[PATCH] add aplaymidi & arecordmidi utils

[PATCH] add aplaymidi & arecordmidi utils

[Clemens Ladisch]

See man pages for details.


add aplaymidi & arecordmidi utilities

Index: alsa-utils/configure.in
===================================================================
RCS file: /cvsroot/alsa/alsa-utils/configure.in,v
retrieving revision 1.76
diff -u -r1.76 configure.in
--- alsa-utils/configure.in	26 Jan 2004 19:30:45 -0000	1.76
+++ alsa-utils/configure.in	23 Feb 2004 08:27:54 -0000
@@ -69,4 +69,5 @@
 AC_OUTPUT(Makefile alsactl/Makefile alsamixer/Makefile amidi/Makefile amixer/Makefile \
 	  alsaconf/alsaconf alsaconf/Makefile \
 	  aplay/Makefile include/Makefile iecset/Makefile utils/Makefile \
-	  utils/alsa-utils.spec seq/Makefile seq/aconnect/Makefile seq/aseqnet/Makefile)
+	  utils/alsa-utils.spec seq/Makefile seq/aconnect/Makefile \
+	  seq/aplaymidi/Makefile seq/aseqnet/Makefile)
Index: alsa-utils/amidi/amidi.1
===================================================================
RCS file: /cvsroot/alsa/alsa-utils/amidi/amidi.1,v
retrieving revision 1.2
diff -u -r1.2 amidi.1
--- alsa-utils/amidi/amidi.1	19 Jan 2004 18:39:12 -0000	1.2
+++ alsa-utils/amidi/amidi.1	23 Feb 2004 08:27:54 -0000
@@ -1,4 +1,4 @@
-.TH AMIDI 1 "18 Jan 2004"
+.TH AMIDI 1 "22 Feb 2004"

 .SH NAME
 amidi \- read from and write to ALSA RawMIDI ports
@@ -16,7 +16,12 @@
 .B amidi
 handles only files containing raw MIDI commands, without timing
 information.
-Standard MIDI (.mid) files are not supported.
+.B amidi
+does not support Standard MIDI (.mid) files, but
+.B aplaymidi(1)
+and
+.B arecordmidi(1)
+do.

 .SH INVOKING
 .B amidi
@@ -68,17 +73,15 @@
 .I -s, --send=filename
 Sends the contents of the specified file to the MIDI port.
 The file must contain raw MIDI commands (e.g. a .syx file);
-you can
-.I not
-use a Standard MIDI (.mid) file for this.
+for Standard MIDI (.mid) files, use
+.B aplaymidi(1).

 .TP
 .I -r, --receive=filename
 Writes data received from the MIDI port into the specified file.
 The file will contain raw MIDI commands (such as in a .syx file);
-this will
-.I not
-create a Standard MIDI (.mid) file.
+to record a Standard MIDI (.mid) file, use
+.B arecordmidi(1).

 .B amidi
 will filter out any Active Sensing bytes (FEh), unless the
@@ -155,6 +158,11 @@
 The
 .I --list-devices
 option pretends that output and input ports are the same.
+
+.SH SEE ALSO
+aplaymidi(1)
+.br
+arecordmidi(1)

 .SH AUTHOR
 Clemens Ladisch <clemens@ladisch.de>
Index: alsa-utils/seq/Makefile.am
===================================================================
RCS file: /cvsroot/alsa/alsa-utils/seq/Makefile.am,v
retrieving revision 1.1
diff -u -r1.1 Makefile.am
--- alsa-utils/seq/Makefile.am	4 Jan 2000 12:39:00 -0000	1.1
+++ alsa-utils/seq/Makefile.am	23 Feb 2004 08:27:54 -0000
@@ -1 +1 @@
-SUBDIRS=aconnect aseqnet
+SUBDIRS=aconnect aplaymidi aseqnet
diff -urN ../../oalsa/alsa/alsa-utils/seq/aplaymidi/Makefile.am alsa-utils/seq/aplaymidi/Makefile.am
--- ../../oalsa/alsa/alsa-utils/seq/aplaymidi/Makefile.am	1970-01-01 01:00:00.000000000 +0100
+++ alsa-utils/seq/aplaymidi/Makefile.am	2004-02-23 09:38:11.051239000 +0100
@@ -0,0 +1,5 @@
+INCLUDES = -I$(top_srcdir)/include
+EXTRA_DIST = aplaymidi.1 arecordmidi.1
+
+bin_PROGRAMS = aplaymidi arecordmidi
+man_MANS = aplaymidi.1 arecordmidi.1
diff -urN ../../oalsa/alsa/alsa-utils/seq/aplaymidi/aplaymidi.1 alsa-utils/seq/aplaymidi/aplaymidi.1
--- ../../oalsa/alsa/alsa-utils/seq/aplaymidi/aplaymidi.1	1970-01-01 01:00:00.000000000 +0100
+++ alsa-utils/seq/aplaymidi/aplaymidi.1	2004-02-23 09:38:11.061243000 +0100
@@ -0,0 +1,55 @@
+.TH APLAYMIDI 1 "15 Feb 2004"
+
+.SH NAME
+aplaymidi \- play Standard MIDI Files
+
+.SH SYNOPSIS
+.B aplaymidi
+-p client:port[,...] [-d delay] midifile ...
+
+.SH DESCRIPTION
+.B aplaymidi
+is a command-line utility that plays the specified MIDI file(s) to one
+or more ALSA sequencer ports.
+
+.SH OPTIONS
+
+.TP
+.I -h, --help
+Prints a list of options.
+
+.TP
+.I -V, --version
+Prints the current version.
+
+.TP
+.I -l, --list
+Prints a list of possible output ports.
+
+.TP
+.I -p, --port=client:port,...
+Sets the sequencer port(s) to which the events in the MIDI file(s) are
+sent.
+
+A client can be specified by its number, its name, or a prefix of its
+name. A port is specified by its number; for port 0 of a client, the
+":0" part of the port specification can be omitted.
+
+For compatibility with
+.B pmidi(1),
+the port specification is taken from the
+.I ALSA_OUTPUT_PORTS
+environment variable if none is given on the command line.
+
+.TP
+.I -d, --delay=seconds
+Specifies how long to wait after the end of each MIDI file,
+to allow the last notes to die away.
+
+.SH SEE ALSO
+pmidi(1)
+.br
+playmidi(1)
+
+.SH AUTHOR
+Clemens Ladisch <clemens@ladisch.de>
diff -urN ../../oalsa/alsa/alsa-utils/seq/aplaymidi/aplaymidi.c alsa-utils/seq/aplaymidi/aplaymidi.c
--- ../../oalsa/alsa/alsa-utils/seq/aplaymidi/aplaymidi.c	1970-01-01 01:00:00.000000000 +0100
+++ alsa-utils/seq/aplaymidi/aplaymidi.c	2004-02-23 09:38:11.091239000 +0100
@@ -0,0 +1,890 @@
+/*
+ * aplaymidi.c - play Standard MIDI Files to sequencer port(s)
+ *
+ * Copyright (c) 2004 Clemens Ladisch <clemens@ladisch.de>
+ *
+ *
+ *  This program is free software; you can redistribute it and/or modify
+ *  it under the terms of the GNU General Public License as published by
+ *  the Free Software Foundation; either version 2 of the License, or
+ *  (at your option) any later version.
+ *
+ *  This program is distributed in the hope that it will be useful,
+ *  but WITHOUT ANY WARRANTY; without even the implied warranty of
+ *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ *  GNU General Public License for more details.
+ *
+ *  You should have received a copy of the GNU General Public License
+ *  along with this program; if not, write to the Free Software
+ *  Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307 USA
+ */
+
+/* TODO: sequencer queue timer selection */
+
+#include <stdio.h>
+#include <stdlib.h>
+#include <stdarg.h>
+#include <string.h>
+#include <getopt.h>
+#include <unistd.h>
+#include <alsa/asoundlib.h>
+#include "aconfig.h"
+#include "version.h"
+
+/*
+ * A MIDI event after being parsed/loaded from the file.
+ * There could be made a case for using snd_seq_event_t instead.
+ */
+struct event {
+	struct event *next;		/* linked list */
+
+	unsigned char type;		/* SND_SEQ_EVENT_xxx */
+	unsigned char port;		/* port index */
+	unsigned int tick;
+	union {
+		unsigned char d[3];	/* channel and data bytes */
+		int tempo;
+		unsigned int length;	/* length of sysex data */
+	} data;
+	unsigned char sysex[0];
+};
+
+struct track {
+	struct event *first_event;	/* list of all events in this track */
+	int end_tick;			/* length of this track */
+
+	struct event *current_event;	/* used while loading and playing */
+};
+
+static snd_seq_t *seq;
+static int client;
+static int port_count;
+static snd_seq_addr_t *ports;
+static int queue;
+static int end_delay = 2;
+static const char *file_name;
+static FILE *file;
+static int file_offset;		/* current offset in input file */
+static int num_tracks;
+static struct track *tracks;
+static int smpte_timing;
+
+/* prints an error message to stderr */
+static void errormsg(const char *msg, ...)
+{
+	va_list ap;
+
+	va_start(ap, msg);
+	vfprintf(stderr, msg, ap);
+	va_end(ap);
+	fputc('\n', stderr);
+}
+
+/* prints an error message to stderr, and dies */
+static void fatal(const char *msg, ...)
+{
+	va_list ap;
+
+	va_start(ap, msg);
+	vfprintf(stderr, msg, ap);
+	va_end(ap);
+	fputc('\n', stderr);
+	exit(EXIT_FAILURE);
+}
+
+/* memory allocation error handling */
+static void check_mem(void *p)
+{
+	if (!p)
+		fatal("Out of memory");
+}
+
+/* error handling for ALSA functions */
+static void check_snd(const char *operation, int err)
+{
+	if (err < 0)
+		fatal("Cannot %s - %s", operation, snd_strerror(err));
+}
+
+static void init_seq(void)
+{
+	int err;
+	snd_seq_client_info_t *info;
+
+	/* open sequencer */
+	err = snd_seq_open(&seq, "default", SND_SEQ_OPEN_DUPLEX, 0);
+	check_snd("open sequencer", err);
+
+	/* set our name (otherwise it's "Client-xxx") */
+	err = snd_seq_set_client_name(seq, "aplaymidi");
+	check_snd("set client name", err);
+
+	/* find out who we actually are */
+	client = snd_seq_client_id(seq);
+	check_snd("get client id", client);
+}
+
+/* parses one or more port addresses from the string */
+static void parse_ports(const char *arg)
+{
+	char *buf, *s, *port_name;
+	int err;
+
+	/* make a copy of the string because we're going to modify it */
+	buf = strdup(arg);
+	check_mem(buf);
+
+	for (port_name = s = buf; s; port_name = s + 1) {
+		/* Assume that ports are separated by commas.  We don't use
+		 * spaces because those are valid in client names. */
+		s = strchr(port_name, ',');
+		if (s)
+			*s = '\0';
+
+		++port_count;
+		ports = realloc(ports, port_count * sizeof(snd_seq_addr_t));
+		check_mem(ports);
+
+		err = snd_seq_parse_address(seq, &ports[port_count - 1], port_name);
+		if (err < 0)
+			fatal("Invalid port %s - %s", port_name, snd_strerror(err));
+	}
+
+	free(buf);
+}
+
+static void create_source_port(void)
+{
+	snd_seq_port_info_t *pinfo;
+	int err;
+
+	snd_seq_port_info_alloca(&pinfo);
+
+	/* the first created port is 0 anyway, but let's make sure ... */
+	snd_seq_port_info_set_port(pinfo, 0);
+	snd_seq_port_info_set_port_specified(pinfo, 1);
+
+	snd_seq_port_info_set_name(pinfo, "aplaymidi");
+
+	snd_seq_port_info_set_capability(pinfo, 0); /* sic */
+	snd_seq_port_info_set_type(pinfo,
+				   SND_SEQ_PORT_TYPE_MIDI_GENERIC |
+				   SND_SEQ_PORT_TYPE_APPLICATION);
+
+	err = snd_seq_create_port(seq, pinfo);
+	check_snd("create port", err);
+}
+
+static void create_queue(void)
+{
+	queue = snd_seq_alloc_named_queue(seq, "aplaymidi");
+	check_snd("create queue", queue);
+	/* the queue is now locked, which is just fine */
+}
+
+static void connect_ports(void)
+{
+	int i, err;
+
+	/*
+	 * We send MIDI events with explicit destination addresses, so we don't
+	 * need any connections to the playback ports.  But we connect to those
+	 * anyway to force any underlying RawMIDI ports to remain open while
+	 * we're playing - otherwise, ALSA would reset the port after every
+	 * event.
+	 */
+	for (i = 0; i < port_count; ++i) {
+		err = snd_seq_connect_to(seq, 0, ports[i].client, ports[i].port);
+		if (err < 0)
+			fatal("Cannot connect to port %d:%d - %s",
+			      ports[i].client, ports[i].port, snd_strerror(err));
+	}
+}
+
+static int read_byte(void)
+{
+	++file_offset;
+	return getc(file);
+}
+
+/* reads a little-endian 32-bit integer */
+static int read_32_le(void)
+{
+	int value;
+	value = read_byte();
+	value |= read_byte() << 8;
+	value |= read_byte() << 16;
+	value |= read_byte() << 24;
+	return !feof(file) ? value : -1;
+}
+
+/* reads a 4-character identifier */
+static int read_id(void)
+{
+	return read_32_le();
+}
+#define MAKE_ID(c1, c2, c3, c4) ((c1) | ((c2) << 8) | ((c3) << 16) | ((c4) << 24))
+
+/* reads a fixed-size big-endian number */
+static int read_int(int bytes)
+{
+	int c, value = 0;
+
+	do {
+		c = read_byte();
+		if (c == EOF)
+			return -1;
+		value = (value << 8) | c;
+	} while (--bytes);
+	return value;
+}
+
+/* reads a variable-length number */
+static int read_var(void)
+{
+	int value, c;
+
+	c = read_byte();
+	value = c & 0x7f;
+	if (c & 0x80) {
+		c = read_byte();
+		value = (value << 7) | (c & 0x7f);
+		if (c & 0x80) {
+			c = read_byte();
+			value = (value << 7) | (c & 0x7f);
+			if (c & 0x80) {
+				c = read_byte();
+				value = (value << 7) | c;
+				if (c & 0x80)
+					return -1;
+			}
+		}
+	}
+	return !feof(file) ? value : -1;
+}
+
+/* allocates a new event */
+static struct event *new_event(struct track *track, int sysex_length)
+{
+	struct event *event;
+
+	event = malloc(sizeof(struct event) + sysex_length);
+	check_mem(event);
+
+	event->next = NULL;
+
+	/* append at the end of the track's linked list */
+	if (track->current_event)
+		track->current_event->next = event;
+	else
+		track->first_event = event;
+	track->current_event = event;
+
+	return event;
+}
+
+static void skip(int bytes)
+{
+	while (bytes > 0)
+		read_byte(), --bytes;
+}
+
+/* reads one complete track from the file */
+static int read_track(struct track *track, int track_end)
+{
+	int tick = 0;
+	unsigned char last_cmd = 0;
+	unsigned char port = 0;
+
+	/* the current file position is after the track ID and length */
+	while (file_offset < track_end) {
+		unsigned char cmd;
+		struct event *event;
+		int delta_ticks, len, c;
+
+		delta_ticks = read_var();
+		if (delta_ticks < 0)
+			break;
+		tick += delta_ticks;
+
+		c = read_byte();
+		if (c < 0)
+			break;
+
+		if (c & 0x80) {
+			/* have command */
+			cmd = c;
+			if (cmd < 0xf0)
+				last_cmd = cmd;
+		} else {
+			/* running status */
+			ungetc(c, file);
+			file_offset--;
+			cmd = last_cmd;
+			if (!cmd)
+				goto _error;
+		}
+
+		switch (cmd >> 4) {
+			/* maps SMF events to ALSA sequencer events */
+			static unsigned char cmd_type[] = {
+				[0x8] = SND_SEQ_EVENT_NOTEOFF,
+				[0x9] = SND_SEQ_EVENT_NOTEON,
+				[0xa] = SND_SEQ_EVENT_KEYPRESS,
+				[0xb] = SND_SEQ_EVENT_CONTROLLER,
+				[0xc] = SND_SEQ_EVENT_PGMCHANGE,
+				[0xd] = SND_SEQ_EVENT_CHANPRESS,
+				[0xe] = SND_SEQ_EVENT_PITCHBEND
+			};
+
+		case 0x8: /* channel msg with 2 parameter bytes */
+		case 0x9:
+		case 0xa:
+		case 0xb:
+		case 0xe:
+			event = new_event(track, 0);
+			event->type = cmd_type[cmd >> 4];
+			event->port = port;
+			event->tick = tick;
+			event->data.d[0] = cmd & 0x0f;
+			event->data.d[1] = read_byte() & 0x7f;
+			event->data.d[2] = read_byte() & 0x7f;
+			break;
+
+		case 0xc: /* channel msg with 1 parameter byte */
+		case 0xd:
+			event = new_event(track, 0);
+			event->type = cmd_type[cmd >> 4];
+			event->port = port;
+			event->tick = tick;
+			event->data.d[0] = cmd & 0x0f;
+			event->data.d[1] = read_byte() & 0x7f;
+			break;
+
+		case 0xf:
+			switch (cmd) {
+			case 0xf0: /* sysex */
+			case 0xf7: /* continued sysex, or escaped commands */
+				len = read_var();
+				if (len < 0)
+					goto _error;
+				if (cmd == 0xf0)
+					++len;
+				event = new_event(track, len);
+				event->type = SND_SEQ_EVENT_SYSEX;
+				event->port = port;
+				event->tick = tick;
+				event->data.length = len;
+				if (cmd == 0xf0) {
+					event->sysex[0] = 0xf0;
+					c = 1;
+				} else {
+					c = 0;
+				}
+				for (; c < len; ++c)
+					event->sysex[c] = read_byte();
+				break;
+
+			case 0xff: /* meta event */
+				c = read_byte();
+				len = read_var();
+				if (len < 0)
+					goto _error;
+
+				switch (c) {
+				case 0x21: /* port number */
+					if (len < 1)
+						goto _error;
+					port = read_byte() % port_count;
+					skip(len - 1);
+					break;
+
+				case 0x2f: /* end of track */
+					track->end_tick = tick;
+					skip(track_end - file_offset);
+					return 1;
+
+				case 0x51: /* tempo */
+					if (len < 3)
+						goto _error;
+					if (smpte_timing) {
+						/* SMPTE timing doesn't change */
+						skip(len);
+					} else {
+						event = new_event(track, 0);
+						event->type = SND_SEQ_EVENT_TEMPO;
+						event->port = port;
+						event->tick = tick;
+						event->data.tempo = read_byte() << 16;
+						event->data.tempo |= read_byte() << 8;
+						event->data.tempo |= read_byte();
+						skip(len - 3);
+					}
+					break;
+
+				default: /* ignore all other meta events */
+					skip(len);
+					break;
+				}
+				break;
+
+			default: /* invalid Fx command */
+				goto _error;
+			}
+			break;
+
+		default: /* cannot happen */
+			goto _error;
+		}
+	}
+_error:
+	errormsg("%s: invalid MIDI data (offset %#x)", file_name, file_offset);
+	return 0;
+}
+
+/* reads an entire MIDI file */
+static int read_smf(void)
+{
+	int header_len, type, time_division, i, err;
+	snd_seq_queue_tempo_t *queue_tempo;
+
+	/* the curren position is immediately after the "MThd" id */
+	header_len = read_int(4);
+	if (header_len < 6) {
+invalid_format:
+		errormsg("%s: invalid file format", file_name);
+		return 0;
+	}
+
+	type = read_int(2);
+	if (type != 0 && type != 1) {
+		errormsg("%s: type %d format is not supported", file_name, type);
+		return 0;
+	}
+
+	num_tracks = read_int(2);
+	if (num_tracks < 1 || num_tracks > 1000) {
+		errormsg("%s: invalid number of tracks (%d)", file_name, num_tracks);
+		num_tracks = 0;
+		return 0;
+	}
+	tracks = calloc(num_tracks, sizeof(struct track));
+	if (!tracks) {
+		errormsg("out of memory");
+		num_tracks = 0;
+		return 0;
+	}
+
+	time_division = read_int(2);
+	if (time_division < 0)
+		goto invalid_format;
+
+	/* interpret and set tempo */
+	snd_seq_queue_tempo_alloca(&queue_tempo);
+	smpte_timing = !!(time_division & 0x8000);
+	if (!smpte_timing) {
+		/* time_division is ticks per quarter */
+		snd_seq_queue_tempo_set_tempo(queue_tempo, 500000); /* default: 120 bpm */
+		snd_seq_queue_tempo_set_ppq(queue_tempo, time_division);
+	} else {
+		/* upper byte is negative frames per second */
+		i = 0x80 - ((time_division >> 8) & 0x7f);
+		/* lower byte is ticks per frame */
+		time_division &= 0xff;
+		/* now pretend that we have quarter-note based timing */
+		switch (i) {
+		case 24:
+			snd_seq_queue_tempo_set_tempo(queue_tempo, 500000);
+			snd_seq_queue_tempo_set_ppq(queue_tempo, 12 * time_division);
+			break;
+		case 25:
+			snd_seq_queue_tempo_set_tempo(queue_tempo, 400000);
+			snd_seq_queue_tempo_set_ppq(queue_tempo, 10 * time_division);
+			break;
+		case 29: /* 30 drop-frame */
+			snd_seq_queue_tempo_set_tempo(queue_tempo, 100000000);
+			snd_seq_queue_tempo_set_ppq(queue_tempo, 2997 * time_division);
+			break;
+		case 30:
+			snd_seq_queue_tempo_set_tempo(queue_tempo, 500000);
+			snd_seq_queue_tempo_set_ppq(queue_tempo, 15 * time_division);
+			break;
+		default:
+			errormsg("%s: invalid number of SMPTE frames per second (%d)",
+				 file_name, i);
+			return 0;
+		}
+	}
+	err = snd_seq_set_queue_tempo(seq, queue, queue_tempo);
+	if (err < 0) {
+		errormsg("Cannot set queue tempo (%u/%i)",
+			 snd_seq_queue_tempo_get_tempo(queue_tempo),
+			 snd_seq_queue_tempo_get_ppq(queue_tempo));
+		return 0;
+	}
+
+	/* read tracks */
+	for (i = 0; i < num_tracks; ++i) {
+		int len;
+
+		/* search for MTrk chunk */
+		for (;;) {
+			int id = read_id();
+			len = read_int(4);
+			if (feof(file)) {
+				errormsg("%s: unexpected end of file", file_name);
+				return 0;
+			}
+			if (len < 0 || len >= 0x10000000) {
+				errormsg("%s: invalid chunk length %d", file_name, len);
+				return 0;
+			}
+			if (id == MAKE_ID('M', 'T', 'r', 'k'))
+				break;
+			skip(len);
+		}
+		if (!read_track(&tracks[i], file_offset + len))
+			return 0;
+	}
+	return 1;
+}
+
+static int read_riff(void)
+{
+	/* skip file length */
+	read_byte();
+	read_byte();
+	read_byte();
+	read_byte();
+
+	/* check file type ("RMID" = RIFF MIDI) */
+	if (read_id() != MAKE_ID('R', 'M', 'I', 'D')) {
+invalid_format:
+		errormsg("%s: invalid file format", file_name);
+		return 0;
+	}
+	/* search for "data" chunk */
+	for (;;) {
+		int id = read_id();
+		int len = read_32_le();
+		if (feof(file)) {
+data_not_found:
+			errormsg("%s: data chunk not found", file_name);
+			return 0;
+		}
+		if (id == MAKE_ID('d', 'a', 't', 'a'))
+			break;
+		if (len < 0)
+			goto data_not_found;
+		skip((len + 1) & ~1);
+	}
+	/* the "data" chunk must contain data in SMF format */
+	if (read_id() != MAKE_ID('M', 'T', 'h', 'd'))
+		goto invalid_format;
+	return read_smf();
+}
+
+static void cleanup_file_data(void)
+{
+	int i;
+	struct event *event;
+
+	for (i = 0; i < num_tracks; ++i) {
+		event = tracks[i].first_event;
+		while (event) {
+			struct event *next = event->next;
+			free(event);
+			event = next;
+		}
+	}
+	num_tracks = 0;
+	free(tracks);
+	tracks = NULL;
+}
+
+static void play_midi(void)
+{
+	snd_seq_event_t ev;
+	int i, max_tick, err;
+
+	/* calculate length of the entire file */
+	max_tick = -1;
+	for (i = 0; i < num_tracks; ++i) {
+		if (tracks[i].end_tick > max_tick)
+			max_tick = tracks[i].end_tick;
+	}
+
+	/* initialize current position in each track */
+	for (i = 0; i < num_tracks; ++i)
+		tracks[i].current_event = tracks[i].first_event;
+
+	/* common settings for all our events */
+	snd_seq_ev_clear(&ev);
+	ev.queue = queue;
+	ev.source.port = 0;
+	ev.flags = SND_SEQ_TIME_STAMP_TICK;
+
+	err = snd_seq_start_queue(seq, queue, NULL);
+	check_snd("start queue", err);
+	/* The queue won't be started until the START_QUEUE event is
+	 * actually drained to the kernel, which is exactly what we want. */
+
+	for (;;) {
+		struct event* event = NULL;
+		struct track* event_track = NULL;
+		int i, min_tick = max_tick + 1;
+
+		/* search next event */
+		for (i = 0; i < num_tracks; ++i) {
+			struct track *track = &tracks[i];
+			struct event *e2 = track->current_event;
+			if (e2 && e2->tick < min_tick) {
+				min_tick = e2->tick;
+				event = e2;
+				event_track = track;
+			}
+		}
+		if (!event)
+			break; /* end of song reached */
+
+		/* advance pointer to next event */
+		event_track->current_event = event->next;
+
+		/* output the event */
+		ev.type = event->type;
+		ev.time.tick = event->tick;
+		ev.dest = ports[event->port];
+		switch (ev.type) {
+		case SND_SEQ_EVENT_NOTEON:
+		case SND_SEQ_EVENT_NOTEOFF:
+		case SND_SEQ_EVENT_KEYPRESS:
+			snd_seq_ev_set_fixed(&ev);
+			ev.data.note.channel = event->data.d[0];
+			ev.data.note.note = event->data.d[1];
+			ev.data.note.velocity = event->data.d[2];
+			break;
+		case SND_SEQ_EVENT_CONTROLLER:
+			snd_seq_ev_set_fixed(&ev);
+			ev.data.control.channel = event->data.d[0];
+			ev.data.control.param = event->data.d[1];
+			ev.data.control.value = event->data.d[2];
+			break;
+		case SND_SEQ_EVENT_PGMCHANGE:
+		case SND_SEQ_EVENT_CHANPRESS:
+			snd_seq_ev_set_fixed(&ev);
+			ev.data.control.channel = event->data.d[0];
+			ev.data.control.value = event->data.d[1];
+			break;
+		case SND_SEQ_EVENT_PITCHBEND:
+			snd_seq_ev_set_fixed(&ev);
+			ev.data.control.channel = event->data.d[0];
+			ev.data.control.value =
+				((event->data.d[1]) |
+				 ((event->data.d[2]) << 7)) - 0x2000;
+			break;
+		case SND_SEQ_EVENT_SYSEX:
+			snd_seq_ev_set_variable(&ev, event->data.length,
+						event->sysex);
+			break;
+		case SND_SEQ_EVENT_TEMPO:
+			snd_seq_ev_set_fixed(&ev);
+			ev.dest.client = SND_SEQ_CLIENT_SYSTEM;
+			ev.dest.port = SND_SEQ_PORT_SYSTEM_TIMER;
+			ev.data.queue.queue = queue;
+			ev.data.queue.param.value = event->data.tempo;
+			break;
+		default:
+			fatal("Invalid event type %d!", ev.type);
+		}
+
+		/* this blocks when the output pool has been filled */
+		err = snd_seq_event_output(seq, &ev);
+		check_snd("output event", err);
+	}
+
+	/* schedule queue stop at end of song */
+	snd_seq_ev_set_fixed(&ev);
+	ev.type = SND_SEQ_EVENT_STOP;
+	ev.time.tick = max_tick;
+	ev.dest.client = SND_SEQ_CLIENT_SYSTEM;
+	ev.dest.port = SND_SEQ_PORT_SYSTEM_TIMER;
+	ev.data.queue.queue = queue;
+	err = snd_seq_event_output(seq, &ev);
+	check_snd("output event", err);
+
+	/* make sure that the sequencer sees all our events */
+	err = snd_seq_drain_output(seq);
+	check_snd("drain output", err);
+
+	/*
+	 * There are three possibilities how to wait until all events have
+	 * been played:
+	 * 1) send an event back to us (like pmidi does), and wait for it;
+	 * 2) wait for the EVENT_STOP notification for our queue which is sent
+	 *    by the system timer port (this would require a subscription);
+	 * 3) wait until the output pool is empty.
+	 * The last is the simplest.
+	 */
+	err = snd_seq_sync_output_queue(seq);
+	check_snd("sync output", err);
+
+	/* give the last notes time to die away */
+	if (end_delay > 0)
+		sleep(end_delay);
+}
+
+static void play_file(void)
+{
+	int ok;
+
+	if (!strcmp(file_name, "-"))
+		file = stdin;
+	else
+		file = fopen(file_name, "rb");
+	if (!file) {
+		errormsg("Cannot open %s - %s", file_name, strerror(errno));
+		return;
+	}
+
+	file_offset = 0;
+	ok = 0;
+
+	switch (read_id()) {
+	case MAKE_ID('M', 'T', 'h', 'd'):
+		ok = read_smf();
+		break;
+	case MAKE_ID('R', 'I', 'F', 'F'):
+		ok = read_riff();
+		break;
+	default:
+		errormsg("%s is not a Standard MIDI File", file_name);
+		break;
+	}
+
+	if (file != stdin)
+		fclose(file);
+
+	if (ok)
+		play_midi();
+
+	cleanup_file_data();
+}
+
+static void list_ports(void)
+{
+	snd_seq_client_info_t *cinfo;
+	snd_seq_port_info_t *pinfo;
+
+	snd_seq_client_info_alloca(&cinfo);
+	snd_seq_port_info_alloca(&pinfo);
+
+	puts(" Port    Client name                      Port name");
+
+	snd_seq_client_info_set_client(cinfo, -1);
+	while (snd_seq_query_next_client(seq, cinfo) >= 0) {
+		int client = snd_seq_client_info_get_client(cinfo);
+
+		snd_seq_port_info_set_client(pinfo, client);
+		snd_seq_port_info_set_port(pinfo, -1);
+		while (snd_seq_query_next_port(seq, pinfo) >= 0) {
+			/* we need both WRITE and SUBS_WRITE */
+			if ((snd_seq_port_info_get_capability(pinfo)
+			     & (SND_SEQ_PORT_CAP_WRITE | SND_SEQ_PORT_CAP_SUBS_WRITE))
+			    != (SND_SEQ_PORT_CAP_WRITE | SND_SEQ_PORT_CAP_SUBS_WRITE))
+				continue;
+			printf("%3d:%-3d  %-32.32s %s\n",
+			       snd_seq_port_info_get_client(pinfo),
+			       snd_seq_port_info_get_port(pinfo),
+			       snd_seq_client_info_get_name(cinfo),
+			       snd_seq_port_info_get_name(pinfo));
+		}
+	}
+}
+
+static void usage(const char *argv0)
+{
+	fprintf(stderr,
+		"Usage: %s -p client:port[,...] [-d delay] midifile ...\n"
+		"-h, --help                  this help\n"
+		"-V, --version               print current version\n"
+		"-l, --list                  list all possible output ports\n"
+		"-p, --port=client:port,...  set port(s) to play to\n"
+		"-d, --delay=seconds         delay after song ends\n",
+		argv0);
+}
+
+static void version(void)
+{
+	fputs("aplaymidi version " SND_UTIL_VERSION_STR "\n", stderr);
+}
+
+int main(int argc, char *argv[])
+{
+	static char short_options[] = "hVlp:d:";
+	static struct option long_options[] = {
+		{"help", 0, NULL, 'h'},
+		{"version", 0, NULL, 'V'},
+		{"list", 0, NULL, 'l'},
+		{"port", 1, NULL, 'p'},
+		{"delay", 1, NULL, 'd'},
+		{}
+	};
+	int c;
+	int do_list = 0;
+
+	init_seq();
+
+	while ((c = getopt_long(argc, argv, short_options,
+				long_options, NULL)) != -1) {
+		switch (c) {
+		case 'h':
+			usage(argv[0]);
+			return 0;
+		case 'V':
+			version();
+			return 0;
+		case 'l':
+			do_list = 1;
+			break;
+		case 'p':
+			parse_ports(optarg);
+			break;
+		case 'd':
+			end_delay = atoi(optarg);
+			break;
+		default:
+			usage(argv[0]);
+			return 1;
+		}
+	}
+
+	if (do_list) {
+		list_ports();
+	} else {
+		if (port_count < 1) {
+			/* use env var for compatibility with pmidi */
+			const char *ports_str = getenv("ALSA_OUTPUT_PORTS");
+			if (ports_str)
+				parse_ports(ports_str);
+			if (port_count < 1) {
+				errormsg("Please specify at least one port with --port.");
+				return 1;
+			}
+		}
+		if (optind >= argc) {
+			errormsg("Please specify a file to play.");
+			return 1;
+		}
+
+		create_source_port();
+		create_queue();
+		connect_ports();
+
+		for (; optind < argc; ++optind) {
+			file_name = argv[optind];
+			play_file();
+		}
+	}
+	snd_seq_close(seq);
+	return 0;
+}
diff -urN ../../oalsa/alsa/alsa-utils/seq/aplaymidi/arecordmidi.1 alsa-utils/seq/aplaymidi/arecordmidi.1
--- ../../oalsa/alsa/alsa-utils/seq/aplaymidi/arecordmidi.1	1970-01-01 01:00:00.000000000 +0100
+++ alsa-utils/seq/aplaymidi/arecordmidi.1	2004-02-23 09:38:11.101243000 +0100
@@ -0,0 +1,62 @@
+.TH ARECORDMIDI 1 "22 Feb 2004"
+
+.SH NAME
+arecordmidi - record Standard MIDI Files
+
+.SH SYNOPSIS
+.B arecordmidi
+-p client:port[,...] [options] midifile
+
+.SH DESCRIPTION
+.B arecordmidi
+is a command-line utility that records a Standard MIDI File from one or
+more ALSA sequencer ports.
+
+To stop recording, press Ctrl+C.
+
+.SH OPTIONS
+
+.TP
+.I -h,--help
+Prints a list of options.
+
+.TP
+.I -V,--version
+Prints the current version.
+
+.TP
+.I -l,--list
+Prints a list of possible input ports.
+
+.TP
+.I -p,--port=client:port,...
+Sets the sequencer port(s) from which events are recorded.
+
+A client can be specified by its number, its name, or a prefix of its
+name. A port is specified by its number; for port 0 of a client, the
+":0" part of the port specification can be omitted.
+
+.TP
+.I -b,--bpm=beats
+Sets the musical tempo of the MIDI file, in beats per minute.
+The default value is 120 BPM.
+
+.TP
+.I -f,--fps=frames
+Sets the SMPTE resolution, in frames per second.
+Possible values are 24, 25, 29.97 (for 30 drop-frame), and 30.
+
+.TP
+.I -t,--ticks=ticks
+Sets the resolution of timestamps (ticks) in the MIDI file,
+in ticks per beat (when using musical tempo) or ticks per frame
+(when using SMPTE timing).
+The default value is 384 ticks/beat or 40 ticks/frame, respectively.
+
+.TP
+.I -s,--split-channels
+Specifies that the data for each MIDI channel should be written to a
+separate track in the MIDI file.
+
+.SH AUTHOR
+Clemens Ladisch <clemens@ladisch.de>
diff -urN ../../oalsa/alsa/alsa-utils/seq/aplaymidi/arecordmidi.c alsa-utils/seq/aplaymidi/arecordmidi.c
--- ../../oalsa/alsa/alsa-utils/seq/aplaymidi/arecordmidi.c	1970-01-01 01:00:00.000000000 +0100
+++ alsa-utils/seq/aplaymidi/arecordmidi.c	2004-02-23 09:38:11.141242000 +0100
@@ -0,0 +1,723 @@
+/*
+ * arecordmidi.c - record standard MIDI files from sequencer ports
+ *
+ * Copyright (c) 2004 Clemens Ladisch <clemens@ladisch.de>
+ *
+ *
+ *  This program is free software; you can redistribute it and/or modify
+ *  it under the terms of the GNU General Public License as published by
+ *  the Free Software Foundation; either version 2 of the License, or
+ *  (at your option) any later version.
+ *
+ *  This program is distributed in the hope that it will be useful,
+ *  but WITHOUT ANY WARRANTY; without even the implied warranty of
+ *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ *  GNU General Public License for more details.
+ *
+ *  You should have received a copy of the GNU General Public License
+ *  along with this program; if not, write to the Free Software
+ *  Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307 USA
+ */
+
+/* TODO: sequencer queue timer selection */
+
+#include <stdio.h>
+#include <stdlib.h>
+#include <stdarg.h>
+#include <string.h>
+#include <signal.h>
+#include <getopt.h>
+#include <sys/poll.h>
+#include <alsa/asoundlib.h>
+#include "aconfig.h"
+#include "version.h"
+
+#define BUFFER_SIZE 4088
+
+/* linked list of buffers, stores data as in the .mid file */
+struct buffer {
+	struct buffer *next;
+	unsigned char buf[BUFFER_SIZE];
+};
+
+struct smf_track {
+	int size;			/* size of entire data */
+	int cur_buf_size;		/* size of cur_buf */
+	struct buffer *cur_buf;
+	snd_seq_tick_time_t last_tick;	/* end of track */
+	unsigned char last_command;	/* used for running status */
+	int used;			/* anything record on this track */
+	struct buffer first_buf;	/* list head */
+};
+
+/* timing/sysex + 16 channels */
+#define TRACKS_PER_PORT 17
+
+
+static snd_seq_t *seq;
+static int client;
+static int port_count;
+static snd_seq_addr_t *ports;
+static int queue;
+static int smpte_timing = 0;
+static int beats = 120;
+static int frames;
+static int ticks = 0;
+static FILE *file;
+static int channel_split;
+static int num_tracks;
+static struct smf_track *tracks;
+static volatile sig_atomic_t stop = 0;
+
+
+/* prints an error message to stderr, and dies */
+static void fatal(const char *msg, ...)
+{
+	va_list ap;
+
+	va_start(ap, msg);
+	vfprintf(stderr, msg, ap);
+	va_end(ap);
+	fputc('\n', stderr);
+	exit(EXIT_FAILURE);
+}
+
+/* memory allocation error handling */
+static void check_mem(void *p)
+{
+	if (!p)
+		fatal("Out of memory");
+}
+
+/* error handling for ALSA functions */
+static void check_snd(const char *operation, int err)
+{
+	if (err < 0)
+		fatal("Cannot %s - %s", operation, snd_strerror(err));
+}
+
+static void init_seq(void)
+{
+	int err;
+
+	/* open sequencer */
+	err = snd_seq_open(&seq, "default", SND_SEQ_OPEN_DUPLEX, 0);
+	check_snd("open sequencer", err);
+
+	/* find out our client's id */
+	client = snd_seq_client_id(seq);
+	check_snd("get client id", client);
+
+	/* set our client's name */
+	err = snd_seq_set_client_name(seq, "arecordmidi");
+	check_snd("set client name", err);
+}
+
+/* parses one or more port addresses from the string */
+static void parse_ports(const char *arg)
+{
+	char *buf, *s, *port_name;
+	int err;
+
+	/* make a copy of the string because we're going to modify it */
+	buf = strdup(arg);
+	check_mem(buf);
+
+	for (port_name = s = buf; s; port_name = s + 1) {
+		/* Assume that ports are separated by commas.  We don't use
+		 * spaces because those are valid in client names. */
+		s = strchr(port_name, ',');
+		if (s)
+			*s = '\0';
+
+		++port_count;
+		ports = realloc(ports, port_count * sizeof(snd_seq_addr_t));
+		check_mem(ports);
+
+		err = snd_seq_parse_address(seq, &ports[port_count - 1], port_name);
+		if (err < 0)
+			fatal("Invalid port %s - %s", port_name, snd_strerror(err));
+	}
+
+	free(buf);
+}
+
+static void init_tracks(void)
+{
+	int i;
+
+	/* MIDI RP-019 says we need at least one track per port */
+	num_tracks = port_count;
+	/* Allocate one track for each possible channel.
+	 * Empty tracks won't be written to the file. */
+	if (channel_split)
+		num_tracks *= TRACKS_PER_PORT;
+
+	tracks = calloc(num_tracks, sizeof(struct smf_track));
+	check_mem(tracks);
+	for (i = 0; i < num_tracks; ++i)
+		tracks[i].cur_buf = &tracks[i].first_buf;
+}
+
+static void create_queue(void)
+{
+	snd_seq_queue_tempo_t *tempo;
+	int err;
+
+	queue = snd_seq_alloc_named_queue(seq, "arecordmidi");
+	check_snd("create queue", queue);
+
+	snd_seq_queue_tempo_alloca(&tempo);
+	if (!smpte_timing) {
+		snd_seq_queue_tempo_set_tempo(tempo, 60000000 / beats);
+		snd_seq_queue_tempo_set_ppq(tempo, ticks);
+	} else {
+		/*
+		 * ALSA doesn't know about the SMPTE time divisions, so
+		 * we pretend to have a musical tempo with the equivalent
+		 * number of ticks/s.
+		 */
+		switch (frames) {
+		case 24:
+			snd_seq_queue_tempo_set_tempo(tempo, 500000);
+			snd_seq_queue_tempo_set_ppq(tempo, 12 * ticks);
+			break;
+		case 25:
+			snd_seq_queue_tempo_set_tempo(tempo, 400000);
+			snd_seq_queue_tempo_set_ppq(tempo, 10 * ticks);
+			break;
+		case 29:
+			snd_seq_queue_tempo_set_tempo(tempo, 100000000);
+			snd_seq_queue_tempo_set_ppq(tempo, 2997 * ticks);
+			break;
+		case 30:
+			snd_seq_queue_tempo_set_tempo(tempo, 500000);
+			snd_seq_queue_tempo_set_ppq(tempo, 15 * ticks);
+			break;
+		default:
+			fatal("Invalid SMPTE frames %d", frames);
+		}
+	}
+	err = snd_seq_set_queue_tempo(seq, queue, tempo);
+	if (err < 0)
+		fatal("Cannot set queue tempo (%u/%i)",
+		      snd_seq_queue_tempo_get_tempo(tempo),
+		      snd_seq_queue_tempo_get_ppq(tempo));
+}
+
+static void create_ports(void)
+{
+	snd_seq_port_info_t *pinfo;
+	int i, err;
+	char name[32];
+
+	snd_seq_port_info_alloca(&pinfo);
+
+	/* common information for all our ports */
+	snd_seq_port_info_set_capability(pinfo,
+					 SND_SEQ_PORT_CAP_WRITE |
+					 SND_SEQ_PORT_CAP_SUBS_WRITE);
+	snd_seq_port_info_set_type(pinfo,
+				   SND_SEQ_PORT_TYPE_MIDI_GENERIC |
+				   SND_SEQ_PORT_TYPE_APPLICATION);
+	snd_seq_port_info_set_midi_channels(pinfo, 16);
+
+	/* we want to know when the events got delivered to us */
+	snd_seq_port_info_set_timestamping(pinfo, 1);
+	snd_seq_port_info_set_timestamp_queue(pinfo, queue);
+
+	/* our port number is the same as our port index */
+	snd_seq_port_info_set_port_specified(pinfo, 1);
+	for (i = 0; i < port_count; ++i) {
+		snd_seq_port_info_set_port(pinfo, i);
+
+		sprintf(name, "arecordmidi port %i", i);
+		snd_seq_port_info_set_name(pinfo, name);
+
+		err = snd_seq_create_port(seq, pinfo);
+		check_snd("create port", err);
+	}
+}
+
+static void connect_ports(void)
+{
+	int i, err;
+
+	for (i = 0; i < port_count; ++i) {
+		err = snd_seq_connect_from(seq, i, ports[i].client, ports[i].port);
+		if (err < 0)
+			fatal("Cannot connect from port %d:%d - %s",
+			      ports[i].client, ports[i].port, snd_strerror(err));
+	}
+}
+
+/* records a byte to be written to the .mid file */
+static void add_byte(struct smf_track *track, unsigned char byte)
+{
+	/* make sure we have enough room in the current buffer */
+	if (track->cur_buf_size >= BUFFER_SIZE) {
+		track->cur_buf->next = calloc(1, sizeof(struct buffer));
+		if (!track->cur_buf->next)
+			fatal("out of memory");
+		track->cur_buf = track->cur_buf->next;
+		track->cur_buf_size = 0;
+	}
+
+	track->cur_buf->buf[track->cur_buf_size++] = byte;
+	track->size++;
+}
+
+/* record a variable-length quantity */
+static void var_value(struct smf_track *track, int v)
+{
+	if (v >= (1 << 28))
+		add_byte(track, 0x80 | ((v >> 28) & 0x03));
+	if (v >= (1 << 21))
+		add_byte(track, 0x80 | ((v >> 21) & 0x7f));
+	if (v >= (1 << 14))
+		add_byte(track, 0x80 | ((v >> 14) & 0x7f));
+	if (v >= (1 << 7))
+		add_byte(track, 0x80 | ((v >> 7) & 0x7f));
+	add_byte(track, v & 0x7f);
+}
+
+/* record the delta time from the last event */
+static void delta_time(struct smf_track *track, const snd_seq_event_t *ev)
+{
+	int diff = ev->time.tick - track->last_tick;
+	if (diff < 0)
+		diff = 0;
+	var_value(track, diff);
+	track->last_tick = ev->time.tick;
+}
+
+/* record a status byte (or not if we can use running status) */
+static void command(struct smf_track *track, unsigned char cmd)
+{
+	if (cmd != track->last_command)
+		add_byte(track, cmd);
+	track->last_command = cmd < 0xf0 ? cmd : 0;
+}
+
+/* put port numbers into all tracks */
+static void record_port_numbers(void)
+{
+	int i;
+
+	for (i = 0; i < num_tracks; ++i) {
+		var_value(&tracks[i], 0);
+		add_byte(&tracks[i], 0xff);
+		add_byte(&tracks[i], 0x21);
+		var_value(&tracks[i], 1);
+		if (channel_split)
+			add_byte(&tracks[i], i / TRACKS_PER_PORT);
+		else
+			add_byte(&tracks[i], i);
+	}
+}
+
+static void record_event(const snd_seq_event_t *ev)
+{
+	unsigned int i;
+	struct smf_track *track;
+
+	/* ignore events without proper timestamps */
+	if (ev->queue != queue || !snd_seq_ev_is_tick(ev))
+		return;
+
+	/* determine which track to record to */
+	i = ev->dest.port;
+	if (channel_split) {
+		i *= TRACKS_PER_PORT;
+		if (snd_seq_ev_is_channel_type(ev))
+			i += 1 + (ev->data.note.channel & 0xf);
+	}
+	if (i >= num_tracks)
+		return;
+	track = &tracks[i];
+
+	switch (ev->type) {
+	case SND_SEQ_EVENT_NOTEON:
+		delta_time(track, ev);
+		command(track, MIDI_CMD_NOTE_ON | (ev->data.note.channel & 0xf));
+		add_byte(track, ev->data.note.note & 0x7f);
+		add_byte(track, ev->data.note.velocity & 0x7f);
+		break;
+	case SND_SEQ_EVENT_NOTEOFF:
+		delta_time(track, ev);
+		command(track, MIDI_CMD_NOTE_OFF | (ev->data.note.channel & 0xf));
+		add_byte(track, ev->data.note.note & 0x7f);
+		add_byte(track, ev->data.note.velocity & 0x7f);
+		break;
+	case SND_SEQ_EVENT_KEYPRESS:
+		delta_time(track, ev);
+		command(track, MIDI_CMD_NOTE_PRESSURE | (ev->data.note.channel & 0xf));
+		add_byte(track, ev->data.note.note & 0x7f);
+		add_byte(track, ev->data.note.velocity & 0x7f);
+		break;
+	case SND_SEQ_EVENT_CONTROLLER:
+		delta_time(track, ev);
+		command(track, MIDI_CMD_CONTROL | (ev->data.control.channel & 0xf));
+		add_byte(track, ev->data.control.param & 0x7f);
+		add_byte(track, ev->data.control.value & 0x7f);
+		break;
+	case SND_SEQ_EVENT_PGMCHANGE:
+		delta_time(track, ev);
+		command(track, MIDI_CMD_PGM_CHANGE | (ev->data.control.channel & 0xf));
+		add_byte(track, ev->data.control.value & 0x7f);
+		break;
+	case SND_SEQ_EVENT_CHANPRESS:
+		delta_time(track, ev);
+		command(track, MIDI_CMD_CHANNEL_PRESSURE | (ev->data.control.channel & 0xf));
+		add_byte(track, ev->data.control.value & 0x7f);
+		break;
+	case SND_SEQ_EVENT_PITCHBEND:
+		delta_time(track, ev);
+		command(track, MIDI_CMD_BENDER | (ev->data.control.channel & 0xf));
+		add_byte(track, (ev->data.control.value + 8192) & 0x7f);
+		add_byte(track, ((ev->data.control.value + 8192) >> 7) & 0x7f);
+		break;
+	case SND_SEQ_EVENT_CONTROL14:
+		/* create two commands for MSB and LSB */
+		delta_time(track, ev);
+		command(track, MIDI_CMD_CONTROL | (ev->data.control.channel & 0xf));
+		add_byte(track, ev->data.control.param & 0x7f);
+		add_byte(track, (ev->data.control.value >> 7) & 0x7f);
+		if ((ev->data.control.param & 0x7f) < 0x20) {
+			delta_time(track, ev);
+			/* running status */
+			add_byte(track, (ev->data.control.param & 0x7f) + 0x20);
+			add_byte(track, ev->data.control.value & 0x7f);
+		}
+		break;
+	case SND_SEQ_EVENT_NONREGPARAM:
+		delta_time(track, ev);
+		command(track, MIDI_CMD_CONTROL | (ev->data.control.channel & 0xf));
+		add_byte(track, MIDI_CTL_NONREG_PARM_NUM_LSB);
+		add_byte(track, ev->data.control.param & 0x7f);
+		delta_time(track, ev);
+		add_byte(track, MIDI_CTL_NONREG_PARM_NUM_MSB);
+		add_byte(track, (ev->data.control.param >> 7) & 0x7f);
+		delta_time(track, ev);
+		add_byte(track, MIDI_CTL_MSB_DATA_ENTRY);
+		add_byte(track, (ev->data.control.value >> 7) & 0x7f);
+		delta_time(track, ev);
+		add_byte(track, MIDI_CTL_LSB_DATA_ENTRY);
+		add_byte(track, ev->data.control.value & 0x7f);
+		break;
+	case SND_SEQ_EVENT_REGPARAM:
+		delta_time(track, ev);
+		command(track, MIDI_CMD_CONTROL | (ev->data.control.channel & 0xf));
+		add_byte(track, MIDI_CTL_REGIST_PARM_NUM_LSB);
+		add_byte(track, ev->data.control.param & 0x7f);
+		delta_time(track, ev);
+		add_byte(track, MIDI_CTL_REGIST_PARM_NUM_MSB);
+		add_byte(track, (ev->data.control.param >> 7) & 0x7f);
+		delta_time(track, ev);
+		add_byte(track, MIDI_CTL_MSB_DATA_ENTRY);
+		add_byte(track, (ev->data.control.value >> 7) & 0x7f);
+		delta_time(track, ev);
+		add_byte(track, MIDI_CTL_LSB_DATA_ENTRY);
+		add_byte(track, ev->data.control.value & 0x7f);
+		break;
+#if 0	/* ignore */
+	case SND_SEQ_EVENT_SONGPOS:
+	case SND_SEQ_EVENT_SONGSEL:
+	case SND_SEQ_EVENT_QFRAME:
+	case SND_SEQ_EVENT_START:
+	case SND_SEQ_EVENT_CONTINUE:
+	case SND_SEQ_EVENT_STOP:
+	case SND_SEQ_EVENT_TUNE_REQUEST:
+	case SND_SEQ_EVENT_RESET:
+	case SND_SEQ_EVENT_SENSING:
+		break;
+#endif
+	case SND_SEQ_EVENT_SYSEX:
+		if (ev->data.ext.len == 0)
+			break;
+		delta_time(track, ev);
+		if (*(unsigned char*)ev->data.ext.ptr == 0xf0)
+			command(track, 0xf0), i = 1;
+		else
+			command(track, 0xf7), i = 0;
+		var_value(track, ev->data.ext.len - i);
+		for (; i < ev->data.ext.len; ++i)
+			add_byte(track, ((unsigned char*)ev->data.ext.ptr)[i]);
+		break;
+	default:
+		return;
+	}
+	track->used = 1;
+}
+
+static void finish_tracks(void)
+{
+	snd_seq_queue_status_t *queue_status;
+	int tick, i, err;
+
+	snd_seq_queue_status_alloca(&queue_status);
+
+	err = snd_seq_get_queue_status(seq, queue, queue_status);
+	check_snd("get queue status", err);
+	tick = snd_seq_queue_status_get_tick_time(queue_status);
+
+	/* make length of first track the recording length */
+	var_value(&tracks[0], tick - tracks[0].last_tick);
+	add_byte(&tracks[0], 0xff);
+	add_byte(&tracks[0], 0x2f);
+	var_value(&tracks[0], 0);
+
+	/* finish other tracks */
+	for (i = 1; i < num_tracks; ++i) {
+		var_value(&tracks[i], 0);
+		add_byte(&tracks[i], 0xff);
+		add_byte(&tracks[i], 0x2f);
+		var_value(&tracks[i], 0);
+	}
+}
+
+static void write_file(void)
+{
+	int used_tracks, time_division, i;
+	struct buffer *buf;
+
+	used_tracks = 0;
+	for (i = 0; i < num_tracks; ++i)
+		used_tracks += !!tracks[i].used;
+
+	/* header id and length */
+	fwrite("MThd\0\0\0\6", 1, 8, file);
+	/* type 0 or 1 */
+	fputc(0, file);
+	fputc(used_tracks > 1, file);
+	/* number of tracks */
+	fputc((used_tracks >> 8) & 0xff, file);
+	fputc(used_tracks & 0xff, file);
+	/* time division */
+	time_division = ticks;
+	if (smpte_timing)
+		time_division |= (0x100 - frames) << 8;
+	fputc(time_division >> 8, file);
+	fputc(time_division & 0xff, file);
+
+	for (i = 0; i < num_tracks; ++i) {
+		if (!tracks[i].used)
+			continue;
+		/* track id */
+		fwrite("MTrk", 1, 4, file);
+		/* data length */
+		fputc((tracks[i].size >> 24) & 0xff, file);
+		fputc((tracks[i].size >> 16) & 0xff, file);
+		fputc((tracks[i].size >> 8) & 0xff, file);
+		fputc(tracks[i].size & 0xff, file);
+		/* track contents */
+		for (buf = &tracks[i].first_buf; buf; buf = buf->next)
+			fwrite(buf->buf, 1, buf == tracks[i].cur_buf
+			       ? tracks[i].cur_buf_size : BUFFER_SIZE, file);
+	}
+}
+
+static void list_ports(void)
+{
+	snd_seq_client_info_t *cinfo;
+	snd_seq_port_info_t *pinfo;
+
+	snd_seq_client_info_alloca(&cinfo);
+	snd_seq_port_info_alloca(&pinfo);
+
+	puts(" Port    Client name                      Port name");
+
+	snd_seq_client_info_set_client(cinfo, -1);
+	while (snd_seq_query_next_client(seq, cinfo) >= 0) {
+		int client = snd_seq_client_info_get_client(cinfo);
+
+		if (client == SND_SEQ_CLIENT_SYSTEM)
+			continue; /* don't show system timer and announce ports */
+		snd_seq_port_info_set_client(pinfo, client);
+		snd_seq_port_info_set_port(pinfo, -1);
+		while (snd_seq_query_next_port(seq, pinfo) >= 0) {
+			/* we need both READ and SUBS_READ */
+			if ((snd_seq_port_info_get_capability(pinfo)
+			     & (SND_SEQ_PORT_CAP_READ | SND_SEQ_PORT_CAP_SUBS_READ))
+			    != (SND_SEQ_PORT_CAP_READ | SND_SEQ_PORT_CAP_SUBS_READ))
+				continue;
+			printf("%3d:%-3d  %-32.32s %s\n",
+			       snd_seq_port_info_get_client(pinfo),
+			       snd_seq_port_info_get_port(pinfo),
+			       snd_seq_client_info_get_name(cinfo),
+			       snd_seq_port_info_get_name(pinfo));
+		}
+	}
+}
+
+static void help(const char *argv0)
+{
+	fprintf(stderr, "Usage: %s [options] outputfile\n"
+		"\nAvailable options:\n"
+		"  -h,--help                  this help\n"
+		"  -V,--version               show version\n"
+		"  -l,--list                  list input ports\n"
+		"  -p,--port=client:port,...  source port(s)\n"
+		"  -b,--bpm=beats             tempo in beats per minute\n"
+		"  -f,--fps=frames            resolution in frames per second (SMPTE)\n"
+		"  -t,--ticks=ticks           resolution in ticks per beat or frame\n"
+		"  -s,--split-channels        create a track for each channel\n",
+		argv0);
+}
+
+static void version(void)
+{
+	fputs("arecordmidi version " SND_UTIL_VERSION_STR "\n", stderr);
+}
+
+static void sighandler(int sig)
+{
+	stop = 1;
+}
+
+int main(int argc, char *argv[])
+{
+	static char short_options[] = "hVlp:b:f:t:s";
+	static struct option long_options[] = {
+		{"help", 0, NULL, 'h'},
+		{"version", 0, NULL, 'V'},
+		{"list", 0, NULL, 'l'},
+		{"port", 1, NULL, 'p'},
+		{"bpm", 1, NULL, 'b'},
+		{"fps", 1, NULL, 'f'},
+		{"ticks", 1, NULL, 't'},
+		{"split-channels", 0, NULL, 's'},
+		{ }
+	};
+
+	char *filename = NULL;
+	int do_list = 0;
+	struct pollfd *pfds;
+	int npfds;
+	int c, err;
+
+	init_seq();
+
+	while ((c = getopt_long(argc, argv, short_options,
+				long_options, NULL)) != -1) {
+		switch (c) {
+		case 'h':
+			help(argv[0]);
+			return 0;
+		case 'V':
+			version();
+			return 0;
+		case 'l':
+			do_list = 1;
+			break;
+		case 'p':
+			parse_ports(optarg);
+			break;
+		case 'b':
+			beats = atoi(optarg);
+			if (beats < 4 || beats > 6000)
+				fatal("Invalid tempo");
+			smpte_timing = 0;
+			break;
+		case 'f':
+			frames = atoi(optarg);
+			if (frames != 24 && frames != 25 &&
+			    frames != 29 && frames != 30)
+				fatal("Invalid number of frames/s");
+			smpte_timing = 1;
+			break;
+		case 't':
+			ticks = atoi(optarg);
+			if (ticks < 1 || ticks > 0x7fff)
+				fatal("Invalid number of ticks");
+			break;
+		case 's':
+			channel_split = 1;
+			break;
+		default:
+			help(argv[0]);
+			return 1;
+		}
+	}
+
+	if (do_list) {
+		list_ports();
+		return 0;
+	}
+
+	if (port_count < 1) {
+		fputs("Pleast specify a source port with --port.\n", stderr);
+		return 1;
+	}
+
+	if (!ticks)
+		ticks = smpte_timing ? 40 : 384;
+	if (smpte_timing && ticks > 0xff)
+		ticks = 0xff;
+
+	if (optind >= argc) {
+		fputs("Please specify a file to record to.\n", stderr);
+		return 1;
+	}
+	filename = argv[optind];
+
+	init_tracks();
+	create_queue();
+	create_ports();
+	connect_ports();
+	if (port_count > 1)
+		record_port_numbers();
+
+	/* record tempo */
+	if (!smpte_timing) {
+		int usecs_per_quarter = 60000000 / beats;
+		var_value(&tracks[0], 0); /* delta time */
+		add_byte(&tracks[0], 0xff);
+		add_byte(&tracks[0], 0x51);
+		var_value(&tracks[0], 3);
+		add_byte(&tracks[0], usecs_per_quarter >> 16);
+		add_byte(&tracks[0], usecs_per_quarter >> 8);
+		add_byte(&tracks[0], usecs_per_quarter);
+	}
+	/* always write at least one track */
+	tracks[0].used = 1;
+
+	file = fopen(filename, "wb");
+	if (!file)
+		fatal("Cannot open %s - %s", filename, strerror(errno));
+
+	err = snd_seq_start_queue(seq, queue, NULL);
+	check_snd("start queue", err);
+	snd_seq_drain_output(seq);
+
+	err = snd_seq_nonblock(seq, 1);
+	check_snd("set nonblock mode", err);
+
+	signal(SIGINT, sighandler);
+	signal(SIGTERM, sighandler);
+
+	npfds = snd_seq_poll_descriptors_count(seq, POLLIN);
+	pfds = alloca(sizeof(*pfds) * npfds);
+	for (;;) {
+		snd_seq_poll_descriptors(seq, pfds, npfds, POLLIN);
+		if (poll(pfds, npfds, 69) < 0)
+			break;
+		do {
+			snd_seq_event_t *event;
+			err = snd_seq_event_input(seq, &event);
+			if (err < 0)
+				break;
+			if (event)
+				record_event(event);
+		} while (err > 0);
+		if (stop)
+			break;
+	}
+
+	finish_tracks();
+	write_file();
+
+	fclose(file);
+	snd_seq_close(seq);
+	return 0;
+}




-------------------------------------------------------
SF.Net is sponsored by: Speed Start Your Linux Apps Now.
Build and deploy apps & Web services for Linux with
a free DVD software kit from IBM. Click Now!
http://ads.osdn.com/?ad_id=1356&alloc_id=3438&op=click

Re: [PATCH] add aplaymidi & arecordmidi utils

[Jaroslav Kysela]

On Mon, 23 Feb 2004, Clemens Ladisch wrote:

> See man pages for details.
> 
> 
> add aplaymidi & arecordmidi utilities

Very nice work. Thank you.

						Jaroslav

-----
Jaroslav Kysela <perex@suse.cz>
Linux Kernel Sound Maintainer
ALSA Project, SuSE Labs


-------------------------------------------------------
SF.Net is sponsored by: Speed Start Your Linux Apps Now.
Build and deploy apps & Web services for Linux with
a free DVD software kit from IBM. Click Now!
http://ads.osdn.com/?ad_id=1356&alloc_id=3438&op=click

Re: [PATCH] add aplaymidi & arecordmidi utils

[Will]

Clemens Ladisch <clemens@ladisch.de> wrote:
>
>add aplaymidi & arecordmidi utilities
>

 I'm getting empty MIDI files containing only the MIDI header.

 $ arecordmidi -l
   Port    Client name                      Port name
   64:0    Rawmidi 0 - Sound Blaster 16 MPU Sound Blaster 16 MPU-401

 $ cat /proc/asound/card0/midi0
   Sound Blaster 16 MPU-401

   Output 0
     Tx bytes     : 9856
   Input 0
     Rx bytes     : 205806

 $ arecordmidi -p 64:0 music.mid
   <Ctrl-C>

 $ cat /proc/asound/card0/midi0
   Sound Blaster 16 MPU-401

   Output 0
     Tx bytes     : 9856
   Input 0
     Rx bytes     : 205946

 $ ls -l music.mid
   -rw-r--r--    1 uab1  gab1          34 Feb 24 17:07 music.mid

 $ file music.mid
   music.mid:  Standard MIDI data

 $ od -c music.mid
   000000  77  84 104 100   0   0   0   6   0   0   0   1   1 128  77  84
            M   T   h   d  \0  \0  \0 006  \0  \0  \0 001 001 200   M   T
   000010 114 107   0   0   0  12   0 255  81   3   7 161  32 150  96 255
            r   k  \0  \0  \0  \f  \0 377   Q 003  \a 241     226   ` 377
   000020  47   0
            /  \0
 
Any ideas?  I can record MIDI files ok using the old OSS sound drivers
and OSS sequencer software.

William


-------------------------------------------------------
SF.Net is sponsored by: Speed Start Your Linux Apps Now.
Build and deploy apps & Web services for Linux with
a free DVD software kit from IBM. Click Now!
http://ads.osdn.com/?ad_id=1356&alloc_id=3438&op=click

Re: [PATCH] add aplaymidi & arecordmidi utils

[Will]

Clemens Ladisch <clemens@ladisch.de> wrote:
>
>add aplaymidi & arecordmidi utilities
>

Having these command-line utilities is very useful.  Previously there
were only the old OSS command-line programs like (x)playmidi and drvmidi.

How easy would it be to make aplaymidi be able to adjust the playback speed
like playmidi?  (playmidi -t 0.5 music.mid plays at twice the default speed)

How easy would it be to make aplaymidi be able to move forwards and backwards
during playback, e.g. by pressing f or b, like Takashi Iwai's drvmidi?

William



-------------------------------------------------------
SF.Net is sponsored by: Speed Start Your Linux Apps Now.
Build and deploy apps & Web services for Linux with
a free DVD software kit from IBM. Click Now!
http://ads.osdn.com/?ad_id=1356&alloc_id=3438&op=click

Re: [PATCH] add aplaymidi & arecordmidi utils

[Clemens Ladisch]

Will wrote:

> Clemens Ladisch <clemens@ladisch.de> wrote:
> >
> >add aplaymidi & arecordmidi utilities
>
>  I'm getting empty MIDI files containing only the MIDI header.

Are you using the latest CVS version of ALSA?
Some of the bugfixes I sent yesterday are needed by arecordmidi.


HTH
Clemens




-------------------------------------------------------
SF.Net is sponsored by: Speed Start Your Linux Apps Now.
Build and deploy apps & Web services for Linux with
a free DVD software kit from IBM. Click Now!
http://ads.osdn.com/?ad_id=1356&alloc_id=3438&op=click

Re: [PATCH] add aplaymidi & arecordmidi utils

[Clemens Ladisch]

Will wrote:
> How easy would it be to make aplaymidi be able to adjust the playback speed
> like playmidi?  (playmidi -t 0.5 music.mid plays at twice the default speed)

Very easy.

> How easy would it be to make aplaymidi be able to move forwards and backwards
> during playback, e.g. by pressing f or b, like Takashi Iwai's drvmidi?

Not so easy; even more so if you want to have correct controller
values.  Adjusting settings changed by sysexs would be almost
impossible.


However, these tools are meant as basic command-line utilities (and
examples).  I think your suggestions would fit better into some GUI
player.


Regards,
Clemens




-------------------------------------------------------
SF.Net is sponsored by: Speed Start Your Linux Apps Now.
Build and deploy apps & Web services for Linux with
a free DVD software kit from IBM. Click Now!
http://ads.osdn.com/?ad_id=1356&alloc_id=3438&op=click

Re: [PATCH] add aplaymidi & arecordmidi utils

[Will]

Clemens Ladisch <clemens@ladisch.de> wrote:
>>
>>  I'm getting empty MIDI files containing only the MIDI header.
>
>Are you using the latest CVS version of ALSA?
>Some of the bugfixes I sent yesterday are needed by arecordmidi.
>

No, but I applied your patches to alsa-lib-1.0.3rc1 and Linux 2.6.3 in
include/sound/asequencer, sound/* etc.  Is that enough to get a
working ALSA sequencer?

Will



-------------------------------------------------------
SF.Net is sponsored by: Speed Start Your Linux Apps Now.
Build and deploy apps & Web services for Linux with
a free DVD software kit from IBM. Click Now!
http://ads.osdn.com/?ad_id=1356&alloc_id=3438&op=click

Re: [PATCH] add aplaymidi & arecordmidi utils

[Clemens Ladisch]

Will wrote:
> Clemens Ladisch <clemens@ladisch.de> wrote:
> >>  I'm getting empty MIDI files containing only the MIDI header.
> >
> >Are you using the latest CVS version of ALSA?
> >Some of the bugfixes I sent yesterday are needed by arecordmidi.
>
> No, but I applied your patches to alsa-lib-1.0.3rc1 and Linux 2.6.3 in
> include/sound/asequencer, sound/* etc.  Is that enough to get a
> working ALSA sequencer?

If you recompiled both alsa-driver and alsa-lib packages, yes.

Perhaps I should test more; the latest tests were made only against
data recorded through the snd-seq-dummy client (this is where the
unpatched sequencer would fail).

If you get correct output when you run "amidi -p virtual -d" and
aconnect your keyboard to that port, then it may be a bug in
arecordmidi.  I'll test when I'm at home.


Regards,
Clemens




-------------------------------------------------------
SF.Net is sponsored by: Speed Start Your Linux Apps Now.
Build and deploy apps & Web services for Linux with
a free DVD software kit from IBM. Click Now!
http://ads.osdn.com/?ad_id=1356&alloc_id=3438&op=click

Re: [PATCH] add aplaymidi & arecordmidi utils

[Will]

Clemens Ladisch <clemens@ladisch.de> wrote:
>
>If you get correct output when you run "amidi -p virtual -d" and
>aconnect your keyboard to that port, then it may be a bug in
>arecordmidi.  I'll test when I'm at home.
>

This is what I get:

  $ amidi -p virtual -d &
 
  $ pmidi -l
    Port     Client name                       Port name
    64:0     Rawmidi 0 - Sound Blaster 16 M    Sound Blaster 16 MPU-401
    65:0     Emu8000 WaveTable                 Emu8000 Port 0
    65:1     Emu8000 WaveTable                 Emu8000 Port 1
    65:2     Emu8000 WaveTable                 Emu8000 Port 2
    65:3     Emu8000 WaveTable                 Emu8000 Port 3
    66:0     OPL3 FM synth                     OPL3 FM Port
    72:0     Virtual Raw MIDI 1-0              VirMIDI 1-0
    73:0     Virtual Raw MIDI 1-1              VirMIDI 1-1
    74:0     Virtual Raw MIDI 1-2              VirMIDI 1-2
    75:0     Virtual Raw MIDI 1-3              VirMIDI 1-3
   128:0     Client-128                        Virtual RawMIDI

 The MIDI keyboard is on port 64:0:

 $ aconnect 64:0 128:0

playing various notes on the MIDI keyboard gives correct MIDI output in ASCII:

   90 29 4A
   90 29 00
   90 2D 4C
      2D 00
   90 2B 5A
   90 2B 00
   90 28 59
   90 28 00
   90 24 26
      24 00
      24 51
   90 24 00
   90 24 63
      24 00
   90 24 75
   90 24 00
   90 24 7B
   90 24 00
      24 65
   90 24 00
      24 50
   90 24 00
   90 24 3D
   90 24 00
   90 24 47
   90 24 00
   90 24 44
   90 24 00
   90 24 3A
   90 24 00
   90 26 6B
   90 26 00
      28 68
   90 28 00
      29 5F
      2B 68
   90 29 00
      29 65
   90 2B 00
      28 5E
      29 00
   90 26 69
      28 00
      24 70
   etc  

William


-------------------------------------------------------
SF.Net is sponsored by: Speed Start Your Linux Apps Now.
Build and deploy apps & Web services for Linux with
a free DVD software kit from IBM. Click Now!
http://ads.osdn.com/?ad_id=1356&alloc_id=3438&op=click

Re: [PATCH] add aplaymidi & arecordmidi utils

[Clemens Ladisch]

Will wrote:

> Clemens Ladisch <clemens@ladisch.de> wrote:
> >
> >If you get correct output when you run "amidi -p virtual -d" and
> >aconnect your keyboard to that port, then it may be a bug in
> >arecordmidi.  I'll test when I'm at home.
>
> This is what I get:
>
>   $ pmidi -l
>     Port     Client name                       Port name
>     64:0     Rawmidi 0 - Sound Blaster 16 M    Sound Blaster 16 MPU-401

You didn't apply all my patches, or recompile; this name would be
different.

Did you change asequencer.h in both the kernel and alsa-lib?

> playing various notes on the MIDI keyboard gives correct MIDI output in ASCII:
>
>    90 29 4A
>    ...

On my machine, arecordmidi works just fine.


Regards,
Clemens




-------------------------------------------------------
SF.Net is sponsored by: Speed Start Your Linux Apps Now.
Build and deploy apps & Web services for Linux with
a free DVD software kit from IBM. Click Now!
http://ads.osdn.com/?ad_id=1356&alloc_id=3438&op=click

Re: [PATCH] add aplaymidi & arecordmidi utils

[Will]

Clemens Ladisch <clemens@ladisch.de> wrote:
>Will wrote:
>> How easy would it be to make aplaymidi be able to adjust the playback speed
>> like playmidi?  (playmidi -t 0.5 music.mid plays at twice the default speed)
>
>Very easy.

1 hour of code changes?

>> How easy would it be to make aplaymidi be able to move forwards and backwards
>> during playback, e.g. by pressing f or b, like Takashi Iwai's drvmidi?
>
>Not so easy; even more so if you want to have correct controller
>values.  Adjusting settings changed by sysexs would be almost
>impossible.
>

I wonder whether drvmidi actually does correctly restore controller and sysex
when you move during MIDI playback.  I haven't studied the drvmidi code in
detail.  I know the sound quality always seems to be correct
even when you move forwards and backwards very often in a piece with
lots of sysex and controller settings.

>However, these tools are meant as basic command-line utilities (and
>examples).  I think your suggestions would fit better into some GUI
>player.

Yes, you're right a GUI player can be much more powerful but I also like
command-line players like drvmidi because you can always run them without X
and moving forwards and backwards using the ncurses interface is easy.
It's a useful minimum functionality (being able to move forwards and backwards
in a MIDI file while it's playing) but it sounds like it'd be difficult
to extend arecordmidi to do this.

William



-------------------------------------------------------
SF.Net is sponsored by: Speed Start Your Linux Apps Now.
Build and deploy apps & Web services for Linux with
a free DVD software kit from IBM. Click Now!
http://ads.osdn.com/?ad_id=1356&alloc_id=3438&op=click

Re: [PATCH] add aplaymidi & arecordmidi utils

[Will]

Clemens Ladisch <clemens@ladisch.de> wrote:
>
>Did you change asequencer.h in both the kernel and alsa-lib?

No, I didn't see my script had actually failed to patch
include/sound/asequencer.h

BTW I know there are several correct ways of updating the ALSA in Linux 2.6.x
according to the docs at Wiki alsa.opensrc.org and alsa-project/alsa-doc/
Is the following a correct way of updating the ALSA in Linux 2.6.3?

1. Get the latest main CVS ALSA:

cvs -z3 -d:pserver:anonymous@cvs.sourceforge.net:/cvsroot/alsa co alsa-driver
cvs -z3 -d:pserver:anonymous@cvs.sourceforge.net:/cvsroot/alsa co alsa-kernel
cvs -z3 -d:pserver:anonymous@cvs.sourceforge.net:/cvsroot/alsa co alsa-firmware
cvs -z3 -d:pserver:anonymous@cvs.sourceforge.net:/cvsroot/alsa co alsa-lib
cvs -z3 -d:pserver:anonymous@cvs.sourceforge.net:/cvsroot/alsa co alsa-oss
cvs -z3 -d:pserver:anonymous@cvs.sourceforge.net:/cvsroot/alsa co alsa-utils
cvs -z3 -d:pserver:anonymous@cvs.sourceforge.net:/cvsroot/alsa co alsa-tools

2. configure and make install everything including
   firstly alsa-driver and secondly alsa-kernel 

3. update Linux kernel headers and source files

  mv alsa-kernel/include/*.h linux-kernel-source/include/sound
  for i in arm core drivers i2c isa oss p* sparc synth usb; do
     pseudo shell script:
     mv alsa-kernel/$i/[./]*/*.[ch] linux-kernel-source/sound/subdirectories
  done


-------------------------------------------------------
SF.Net is sponsored by: Speed Start Your Linux Apps Now.
Build and deploy apps & Web services for Linux with
a free DVD software kit from IBM. Click Now!
http://ads.osdn.com/?ad_id=1356&alloc_id=3438&op=click

Re: [PATCH] add aplaymidi & arecordmidi utils

[James Courtier-Dutton]

Will wrote:
> Clemens Ladisch <clemens@ladisch.de> wrote:
> 
>>Did you change asequencer.h in both the kernel and alsa-lib?
> 
> 
> No, I didn't see my script had actually failed to patch
> include/sound/asequencer.h
> 
> BTW I know there are several correct ways of updating the ALSA in Linux 2.6.x
> according to the docs at Wiki alsa.opensrc.org and alsa-project/alsa-doc/
> Is the following a correct way of updating the ALSA in Linux 2.6.3?

I have just updated the alsa.opensrc.org wiki.
Use option 1] on
http://alsa.opensrc.org/index.php?page=AlsaBuild2.6

Cheers
James




-------------------------------------------------------
SF.Net is sponsored by: Speed Start Your Linux Apps Now.
Build and deploy apps & Web services for Linux with
a free DVD software kit from IBM. Click Now!
http://ads.osdn.com/?ad_id=1356&alloc_id=3438&op=click

Re: [PATCH] add aplaymidi & arecordmidi utils

[Takashi Iwai]

At Wed, 25 Feb 2004 13:02:15 GMT,
Will wrote:
> 
> >> How easy would it be to make aplaymidi be able to move forwards and backwards
> >> during playback, e.g. by pressing f or b, like Takashi Iwai's drvmidi?
> >
> >Not so easy; even more so if you want to have correct controller
> >values.  Adjusting settings changed by sysexs would be almost
> >impossible.
> >
> 
> I wonder whether drvmidi actually does correctly restore controller and sysex
> when you move during MIDI playback.  I haven't studied the drvmidi code in
> detail.  I know the sound quality always seems to be correct
> even when you move forwards and backwards very often in a piece with
> lots of sysex and controller settings.

it does.  when skipping forward/backward, it will send all necessary
ctl/sysex messages from the beginning.  i implemented in this way
because it's designed only for the internal MIDI wavetable, which can
handle many events immediately.  that's not for general MIDI devices.


Takashi


-------------------------------------------------------
SF.Net is sponsored by: Speed Start Your Linux Apps Now.
Build and deploy apps & Web services for Linux with
a free DVD software kit from IBM. Click Now!
http://ads.osdn.com/?ad_id=1356&alloc_id=3438&op=click

Re: [PATCH] add aplaymidi & arecordmidi utils

[Will]

James Courtier-Dutton <James@superbug.demon.co.uk> wrote:
>Will wrote:
>> Is the following a correct way of updating the ALSA in Linux 2.6.3?
>
>I have just updated the alsa.opensrc.org wiki.
>Use option 1] on
>http://alsa.opensrc.org/index.php?page=AlsaBuild2.6

Thanks.  The section describing option 1 is still a bit confusing because
it seems to contradict itself firstly by saying alsa-driver is needed (at
the first paragraph) and secondly by saying it might not be needed (at
the end of the section).  It also doesn't say you should not install
alsa-kernel which presumably you shouldn't do.

Quick summary
1. Install alsa-driver from ALSA CVS
2. Disable ALSA in kernel (#CONFIG_SND is not set)
3. Enable sound in kernel (CONFIG_SOUND=y)
4. make and install from kernel source
5. configure/install alsa-lib alsa-firmware alsa-oss alsa-tools alsa-utils



-------------------------------------------------------
SF.Net is sponsored by: Speed Start Your Linux Apps Now.
Build and deploy apps & Web services for Linux with
a free DVD software kit from IBM. Click Now!
http://ads.osdn.com/?ad_id=1356&alloc_id=3438&op=click

Re: [PATCH] add aplaymidi & arecordmidi utils

[Will]

Takashi Iwai wrote:
>>Will wrote:
>> I wonder whether drvmidi actually does correctly restore controller and sysex
>> when you move during MIDI playback.  I haven't studied the drvmidi code in
>> detail.  I know the sound quality always seems to be correct
>> even when you move forwards and backwards very often in a piece with
>> lots of sysex and controller settings.
>
>it does.  when skipping forward/backward, it will send all necessary
>ctl/sysex messages from the beginning.  i implemented in this way
>because it's designed only for the internal MIDI wavetable, which can
>handle many events immediately.  that's not for general MIDI devices.

That's great.  How easy would it be to port drvmidi to the ALSA sequencer?
It's a fantastic little player.  It has the very useful functionality of
a GUI player (moving forwards and backwards during playback).
The only problem is it's for the OSS sequencer...

William



-------------------------------------------------------
SF.Net is sponsored by: Speed Start Your Linux Apps Now.
Build and deploy apps & Web services for Linux with
a free DVD software kit from IBM. Click Now!
http://ads.osdn.com/?ad_id=1356&alloc_id=3438&op=click

Re: [PATCH] add aplaymidi & arecordmidi utils

[Takashi Iwai]

At Wed, 25 Feb 2004 14:19:36 GMT,
Will wrote:
> 
> Takashi Iwai wrote:
> >>Will wrote:
> >> I wonder whether drvmidi actually does correctly restore controller and sysex
> >> when you move during MIDI playback.  I haven't studied the drvmidi code in
> >> detail.  I know the sound quality always seems to be correct
> >> even when you move forwards and backwards very often in a piece with
> >> lots of sysex and controller settings.
> >
> >it does.  when skipping forward/backward, it will send all necessary
> >ctl/sysex messages from the beginning.  i implemented in this way
> >because it's designed only for the internal MIDI wavetable, which can
> >handle many events immediately.  that's not for general MIDI devices.
> 
> That's great.  How easy would it be to port drvmidi to the ALSA sequencer?

a bit touch.  it's a quite old code.

> It's a fantastic little player.  It has the very useful functionality of
> a GUI player (moving forwards and backwards during playback).
> The only problem is it's for the OSS sequencer...

sorry, i have no time for that...
patches are always welcome, though :)

--
Takashi Iwai <tiwai@suse.de>		ALSA Developer - www.alsa-project.org


-------------------------------------------------------
SF.Net is sponsored by: Speed Start Your Linux Apps Now.
Build and deploy apps & Web services for Linux with
a free DVD software kit from IBM. Click Now!
http://ads.osdn.com/?ad_id=1356&alloc_id=3438&op=click

CVS ALSA failed upgrade in Linux 2.6.3

[William]

Jaroslav wrote on alsa-project.org:

   "simply copy files from the ALSA's alsa-kernel CVS module to relevant
   locations in the 2.6 kernel tree."

This is one method of upgrading to CVS ALSA.
However, using this method with Linux 2.6.3 and current CVS alsa-kernel
gives errors:

 $ cd linux-2.6.3

Optionally remove the old sound: [ $ mv sound sound-default-2.6.3 ]

 $ date
Mon 01 March 2004
 $ cvs -d:pserver:anonymous@cvs.sourceforge.net:/cvsroot/alsa login
[...]
 $ cvs -z3 -d:pserver:anonymous@cvs.sourceforge.net:/cvsroot/alsa co alsa-kernel
[...]
 $ cp -fR alsa-kernel/* sound

 $ make oldconfig
 $ make xconfig
make[1]: `scripts/fixdep' is up to date.
scripts/kconfig/qconf arch/i386/Kconfig
sound/Kconfig:35: can't open file "sound/oss/dmasound/Kconfig"
make[1]: *** [xconfig] Error 1
make: *** [xconfig] Error 2

 $ make clean
  CLEAN   arch/i386/boot/compressed
  CLEAN   arch/i386/boot
  CLEAN   arch/i386/kernel
  CLEAN   drivers/char
  CLEAN   drivers/ieee1394
  CLEAN   drivers/pci
  CLEAN   drivers/scsi/aic7xxx
  CLEAN   init
  CLEAN   lib
sound/oss/Makefile:99: linux-2.6.3/Rules.make: No such file or directory
make[2]: *** No rule to make target `linux-2.6.3/Rules.make'.  Stop.
make[1]: *** [sound/oss] Error 2
make: *** [_clean_sound] Error 2

This doesn't work.
So, after recursively copying the contents of the old sound-default-2.6.3/oss
into the new ALSA CVS directory:

 $ cp -iR sound-default-2.6.3/oss sound
cp: overwrite `sound/oss/Makefile'? y

 $ make clean
  CLEAN   usr
  CLEAN   scripts/genksyms
  CLEAN   scripts/kconfig
  CLEAN   scripts
  RM  $(CLEAN_FILES)

 $ make oldconfig
  HOSTCC  scripts/fixdep
  SHIPPED scripts/kconfig/zconf.tab.h
  HOSTCC  scripts/kconfig/conf.o
  HOSTCC  scripts/kconfig/mconf.o
  SHIPPED scripts/kconfig/zconf.tab.c
  SHIPPED scripts/kconfig/lex.zconf.c
  HOSTCC  -fPIC scripts/kconfig/zconf.tab.o
  HOSTLLD -shared scripts/kconfig/libkconfig.so
  HOSTLD  scripts/kconfig/conf
scripts/kconfig/conf -o arch/i386/Kconfig
[...]
  Digigram miXart (SND_MIXART) [N/m/?] (NEW)

 $ make
  CC [M]  sound/core/pcm.o
sound/core/pcm.c: In function `snd_pcm_substream_proc_hw_params_read':
sound/core/pcm.c:332: structure has no member named `period_frames'
sound/core/pcm.c: At top level:
sound/core/pcm.c:1064: `snd_pcm_limit_hw_rates' undeclared here (not in a function)
sound/core/pcm.c:1064: initializer element is not constant
sound/core/pcm.c:1064: (near initialization for `__ksymtab_snd_pcm_limit_hw_rates.value')
make[2]: *** [sound/core/pcm.o] Error 1
make[1]: *** [sound/core] Error 2
make: *** [sound] Error 2




-------------------------------------------------------
SF.Net is sponsored by: Speed Start Your Linux Apps Now.
Build and deploy apps & Web services for Linux with
a free DVD software kit from IBM. Click Now!
http://ads.osdn.com/?ad_id=1356&alloc_id=3438&op=click

Re: CVS ALSA failed upgrade in Linux 2.6.3

[Takashi Iwai]

At Tue, 2 Mar 2004 15:20:28 GMT,
William wrote:
> 
> Jaroslav wrote on alsa-project.org:
> 
>    "simply copy files from the ALSA's alsa-kernel CVS module to relevant
>    locations in the 2.6 kernel tree."
> 
> This is one method of upgrading to CVS ALSA.

"simply copy" doesn't mean so simply copy :)
for example, alsa-kernel/include directory goes to
linux/include/sound.  also, it might be safer to leave oss things.

if you're not sure, just build alsa-driver with 2.6 kernels as well as
you did for 2.4 kernels.  it should be ok now.


Takashi


-------------------------------------------------------
SF.Net is sponsored by: Speed Start Your Linux Apps Now.
Build and deploy apps & Web services for Linux with
a free DVD software kit from IBM. Click Now!
http://ads.osdn.com/?ad_id=1356&alloc_id=3438&op=click

Re: CVS ALSA failed upgrade in Linux 2.6.3

[William]

Takashi Iwai <tiwai@suse.de> wrote:
>>William wrote:
>> Jaroslav wrote on alsa-project.org:
>> 
>>    "simply copy files from the ALSA's alsa-kernel CVS module to relevant
>>    locations in the 2.6 kernel tree."
>> 
>> This is one method of upgrading to CVS ALSA.
>
>"simply copy" doesn't mean so simply copy :)

Thanks for explaining that.  Please would you delete the word "simply" in
the explanation on alsa-project.org because it's misleading.

>for example, alsa-kernel/include directory goes to
>linux/include/sound.

Ok, so is my summary below correct?

   $ cp -fR alsa-kernel/include/*  linux/include/sound
   $ mv -f alsa-kernel/include  out-of-the-way-temporarily
   $ cp -fR alsa-kernel/*  linux/sound

> also, it might be safer to leave oss things.

What do you mean "might be safer"?  That it's really necessary to leave oss?
Or that you're not certain?

>if you're not sure, just build alsa-driver with 2.6 kernels as well as
>you did for 2.4 kernels.  it should be ok now.

William


-------------------------------------------------------
SF.Net is sponsored by: Speed Start Your Linux Apps Now.
Build and deploy apps & Web services for Linux with
a free DVD software kit from IBM. Click Now!
http://ads.osdn.com/?ad_id=1356&alloc_id=3438&op=click

Re: CVS ALSA failed upgrade in Linux 2.6.3

[Takashi Iwai]

At Tue, 2 Mar 2004 15:52:24 GMT,
William wrote:
> 
> Takashi Iwai <tiwai@suse.de> wrote:
> >>William wrote:
> >> Jaroslav wrote on alsa-project.org:
> >> 
> >>    "simply copy files from the ALSA's alsa-kernel CVS module to relevant
> >>    locations in the 2.6 kernel tree."
> >> 
> >> This is one method of upgrading to CVS ALSA.
> >
> >"simply copy" doesn't mean so simply copy :)
> 
> Thanks for explaining that.  Please would you delete the word "simply" in
> the explanation on alsa-project.org because it's misleading.

maybe better to remove the whole sentense.

> >for example, alsa-kernel/include directory goes to
> >linux/include/sound.
> 
> Ok, so is my summary below correct?
> 
>    $ cp -fR alsa-kernel/include/*  linux/include/sound
>    $ mv -f alsa-kernel/include  out-of-the-way-temporarily
>    $ cp -fR alsa-kernel/*  linux/sound

the scripts directory shouldn't be copied.
also, Documentation directory goes to linux/sound/Documentation/sound/alsa

> > also, it might be safer to leave oss things.
> 
> What do you mean "might be safer"?  That it's really necessary to leave oss?
> Or that you're not certain?

don't copy the files under oss directory.


Takashi


-------------------------------------------------------
SF.Net is sponsored by: Speed Start Your Linux Apps Now.
Build and deploy apps & Web services for Linux with
a free DVD software kit from IBM. Click Now!
http://ads.osdn.com/?ad_id=1356&alloc_id=3438&op=click

Re: CVS ALSA failed upgrade in Linux 2.6.3

[William]

Takashi Iwai <tiwai@suse.de> wrote:
>> Thanks for explaining that.  Please would you delete the word "simply" in
>> the explanation on alsa-project.org because it's misleading.
>
>maybe better to remove the whole sentense.

I think it would be useful to keep an explanation of how to install
CVS alsa-kernel into a Linux 2.6.x source tree.

BTW, why is there still an alsa-driver package?
Is there any difference between (a) installing alsa-kernel into Linux 2.6.x
and (b) installing alsa-driver in the normal way with the configure option
--with-kernel=somewhere/linux

>the scripts directory shouldn't be copied.
>also, Documentation directory goes to linux/sound/Documentation/sound/alsa

Don't you mean "linux/sound/Documentation"?

>don't copy the files under oss directory.

So, perhaps below is a correct summary?

    cp -fR alsa-kernel/include/*  linux/include/sound
    mv -f alsa-kernel/include  out-of-the-way-temporarily-include

    mkdir -p linux/sound/Documentation/sound/alsa
    cp -fR alsa-kernel/Documentation/*  linux/sound/Documentation/sound/alsa
    mv -f alsa-kernel/Documentation out-of-the-way-temporarily-Documentation

    mv -f alsa-kernel/oss out-of-the-way-temporarily-oss

    mv -f alsa-kernel/scripts out-of-the-way-temporarily-scripts

    cp -fR alsa-kernel/*  linux/sound

Thanks,

William


-------------------------------------------------------
SF.Net is sponsored by: Speed Start Your Linux Apps Now.
Build and deploy apps & Web services for Linux with
a free DVD software kit from IBM. Click Now!
http://ads.osdn.com/?ad_id=1356&alloc_id=3438&op=click

Re: CVS ALSA failed upgrade in Linux 2.6.3

[Takashi Iwai]

At Tue, 2 Mar 2004 17:28:18 GMT,
William wrote:
> 
> Takashi Iwai <tiwai@suse.de> wrote:
> >> Thanks for explaining that.  Please would you delete the word "simply" in
> >> the explanation on alsa-project.org because it's misleading.
> >
> >maybe better to remove the whole sentense.
> 
> I think it would be useful to keep an explanation of how to install
> CVS alsa-kernel into a Linux 2.6.x source tree.
>  
> BTW, why is there still an alsa-driver package?
> Is there any difference between (a) installing alsa-kernel into Linux 2.6.x
> and (b) installing alsa-driver in the normal way with the configure option
> --with-kernel=somewhere/linux

yes, alsa-kernel is just for update of 2.6.x tree, and it doesn't
include the experimental new drivers included in alsa-driver tree.
alsa-driver tree also includes bunch of wrappers for 2.2/2.4 kernels.

that's why i wrote above to remove the sentense.
if you copy the files of alsa-kernel, you'll have no such new
drivers.


> >the scripts directory shouldn't be copied.
> >also, Documentation directory goes to linux/sound/Documentation/sound/alsa
> 
> Don't you mean "linux/sound/Documentation"?

sorry, linux/Documentation/sound/alsa.


Takashi


-------------------------------------------------------
SF.Net is sponsored by: Speed Start Your Linux Apps Now.
Build and deploy apps & Web services for Linux with
a free DVD software kit from IBM. Click Now!
http://ads.osdn.com/?ad_id=1356&alloc_id=3438&op=click

Re: CVS ALSA failed upgrade in Linux 2.6.3

[William]

Takashi Iwai <tiwai@suse.de> wrote:
>>William wrote:
>> I think it would be useful to keep an explanation of how to install
>> CVS alsa-kernel into a Linux 2.6.x source tree.
>>  
>> BTW, why is there still an alsa-driver package?
>> Is there any difference between (a) installing alsa-kernel into Linux 2.6.x
>> and (b) installing alsa-driver in the normal way with the configure option
>> --with-kernel=somewhere/linux
>
>yes, alsa-kernel is just for update of 2.6.x tree, and it doesn't
>include the experimental new drivers included in alsa-driver tree.
>alsa-driver tree also includes bunch of wrappers for 2.2/2.4 kernels.

This is very important information. How about explaining this on the
alsa-project.org website where the current sentence reads "simply copy ..." ?

>that's why i wrote above to remove the sentense.
>if you copy the files of alsa-kernel, you'll have no such new drivers.

I probably want the new/experimental drivers.
  cd alsa-driver-cvs-20040302
  ./configure --with-cards=.... --with-kernel=somewhere/linux-2.6.x
Does this mean the both ALSA and OSS sound modules in the Linux 2.6.x source
tree are automatically replaced and updated by the ALSA and OSS modules
from alsa-driver?  I want to have the old OSS sound drivers so
I can change from using ALSA to OSS (and vice-versa) by modules unloading
and re-loading.

>> Don't you mean "linux/sound/Documentation"?
>sorry, linux/Documentation/sound/alsa.

Ok, so is this a correct summary of installing alsa-kernel into Linux 2.6.x?

#!/bin/sh
# Install CVS alsa-kernel into Linux 2.6.x source tree:
    cp -fR alsa-kernel/include/*  linux/include/sound
    mv -f alsa-kernel/include  out-of-the-way-temporarily-include

    cp -fR alsa-kernel/Documentation/*  linux/Documentation/sound/alsa
    mv -f alsa-kernel/Documentation out-of-the-way-temporarily-Documentation

    mv -f alsa-kernel/oss out-of-the-way-temporarily-oss

    mv -f alsa-kernel/scripts out-of-the-way-temporarily-scripts

    cp -fR alsa-kernel/*  linux/sound

Thanks,

William


-------------------------------------------------------
SF.Net is sponsored by: Speed Start Your Linux Apps Now.
Build and deploy apps & Web services for Linux with
a free DVD software kit from IBM. Click Now!
http://ads.osdn.com/?ad_id=1356&alloc_id=3438&op=click

Re: CVS ALSA failed upgrade in Linux 2.6.3

[Takashi Iwai]

At Tue, 2 Mar 2004 17:59:55 GMT,
William wrote:
> 
> Takashi Iwai <tiwai@suse.de> wrote:
> >>William wrote:
> >> I think it would be useful to keep an explanation of how to install
> >> CVS alsa-kernel into a Linux 2.6.x source tree.
> >>  
> >> BTW, why is there still an alsa-driver package?
> >> Is there any difference between (a) installing alsa-kernel into Linux 2.6.x
> >> and (b) installing alsa-driver in the normal way with the configure option
> >> --with-kernel=somewhere/linux
> >
> >yes, alsa-kernel is just for update of 2.6.x tree, and it doesn't
> >include the experimental new drivers included in alsa-driver tree.
> >alsa-driver tree also includes bunch of wrappers for 2.2/2.4 kernels.
> 
> This is very important information. How about explaining this on the
> alsa-project.org website where the current sentence reads "simply copy ..." ?
 
maybe.  the statement is a bit old.

> >that's why i wrote above to remove the sentense.
> >if you copy the files of alsa-kernel, you'll have no such new drivers.
> 
> I probably want the new/experimental drivers.
>   cd alsa-driver-cvs-20040302
>   ./configure --with-cards=.... --with-kernel=somewhere/linux-2.6.x
> Does this mean the both ALSA and OSS sound modules in the Linux 2.6.x source
> tree are automatically replaced and updated by the ALSA and OSS modules
> from alsa-driver?  I want to have the old OSS sound drivers so
> I can change from using ALSA to OSS (and vice-versa) by modules unloading
> and re-loading.

OSS drivers won't be replaced.  they will be left as they were.

> 
> >> Don't you mean "linux/sound/Documentation"?
> >sorry, linux/Documentation/sound/alsa.
> 
> Ok, so is this a correct summary of installing alsa-kernel into Linux 2.6.x?
> 
> #!/bin/sh
> # Install CVS alsa-kernel into Linux 2.6.x source tree:
>     cp -fR alsa-kernel/include/*  linux/include/sound
>     mv -f alsa-kernel/include  out-of-the-way-temporarily-include
> 
>     cp -fR alsa-kernel/Documentation/*  linux/Documentation/sound/alsa
>     mv -f alsa-kernel/Documentation out-of-the-way-temporarily-Documentation
> 
>     mv -f alsa-kernel/oss out-of-the-way-temporarily-oss
> 
>     mv -f alsa-kernel/scripts out-of-the-way-temporarily-scripts
> 
>     cp -fR alsa-kernel/*  linux/sound

does recursive copy work in this way?
also omit CVS and kbuild directories if you have.

another easy way to update is to use the bk tree patch...


Takashi


-------------------------------------------------------
SF.Net is sponsored by: Speed Start Your Linux Apps Now.
Build and deploy apps & Web services for Linux with
a free DVD software kit from IBM. Click Now!
http://ads.osdn.com/?ad_id=1356&alloc_id=3438&op=click

Re: CVS ALSA failed upgrade in Linux 2.6.3

[James Courtier-Dutton]

William wrote:
> Jaroslav wrote on alsa-project.org:
> 
>    "simply copy files from the ALSA's alsa-kernel CVS module to relevant
>    locations in the 2.6 kernel tree."
> 
> This is one method of upgrading to CVS ALSA.
> However, using this method with Linux 2.6.3 and current CVS alsa-kernel
> gives errors:
> 

Use directions at: -
http://alsa.opensrc.org/index.php?page=AlsaBuild2.6

Cheers
James


-------------------------------------------------------
SF.Net is sponsored by: Speed Start Your Linux Apps Now.
Build and deploy apps & Web services for Linux with
a free DVD software kit from IBM. Click Now!
http://ads.osdn.com/?ad_id=1356&alloc_id=3438&op=click

Re: CVS ALSA failed upgrade in Linux 2.6.3

[William]

James Courtier-Dutton <James@superbug.demon.co.uk> wrote:
>
>Use directions at: -
>http://alsa.opensrc.org/index.php?page=AlsaBuild2.6
>

Thanks, James.  Like I said in my previous reply to you, the section
describing option 1 is confusing because it seems to contradict itself
firstly by saying alsa-driver is needed (at the first paragraph) and
secondly by saying alsa-driver might not be needed (at the end of the section)
because Linux 2.6.3 includes alsa-driver-1.0.2,
which misses the point of wanting to use CVS alsa-driver in the first place,
i.e. to get the very latest drivers -- not those in alsa-driver-1.0.2.
Basically it doesn't explain why you should install CVS alsa-drivers.
Also it doesn't say what to do with alsa-kernel yet the cvs example shows
alsa-kernel being checked out.  As Takashi Iwai said earlier, alsa-kernel
includes stable drivers only and excludes new/experimental drivers.
My understanding is you should install either alsa-kernel or CVS alsa-drivers,
but not both of them.

William


-------------------------------------------------------
SF.Net is sponsored by: Speed Start Your Linux Apps Now.
Build and deploy apps & Web services for Linux with
a free DVD software kit from IBM. Click Now!
http://ads.osdn.com/?ad_id=1356&alloc_id=3438&op=click

Re: CVS ALSA failed upgrade in Linux 2.6.3

[James Courtier-Dutton]

William wrote:
> James Courtier-Dutton <James@superbug.demon.co.uk> wrote:
> 
>>Use directions at: -
>>http://alsa.opensrc.org/index.php?page=AlsaBuild2.6
>>
> 
> 
> Thanks, James.  Like I said in my previous reply to you, the section
> describing option 1 is confusing because it seems to contradict itself
> firstly by saying alsa-driver is needed (at the first paragraph) and
> secondly by saying alsa-driver might not be needed (at the end of the section)
> because Linux 2.6.3 includes alsa-driver-1.0.2,
> which misses the point of wanting to use CVS alsa-driver in the first place,
> i.e. to get the very latest drivers -- not those in alsa-driver-1.0.2.
> Basically it doesn't explain why you should install CVS alsa-drivers.
> Also it doesn't say what to do with alsa-kernel yet the cvs example shows
> alsa-kernel being checked out.  As Takashi Iwai said earlier, alsa-kernel
> includes stable drivers only and excludes new/experimental drivers.
> My understanding is you should install either alsa-kernel or CVS alsa-drivers,
> but not both of them.
> 
> William
> 
> 
The point of option 1 is to explain how to upgrade alsa if you need to. 
Not everyone needs to upgrade.
If you wish to use the cvs, you have to get alsa-driver and alsa-kernel, 
because alsa-driver uses files from ../alsa-kernel
So, if you follow the cvs command exactly it will work.

Cheers
James



-------------------------------------------------------
SF.Net is sponsored by: Speed Start Your Linux Apps Now.
Build and deploy apps & Web services for Linux with
a free DVD software kit from IBM. Click Now!
http://ads.osdn.com/?ad_id=1356&alloc_id=3438&op=click

Re: CVS ALSA failed upgrade in Linux 2.6.3

[William]

Takashi Iwai <tiwai@suse.de> wrote:
>William wrote:
>> This is very important information. How about explaining this on the
>> alsa-project.org website where the current sentence reads "simply copy ..." ?
> 
>maybe.  the statement is a bit old.

And a bit confusing too...

>> I probably want the new/experimental drivers.
>>   cd alsa-driver-cvs-20040302
>>   ./configure --with-cards=.... --with-kernel=somewhere/linux-2.6.x
>> Does this mean the both ALSA and OSS sound modules in the Linux 2.6.x source
>> tree are automatically replaced and updated by the ALSA and OSS modules
>> from alsa-driver?  I want to have the old OSS sound drivers so
>> I can change from using ALSA to OSS (and vice-versa) by modules unloading
>> and re-loading.
>
>OSS drivers won't be replaced.  they will be left as they were.

And, presumably, the in-kernel ALSA sound modules are also not replaced when
you install CVS alsa-driver?  Why does CVS alsa-driver require CVS alsa-kernel
for the compilation even with the config option --with-kernel=/somewhere/linux?

>>     cp -fR alsa-kernel/*  linux/sound
>
>does recursive copy work in this way?
>also omit CVS and kbuild directories if you have.

Yes, if I understand your question, "cp -R" will copy all of the subdirectories
recursively.  It's probably better to use "-p" too to preserve modes, etc.

>another easy way to update is to use the bk tree patch...

Does the bk tree patch have older code than the very new drivers in
CVS alsa-driver?

William

Finally, a correct summary of installing CVS alsa-kernel into Linux 2.6.x:

#!/bin/sh
# Install CVS alsa-kernel into Linux 2.6.x source tree:

read linux
cd $linux

cvs -d:pserver:anonymous@cvs.sourceforge.net:/cvsroot/alsa login
cvs -z3 -d:pserver:anonymous@cvs.sourceforge.net:/cvsroot/alsa co alsa-kernel alsa-driver

echo -n "Enter kernel source directory: "
cp -fpR alsa-kernel/include/*  $linux/include/sound
mv -f alsa-kernel/include  out-of-the-way-temporarily-include   

cp -fpR alsa-kernel/Documentation/*  $linux/Documentation/sound/alsa
mv -f alsa-kernel/Documentation out-of-the-way-temporarily-Documentation

mv -f alsa-kernel/oss out-of-the-way-temporarily-oss

mv -f alsa-kernel/scripts out-of-the-way-temporarily-scripts
mv -f alsa-kernel/CVS out-of-the-way-temporarily-CVS
mv -f alsa-kernel/kbuild out-of-the-way-temporarily-kbuild

cp -fpR alsa-kernel/*  $linux/sound



-------------------------------------------------------
SF.Net is sponsored by: Speed Start Your Linux Apps Now.
Build and deploy apps & Web services for Linux with
a free DVD software kit from IBM. Click Now!
http://ads.osdn.com/?ad_id=1356&alloc_id=3438&op=click

Re: CVS ALSA failed upgrade in Linux 2.6.3

[William]

James Courtier-Dutton <James@superbug.demon.co.uk> wrote:
>
>The point of option 1 is to explain how to upgrade alsa if you need to. 
>Not everyone needs to upgrade.

My point was that the explanation given in option is actually confusing.
I guess it's the wiki problem of many different people collaboratively
editing a wiki page, leading to incoherency.  That wiki page would probably
benefit from having just one person rewrite it from scratch.

>If you wish to use the cvs, you have to get alsa-driver and alsa-kernel, 
>because alsa-driver uses files from ../alsa-kernel

Yes, I knew that already but what I didn't know was that it is best
to install alsa-driver not alsa-kernel to get all the new/experimental drivers.

>So, if you follow the cvs command exactly it will work.

It didn't work (Unknown symbols in snd*) but at least I know why (Takashi said
the CVS alsa-driver is currently in the middle of a major rewrite).
Thanks anyway.

William


-------------------------------------------------------
SF.Net is sponsored by: Speed Start Your Linux Apps Now.
Build and deploy apps & Web services for Linux with
a free DVD software kit from IBM. Click Now!
http://ads.osdn.com/?ad_id=1356&alloc_id=3438&op=click