[FFmpeg-devel] [PATCH] lzo build system checks

[Diego Biurrun]

On Thu, Apr 09, 2009 at 04:36:41PM +0200, Reimar D?ffinger wrote:
> On Thu, Apr 09, 2009 at 04:25:36PM +0200, Diego Biurrun wrote:
> > --- configure	(revision 18382)
> > +++ configure	(working copy)
> > @@ -895,6 +895,7 @@
> >      lrintf
> > +    lzo1x_999_compress
> >      machine_ioctl_bt848_h
> > @@ -1970,6 +1971,7 @@
> >  check_func_headers io.h setmode
> > +check_func_headers lzo/lzo1x.h lzo1x_999_compress
> >  check_func_headers windows.h GetProcessTimes
> >  
> > --- libavutil/Makefile	(revision 18382)
> > +++ libavutil/Makefile	(working copy)
> > @@ -42,6 +42,7 @@
> >  
> >  TESTPROGS = adler32 aes base64 crc des lls md5 pca sha1 softfloat tree
> > +TESTPROGS-$(HAVE_LZO1X_999_COMPRESS) += lzo
> 
> Looks right to me.

I committed the build system part.

> > --- libavutil/lzo.c	(revision 18382)
> > +++ libavutil/lzo.c	(working copy)
> > @@ -239,6 +239,10 @@
> >  #include "log.h"
> >  #define MAXSZ (10*1024*1024)
> > +
> > +//#define HAVE_LZO1X_DECOMPRESS_SAFE 1
> > +//#define HAVE_LZO1X_DECOMPRESS 1
> 
> I meant those would have to be
> #define HAVE_LZO1X_DECOMPRESS_SAFE 0
> #define HAVE_LZO1X_DECOMPRESS 0

They are 0 by default if undefined, so uncommenting one of the above
lines would turn them on.

> But I don't really like it since if someone changes them, lzo-test will
> no longer test our lzo implementation, thus I prefer if they have to be
> defined on the command-line (which is also the reason for using
> defined/undefined instead of 0/1).
> And while I agree the names are probably bad I think yours are worse, I
> would propose using e.g.
> TEST_LIBLZO or BENCHMARK_LIBLZO if you consider those any better.

I painted the bikeshed pink according to your taste ;)

Diego