[Mplayer-cvslog] CVS: main configure,1.910,1.911

[Alexander Strasser]

Joey Parrish wrote ( On Wed, Sep 15, 2004 at 05:58:13PM -0500 ):
> On Thu, Sep 16, 2004 at 12:04:55AM +0200, Alexander Strasser wrote:
> > Alexander Strasser wrote ( On Thu, Sep 16, 2004 at 12:01:07AM +0200 ):
> > > Didn't work for me.
> > > What do you think about this:
> > > if ( ( test `bio2jack-config --version | cut -d '.' -f 2` -ge 3 ) ) >/dev/null 2>&1 ; then

Oh! I just saw I forgot `!' after test... but shouldn't matter I'll apply
a correct version, when the points below are clear.

> > 
> > Please note that I'm no configure/shell expert, just a suggestion.
> 
> Works for me.  I say do it.
Would like to commit the fix, but I think this is flawed anyway:

echocheck "JACK"
if test "$_jack" = auto ; then
  _jack=no
  if ( ( pkg-config --modversion jack ) > /dev/null 2>&1 ) &&
     ( jackd --version | grep version | awk '{ print $3 }' ) >> "$TMPLOG" 2>&1 ; then

cat > $TMPC << EOF
#include <jack/jack.h>
int main(void) { JACK_Init(); return 0; }
EOF
        # This test only checks the minor version number.
        if ( ( test ! `bio2jack-config --version | cut -d '.' -f 2` -ge 3 ) ) >/dev/null 2>&1 ; then
            _jack=no;
        elif test -z "$_bio2jackdir" ; then
                cc_check -lbio2jack `pkg-config --libs --cflags jack` && ( "$TMPO" >> "$TMPLOG" 2>&1 ) && _jack=yes
        else
                cc_check -L "$_bio2jackdir" -lbio2jack `pkg-config --libs --cflags jack` && ( "$TMPO" >> "$TMPLOG" 2>&1 ) && _jack=yes
        fi
  fi
fi

- Why do we set _jack to no, even it was initialized to no at beginning
  for the tests to change it to yes eventually, or not?

- And why is the version test in a `if-elif-else'-conditional? That makes
  the version check unable to fail the whole thing alone? Is that really
  intended?

As I thought the version check is there to ensure that a specific version
must be used, but maybe I'm completely wrong... but confused for sure ;)

  Alex (beastd)