[MPlayer-dev-eng] [PATCH] FreeBSD DVD Fix

[Reimar Döffinger]

<patch_ad:

> @@ -5909,7 +5856,7 @@
>    _stripbinaries=no
>  elif test -z "$CFLAGS" ; then
>    if test "$host_arch" != "mips" ; then
> -    CFLAGS="-O4 $_march $_mcpu -pipe -ffast-math -fomit-frame-pointer"
> +    CFLAGS="-O"
>    else
>      CFLAGS="-O4 $_march $_mcpu -ffast-math -fomit-frame-pointer"
>    fi
> @@ -5919,6 +5866,9 @@
>      _stripbinaries=no
>    fi
>  else
> +  if test "$_runtime_cpudetection" = no ; then
> +    CFLAGS="$CFLAGS $_march $_mcpu"
> +  fi
>    cat <<EOF

Are they serious about this? -O seems a bit too little, and at least -O2 
should not break most compilers (not to mention that you simply 
shouldn't use a broken compiler...)
And the $_runtime_cpudetection check seems completely useless to me, as 
the default for runtime_cpudetection is around line 801:
     # i686 is probably the most common CPU - optimize for it
     _mcpu="$cpuopt=i686"
     # at least i486 required, for bswap instruction
     _march="-march=i486"

What sense does it make to compile for 80386 as MPlayer won't run on it 
anyway?

patch-ae:
> +       @${BSD_INSTALL_MAN} DOCS/man/en/mplayer.1 $(MANDIR)/man1/mplayer.1
no support for other languages or done differently or just lazyness?

patch-libmpdemux-demux_rtp.cpp:
Just as Jan though, this is in CVS now.

patch-libvo-vo_md5.c:
CVS version tries both md5sum and md5

patch-vidix-drivers-Makefile:
Are the X11 headers really needed for all vidix drivers? For RADEON and 
RAGE128 it makes sense, as they link against the X11 libs, but the 
others...?

Greetings,
Reimar Döffinger