[Mplayer-users] MPlayer on non x86 platform

Felix von Leitner leitner at convergence.de
Thu Jun 14 15:20:17 CEST 2001

Thus spake Maksim Krasnyanskiy (maxk at qualcomm.com):
> >Just to make sure you actually know what you are talking about: please
> >point me to a project of yours with autoconf support so I can see if the
> >autoconf actually suck as little as possible.
> http://vtun.sf.net (IP/Ethernet tunneler for Linux, Solaris, *BSD)
> http://vtun.sf.net/tun (Virtual net device driver, Linux, Solaris, FreeBSD) 
> http://bluez.sf.net (Linux Bluetooth stack)

> If you have time for analyzing configure scripts in my projects go ahead.

> >And let's see if your tests are actually portable.
> No complains so far :)

OK.  Please go home now.
Whether people complained is absolutely _no_ measure for code quality.

BTW: I think we actually talked before about vtun.  I complained that
the "encryption" was not secure and your random number generation was
predictable.  Your reply was "I don't care".  That's when I stopped
looking at vtun.

Your vtun configure.in looks quite clean now, but you don't actually do
any tests that need to compile anything like the MMX tests would need.

> >You know, the more one understands of autoconf, the more complicated it gets.  
> >Trust me, I know it well enough to lecture on it, and I refuse to use it in my projects 
> >because of the bloat it adds.
> Yeah, I guess hundreds of people are totally blind and dumb otherwise
> there is no explanation why so many projects use autoconf.

Millions of flies can't be wrong: eat shit!
Why are you using Linux, by the way?  Many more people use Windows and
you seem to believe that that is a measure for quality, right?

> >A hello-world style autoconf that looks for three headers and two libraries generates >100k
> That's soo not true.
> Not sure how you write your configure.in. I have several configure.in's that look for bunch of headers and libraries, etc
> use local defined macros, and --with options and they are about 64K - 82K.

It's the configure.in from ffingerd, a fingerd I wrote years ago.
It does some 30 checks, looking for programs, headers and libraries.
Nothing special.  110k configure.

> >And generating configure takes >4 seconds on my dual P-II w/ 400 MHz each (i.e. is completely unacceptably slow).
> Takes less than a sec for me.

It's a trivial substitution, for God's sake!
It should take no more than 0.01 seconds!
And unless you use a 400 MHz P-II, it's completely irrelevant for the
comparison how long it takes for you.


Mplayer-users mailing list
Mplayer-users at lists.sourceforge.net

More information about the MPlayer-users mailing list