rplay PORTING - How to port rplay to other systems.
Last updated: Mon Nov 7 13:13:50 PST 1994
0. PORTING OVERVIEW
--------------------
This file describes the following rplay porting topics:
AUTO CONFIGURATION(1), AUDIO CONFIGURATION(2)
Please email new audio configurations to boyns@sdsu.edu.
1. AUTO CONFIGURATION
---------------------
Starting with 3.2.0, all system audio specific code is located in
separate files in the rplayd/audio directory. GNU autoconf is used to
determine what type of system you have and therefore what audio code
to use. Currently the following systems are automatically determined:
sun, solaris, hpux, sgi, linux, 386bsd, Free_BSD
New systems should be added to configure.in using the following linux example:
dnl
dnl Check for Linux
dnl
if test -z "$RPLAY_TARGET"; then
AC_MSG_CHECKING(for linux)
AC_EGREP_CPP(yes,
[
#ifdef linux
yes
#endif
], is_linux=yes, is_linux=no)
AC_MSG_RESULT($is_linux)
if test $is_linux = "yes"; then
RPLAY_TARGET="linux"
fi
fi
You will see the above in configure.in. The important line is "#ifdef
linux" which references a unique preprocessor symbol for the system.
You will have to find a unique symbol that your preprocessor defines
so your system type can be automatically determined. If you have gcc
you can probably find your system's preprocessor defines in the specs
file.
Assume you have a foobar system and it defines __foobar__. You would
then have to add the following to configure.in:
dnl
dnl Check for foobar
dnl
if test -z "$RPLAY_TARGET"; then
AC_MSG_CHECKING(for foobar)
AC_EGREP_CPP(yes,
[
#ifdef __foobar__
yes
#endif
], is_foobar=yes, is_foobar=no)
AC_MSG_RESULT($is_foobar)
if test $is_foobar = "yes"; then
RPLAY_TARGET="foobar"
fi
fi
This assumes that the foobar audio files will be called audio_foobar.c
and audio_foobar.h in the rplayd/audio directory. Now that you have a
new configure.in, you will have to create a new configure script. For
this, you will need to get the GNU autoconf package from your favorite
GNU archive site (prep.ai.mit.edu:/pub/gnu). After installing
autoconf, execute `autoconf' and a new `configure' script will be
created.
If you do not want to deal with installing autoconf, email me the
unique preprocessor symbol for your system and I'll create a configure
script for you. You can also experiment with the generic audio
configuration files.
2. AUDIO CONFIGURATION
-----------------------
The audio configuration files are located in the rplayd/audio
directory. Each system should have its own audio configuration. The
system type is determined using GNU autoconf (see section 1).
audio_generic.c and audio_generic.h should be used as a template.
audio_<system>.h must define the following:
RPLAY_AUDIO_DEVICE <string> - the name of the audio device.
RPLAY_AUDIO_TIMEOUT <int> - close the audio device when it hasn't been
used for <int> seconds.
RPLAY_AUDIO_RATE <int> - number of times per second that the audio
data should be written to the audio device.
RPLAY_AUDIO_BYTE_ORDER <val> - <val> should be RPLAY_BIG_ENDIAN or
RPLAY_LITTLE_ENDIAN.
RPLAY_AUDIO_FLUSH_TIMEOUT <int> - flush audio device when it hasn't been
used for <int> seconds. There are the
following exceptions:
<int> == 0 : flushing is disabled
<int> == -1 : flush when spool is empty
(this should be the default)
<int> == -2 : flush after each audio write
audio_<system>.c has the following routines:
int rplay_audio_init()
int rplay_audio_open()
int rplay_audio_isopen()
int rplay_audio_flush()
int rplay_audio_write(char *buf, int nbytes)
int rplay_audio_close()
int rplay_audio_get_volume()
int rplay_audio_set_volume(int volume)
One of the most important routines is rplay_audio_init(). This
routine sets the sample rate, precision, channels, audio port(s), and
the format of the audio device. Generic defaults are 8000, 8, 1,
speaker, and RPLAY_FORMAT_ULAW (see rplay.h) respectively.
The audio_sun rplay_audio_init detects an external speakerbox which
supports 16-bit audio and high sample rates so it could be a good
example for other configurations.
Systems that don't have hardware volume control should enable software
volume control by defining `FAKE_VOLUME'. Look at rplay_get_volume
and rplay_set_volume in audio_linux.c to see how FAKE_VOLUME should be
used.
$Id: PORTING,v 1.2 1998/08/13 06:13:01 boyns Exp $