[FFmpeg-devel] [PATCH] small configure simplification

Aurelien Jacobs aurel
Sun Aug 26 18:41:47 CEST 2007


On Tue, 21 Aug 2007 00:46:44 +0200
Aurelien Jacobs <aurel at gnuage.org> wrote:

> On Mon, 20 Aug 2007 23:25:22 +0100
> M?ns Rullg?rd <mans at mansr.com> wrote:
> 
> > Aurelien Jacobs <aurel at gnuage.org> writes:
> > 
> > > Hi,
> > >
> > > The attached patch simplify configure using the new $COMPONENT_LIST.
> > > Ok to apply?
> > >
> > > Aurel
> > >
> > > Index: configure
> > > ===================================================================
> > > --- configure	(revision 10153)
> > > +++ configure	(working copy)
> > > @@ -995,11 +995,9 @@
> > >    --disable-protocols) disable $PROTOCOL_LIST
> > >    ;;
> > >    --enable-*=*|--disable-*=*)
> > > -  eval `echo "$opt" | sed 's/=/-/;s/--/action=/;s/-/ thing=/;s/-/ name=/'`
> > > -  case "$thing" in
> > > -      encoder|decoder|muxer|demuxer|parser|bsf|protocol) $action ${optval}_${thing} ;;
> > > -      *) die_unknown "$opt" ;;
> > > -  esac
> > > +  eval `echo "$opt" | sed 's/=/-/;s/--/action=/;s/-/ thing=/;s/-/s name=/'`
> > > +  is_in "$thing" $COMPONENT_LIST || die_unknown $opt
> > > +  $action ${optval}_${thing%s}
> > >    ;;
> > >    --enable-?*|--disable-?*)
> > >    eval `echo "$opt" | sed 's/--/action=/;s/-/ option=/;s/-/_/g'`
> > > @@ -1796,18 +1794,16 @@
> > >      echo 'asm (".align 3");' | check_cc && asmalign_pot="yes"
> > >  fi
> > >  
> > > -enabled_any $ENCODER_LIST  && enable encoders
> > > -enabled_any $DECODER_LIST  && enable decoders
> > > -enabled_any $MUXER_LIST    && enable muxers
> > > -enabled_any $DEMUXER_LIST  && enable demuxers
> > > -enabled_any $PROTOCOL_LIST && enable protocols
> > > -enabled_any $BSF_LIST      && enable bsfs
> > > -
> > >  enabled_any $THREADS_LIST  && enable threads
> > >  
> > > -check_deps $CONFIG_LIST $HAVE_LIST $DECODER_LIST $ENCODER_LIST $PARSER_LIST \
> > > -    $BSF_LIST $DEMUXER_LIST $MUXER_LIST $PROTOCOL_LIST
> > > +for component in $COMPONENT_LIST; do
> > > +    LIST_NAME=`toupper ${component%s}`_LIST
> > > +    eval enabled_any \$$LIST_NAME && enable $component
> > > +    eval check_deps \$$LIST_NAME
> > > +done
> > >  
> > > +check_deps $CONFIG_LIST $HAVE_LIST
> > > +
> > >  enabled libogg    && append pkg_requires "ogg >= 1.1"
> > >  enabled libtheora && append pkg_requires "theora"
> > >  enabled libvorbis && append pkg_requires "vorbis vorbisenc"
> > > @@ -1975,14 +1971,11 @@
> > >  
> > >  print_config ARCH_   $TMPH config.mak $ARCH_LIST
> > >  print_config HAVE_   $TMPH config.mak $HAVE_LIST
> > > -print_config CONFIG_ $TMPH config.mak $CONFIG_LIST   \
> > > -                                      $DECODER_LIST  \
> > > -                                      $ENCODER_LIST  \
> > > -                                      $PARSER_LIST   \
> > > -                                      $BSF_LIST      \
> > > -                                      $DEMUXER_LIST  \
> > > -                                      $MUXER_LIST    \
> > > -                                      $PROTOCOL_LIST \
> > > +print_config CONFIG_ $TMPH config.mak $CONFIG_LIST
> > > +for component in $COMPONENT_LIST; do
> > > +    LIST_NAME=`toupper ${component%s}`_LIST
> > > +    eval print_config CONFIG_ $TMPH config.mak \$$LIST_NAME
> > > +done
> > >  
> > >  if test "$targetos" = darwin; then
> > >    echo "#define CONFIG_DARWIN 1"  >> $TMPH
> > 
> > This adds a level of indirection only to marginally reduce the amount
> > of code.  The new versions is probably slower too.  I don't like it.
> 
> I admit I havn't benchmarked it... but I doubt there is any significant
> speed difference (at least with any decent shell).
> The advantage here IMO is not the amount of code, but the reduced
> duplication. Now, when we will have to add a $AVFILTER_LIST, we will
> have to add it at dozen of places in configure. This patch simplify
> addition of new kind of "component". Don't you think it's worth ?
> 
> BTW, where your comment related only to the end of this patch or also
> to the first hunk ?
> 
> Anyway, attached a new patch containing further simplifications.
> Please tell me if at least the first hunk is acceptable.

No comments ?
Well I will apply only first hunk very soon.

Aurel




More information about the ffmpeg-devel mailing list