__ void regw(unsigned int regindex, unsigned long regdata)
>  {
>  unsigned long appaddr;
> @@ -175,6 +176,7 @@
>  return (*(__volatile__ int *)appaddr);
>  #endif
>  }
> +#endif /* __powerpc__ */
>  
>  extern __inline__ void regwb(unsigned int regindex, unsigned char regdata)
>  {

I will apply this one as well.

Can someone tell me the whys and wherefores of "inline" versus
"__inline__"?  The New Testament is silent on the subject.  Is it a GNU
thing?

Also, for m68k I am disabling *all* framebuffer accelerated drivers.  There
are presently four: the NCR 77C32BLT, the IMS Twin Turbo, the ATI Mach64,
and the Permedia 2.

Christian, I am not sure I can get 3.3.5-1 building with *no* modifications
for you yet.  The only one I'm stuck on is the rules file fix, which
obviously cannot be done for everyone else.

However, I have a new rule called "environment" which people should run
before trying to build.  It will tell you a lot about how sane the build
environment will be.  Keep in mind that dpkg-buildpackage sets a lot of
variables on the debian/rules command like these days...DEB_HOST_ARCH et
al.

A version of 3.3.5-1 has already been built and BenC says it works great on
the SPARC.  So let's keep our fingers crossed.

I'll be starting a "final" build of 3.3.5-1 in a few minutes.  I will bring
the .debs home for testing and if nothing hideous happens this will be
uploaded to master.

-- 
G. Branden Robinson  |
Debian GNU/Linux |   The noble soul has reverence for itself.
[EMAIL PROTECTED]   |   -- Friedrich Nietzsche
cartoon.ecn.purdue.edu/~branden/ |


pgp3VmtiE4BeO.pgp
Description: PGP signature