changelog

Windows NT声卡驱动VXD

	set O_NONBLOCK on fallback only for the sound device.

	* esdlib.c: removed fsync's here too

1998-11-06  Yo Ric Dude  <[e-mail]>

	* samples.c, players.c, proto.c: fixed rounding to multiple of 4
	algorithm, and removed fsync's as they don't do anything for a
	socket.  patches suggested by Miodrag Vallat.

1998-11-05  Yo Ric Dude  <[e-mail]>

	* audio_aix.c: typo fixes courtest of Philippe Defert

1998-11-04  Yo Ric Dude  <[e-mail]>

	* TODO, esd.h, esdctl.c, esdmgr.c, proto.c: It really should have
 	been esd_free_all_info(), and while we're at it, make it do
 	something potentially useful.
  	
	* esound.spec: set it up for autoconf to update the version
 	number, then lost the changes. thanks to Michael Sterrett.

1998-11-03  Christopher Blizzard  <[e-mail]>

	* esdctl.c (main): Change call from esd_free_all_info() to
 	esd_free_esd_info() so that this will build.

1998-11-03  Yo Ric Dude  <[e-mail]>

	* TODO, configure.in: update version number
	* Makefile.am, esd-server.h, esd.h, esdctl.c, proto.c, esdmgr.c: 
	add initial framework for remote management of server, filter out
	diagnostic/debugging information from binary if we don't want it.
	also inlcude esound.spec file in distribution.

Sat Oct 31 21:36:07 PST 1998 Manish Singh <[e-mail]>

	* esddsp.c: don't try to open local sound device if ESPEAKER is set

1998-10-31  Yo Ric Dude  <[e-mail]>

	* EsounD.html, EsounD.ps, Makefile.am: replaced EsounD.html with
 	EsounD.ps until the doc subdirectory is functional with docbook.
	* configure.in: call it 0.2.6.

Fri Oct 30 20:39:16 PST 1998 Manish Singh <[e-mail]>

	* esddsp.c: provide implementation for close() to expire sndfd

1998-10-28  Yo Ric Dude  <[e-mail]>

	* esound.spec.in: reordered the changelog entries, as rpm likes to 
	complain about it if the order is backwards...

1998-10-24  Yo Ric Dude  <[e-mail]>

	* audio_alsa.c: reworked alsa card detection per patch from 
	Andrew Clausen <[e-mail]>

Wed Oct 21 23:57:11 1998  Raph Levien  <[e-mail]>

	* esdplay.c (play_file): The computation of bytes_per_frame was
	not taking the number of channels into account. It does now.

1998-10-21  Yo Ric Dude  <[e-mail]>

	* esddsp.c, esddsp.in: give esddsp some class, allow for passing
	of identifying name to esd via environment variable

Tue Oct 20 20:43:48 PDT 1998 Manish Singh <[e-mail]>

	* clients.c: make optional use of tcp_wrappers

	* Makefile.am, configure.in: OSS runs on other platforms too, so
 	explictly special case for linux for esddsp. Also don't install
 	the wrapper script if esddsp isn't supported. Add a check for
 	libwrap and friends. Some reorganization.

	* esddsp.c: respond to SNDCTL_DSP_GETCAPS (and say we can't do much)
	More evilness to get multiple x11amps to coexist (optional)

	* esddsp.in: exec instead of just run

1998-10-19  Yo Ric Dude  <[e-mail]>

	* Makefile.am, esddsp.in, configure.in: automatically build the
	esddsp wrapper script for Linux only

Sun Oct 18 21:33:26 PDT 1998 Manish Singh <[e-mail]>

	* Makefile.am: add TODO and esddsp.c to EXTRA_DIST

	* esddsp.c: somewhat of a rewrite. We only try to use esound
	if an open() fails. This lets the no esound case fall back to
	/dev/dsp, and also doesn't do a strcmp on every call to open.
	Got rid of the varargs junk in ioctl, I've never seen an ioctl
	with more than one extra arg anyway (please tell me if I'm wrong).
	Reworked the code, so we're not so x11amp specific, i.e. the
	esound socket is opened immediately after all the info is set.
	We also make sure we don't try to set it again. Default to mono
	output per OSS specs. open/close/open with multiple formats
	works. Provide a default value for GETFMTS. The result of all
	this is that libesddsp works with a lot more stuff now. I've
	tested x11amp, mpg123, amp, freeamp, xanim, rvplayer, and
	esound itself. ;) (a side effect is that rvplayer works on
	Linux 2.1.120+ kernels as well)

1998-10-18  Yo Ric Dude  <[e-mail]>

	* esddsp.c, proto.c: fixed inconsistency in esd_on_standby
 	handling withing esd_proto_resume and esd_proto_standby. 

	* esd.c: removed unused variable, audio

	* esd-server.h, esd.c, esddsp.c, players.c: how did monitoring
	ever work like that? corrected the header, actually passed in the
	length of the monitor buffer.

1998-10-17  Raja R Harinath  <[e-mail]>

	* configure.in (nanosleep): Make this work on Solaris.
	From Brandon S. Allbery.

1998-10-17  Yo Ric Dude  <[e-mail]>

	* esddsp.c: added Yosh's slick trick for transparently
	rerouting /dev/dsp to esd.  Only tested with Linux and
	x11amp.  Directions are in the source file for the brave.

Sat Oct 17 02:20:42 PDT 1998 Manish Singh <[e-mail]>

	* esd.c, acconfig.h, configure.in: emulated nanosleep with select
	for systems that don't have it. Compiles on DU4, FreeBSD

	* audio_oss.c: FreeBSD uses __FreeBSD__, not FREEBSD

1998-10-16  Yo Ric Dude  <[e-mail]>

	* filter.c, mix.c, players.c: misc optimizations courtesy of Bibek.

	* audio_hpux.c: HP-UX patches courtest of Jimmy Olsen <[e-mail]>.
	It is claimed that esd now runs on HP-Ux

	* esd-server.h, esd.c, filter.c, players.c, proto.c: added support 
	for multiple simultaneous monitor streams.

Fri Oct 16 02:48:26 PDT 1998 Manish Singh <[e-mail]>

	* esd.h: removed system includes so it doesn't "pollute"

	* audio.c, esd-server.h, esd.h, essrldcat.c esdctl.c, esdfilt.c,
	esdloop.c, esdmon.c, esdplay.c, esdrec.c, esdsample.c: added
	in specifc #includes so it compiles again

1998-10-14  Yo Ric Dude  <[e-mail]>

	* Makefile.am: esdctl now uses getopt_long, and therefore needs
	the LIBOBJS in it's LDADD.

	* esdctl.c: beefed up esdctl for industrial grade use.  added an
 	exit_usage function, and getopt style option parsing.  It was
 	destined to become esdsh, but this in combination with a real
 	scripting language should be suffient.

1998-10-13  Yo Ric Dude  <[e-mail]>

	* esdlib.c, proto.c: trap for NULL name sent to sample_getid(),
 	and check for sample names beyond the head of the list.

Mon Oct 12 23:12:03 CDT 1998 Frank Belew <[e-mail]>

	* esd.h: removed anything resembling a socket header
	* esd-server.h: added everything removed from esd.h
	* audio_alsa.c: #include esd.h instead of esd-server.h since sockets 
	are not needed.
	* esdlib.c: removed superfluous headers that are already included from
	esd-server.h

1998-10-10  Yo Ric Dude  <[e-mail]>

	* esd-server.h, esd.c, esd.h, esdlib.c, mix.c, proto.c, filter.c,
 	esdfilt.c Makefile.am, .cvsignore: filter patch courtesy of Bibek
 	Sahu. ten points to the first person to create esdequalizer. does
 	libgsl do digital filters?

Tue Oct  6 20:57:37 PDT 1998 Manish Singh <[e-mail]>

	* esd.c: nanosleep in the no data ready case too, so we don't
	suck up CPU

Sat Oct  3 13:31:29 PDT 1998 Manish Singh <[e-mail]>

	* audio.c: added driver_trace variable for sound driver debug
	messages

	* audio_alsa.c: use driver_trace instead of esdbg_trace, since
	that's limited to the server. Fix Pav's completely broken fix.

Thu Oct  1 19:27:22 PDT 1998 Manish Singh <[e-mail]>

	* esd.m4: make it work right

1998-10-01  Yo Ric Dude  <[e-mail]>

	* esound.spec[.in]: make autoconf do the version updating for us.

	* Makefile.am, configure.in: move version updating to same place.

1998-10-01  Yo Ric Dude  <[e-mail]>

	* audio_alsa.c: multiple audio card detection patches from Bibek.
  	This should open and use the first available card under ALSA.  It
 	has been tested on single card systems, and works fine.  However,
 	no one with multiple audio cards has tested this, so if anyone is
 	feeling brave and adventurous... =)

1998-10-01  Yo Ric Dude  <[e-mail]>

	* esd.h, esdlib.c: why sample size is treated as a long in esdlib,
 	but an int everywhere else is beyond me.  64 bit platforms should
 	be happier now. still have to check 32-64 bit interoperability

	* configure.in: touch-ups and commenting *shudder*.

Thu Oct  1 22:58:03 CEST 1998 Jochen Friedrich <[e-mail]>

	* esdlib.c: added kludge to esd_sample_cache() to make it
	work under Alpha. Problem was to send a long and receive an int.
	On Alpha this makes extra 4 bytes of crap enough to bring
	client and server out of sync. Probably a better solution is
	to change the prototype for size from long to int.

1998-10-01  Raja R Harinath  <[e-mail]>

	* configure.in (esd-config): Turn on execute bit only if
	`esd-config' was generated.
	(getopt_long): New check.
	* getopt.h, getopt.c, getopt1.c: New files.
	* Makefile.am (esdplay_LDADD): Add LIBOBJS.
	
Wed Sep 30 21:08:19 PDT 1998 Manish Singh <[e-mail]>

	* esd.m4
	* esd-config
	* configure.in: Added support for esd-config (like GTK+). Added
	explict disabling of local sound (for an audio data server). Fixed
	NEW_ALSA detection. Define HAVE_INET_ATON if we have it.

	* audio.c
	* audio_none.c: support for no local sound

	* esdlib.c: provide inet_aton implementation if it's not avaliable

1998-09-29  Yo Ric Dude  <[e-mail]>

	* esd-server.h, esd.c: updated obsolete info. As noted by Sopwith,
 	termination procedure, clean_exit(), needed tweaking.

	* esd.h, esdlib.c, esdsample.c, proto.c: esd_sample_getid(),
	for fetching a sample id from the name, courtesy of Sopwith.

	* esdcat.c: FILE *source = source; /* hello, mcfly? anybody 
	in there? */ pointed out by Bibek, Sopwith, and others...

1998-09-27  Yo Ric Dude  <[e-mail]>

	* configure.in: `AC_PROG_RANLIB' is rendered obsolete by `AM_PROG_LIBTOOL'

	* esdlib.c: added the missing esd_close() function.  no need for 
	clients to know they're dealing with a socket.

1998-09-25  Mandrake <[e-mail]>

	* autogen.sh: added libtoolize --copy --force to autogen.sh per
	Andreas Jellinghaus <[e-mail]>

1998-09-26  Yo Ric Dude  <[e-mail]>

	* clients.c: artistic changes 

	* configure.in: add check for nanosleep, Solaris needs -lposix4.
 	Also included potentially useful information for obtaining the
 	audiofile library.

	* esd.c: forced a wait in esd.c while in standby mode, so there's
	still some CPU to go around.

	* esound.spec: version number update, maybe make a esound.spec.in?

Wed Sep 23 13:05:11 CDT 1998  mej <[e-mail]>

	* clients.c, configure.in - Fixes needed to find FIONBIO macro,
	  at least on Solaris
	
1998-09-23  Kimball Thurston <[e-mail]>

	* clients.c - Make sure that the socket when connected is marked
	  as non-blocking

	* audio_irix.c - Fix to reset the output rate every time a sample
	  is played, so other utilities that play sounds and rudely
	  change the output rate on SGI systems can't mess with esd too
	  much

1998-09-22  Yo Ric Dude  <[e-mail]>

	* AUTHORS, TODO: updates/formatting

	* acconfig.h, audio.c, audio_alsa.c, configure.in: "New" ALSA
 	driver detection courtesy of Bibek Sahu.

1998-09-20  Yo Ric Dude  <[e-mail]>

	* configure.in: made esdplay a --enable-able option, that was fun.
	let's bump up the revision level here, too, while we're at it.

	* esd.c, mix.c, samples.c: i think i've got the 8bit stuff working,
	but it needs to be cleaned up, and preferably use function pointers 
	over if-else blocks to pick the appropriate 8bit/16bit algorithms.

1998-09-19  Yo Ric Dude  <[e-mail]>