doc: Move filter examples to man page.
[paraslash.git] / web / manual.md
1 **Paraslash user manual**
2
3 This document describes how to install, configure and use the paraslash
4 network audio streaming system. Most chapters start with a chapter
5 overview and conclude with an example section. We try to focus on
6 general concepts and on the interaction of the various pieces of the
7 paraslash package. Hence this user manual is not meant as a replacement
8 for the manual pages that describe all command line options of each
9 paraslash executable.
10
11 ============
12 Introduction
13 ============
14
15 In this chapter we give an [overview](#Overview) of the interactions
16 of the programs contained in the paraslash package, followed by
17 [brief descriptions](#The.paraslash.executables) of all executables.
18
19 Overview
20 --------
21
22 The core functionality of the para suite is provided by two main
23 applications, para_server and para_audiod. para_server maintains
24 the audio file database and acts as the streaming source, while
25 para_audiod is the streaming client. Usually, both run in the
26 background on different hosts but a local setup is also possible.
27
28 A simplified picture of a typical setup is as follows
29
30
31 .____________________.
32 | ______ |
33 .-----------------------. | .d########b. |
34 |.---------------------.| | .d############b |
35 || || | .d######""####//b. |
36 || || | 9######( )######P |
37 || || | 'b######++######d' |
38 || Screen || | "9############P" |
39 || || | "9a########P" |
40 || || | `""""'' |
41 |`---------------------'| | ________________ |
42 `-----------------------' | |________________| |
43 ___) (___ |____________________|
44 `-._______.-' loudspeaker
45 | |
46 | |
47 | |
48 .____/ \___. ._____________. ._____/ \_____.
49 | | | | | |
50 | para_gui |-----| para_audioc |-----| para_audiod |
51 |____ ___| |_____________| |_____ _____|
52 \ / \ /
53 | |
54 | |
55 | |
56 ._____/ \_____. ._____/ \_____.
57 | | | |
58 | para_client |-----------------------| para_server |
59 |_____________| |_____ _____|
60 \ /
61 |
62 |
63 .-'"""`-.
64 ( )
65 |`-.___.-'|
66 | |
67 |. ' " ` .|
68 | |
69 `-.___.-'
70 Database
71
72 The two client programs, para_client and para_audioc communicate with
73 para_server and para_audiod, respectively.
74
75 para_gui controls para_server and para_audiod by executing para_client
76 and para_audioc. In particular, it runs a command to obtain the state
77 of para_audiod and para_server, and the metadata of the current audio
78 file. This information is pretty-printed in a curses window.
79
80 The paraslash executables
81 -------------------------
82
83 <h3> para_server </h3>
84
85 para_server streams binary audio data (MP3, ...) over local and/or
86 remote networks. It listens on a TCP port and accepts commands such
87 as play, stop, pause, next from authenticated clients. The components
88 of para_server are illustrated in the following diagram:
89
90 ______________________________________________________________________ network
91 | | | | |
92 | .-'""""`-. | | | |
93 | ( ) | | | |
94 .____/ \_____. |`-.____.-'| .____/ \____/ \____/ \____. |
95 | | | | | | |
96 | dispatcher | | database | | senders (http/udp/dccp) | |
97 |____ _____| | | |___________ ___________| |
98 \ / |. ' "" ` .| \ / |
99 | | | | |
100 | `-.____.-' | |
101 | | | |
102 | | | |
103 | | | |
104 | ._____/ \_____. .________/ \________. |
105 | | | | | |
106 | | audio file |________| virtual streaming | |
107 | | selector | | system | |
108 | |_____ _____| |________ ________| |
109 | \ / \ / |
110 | | | |
111 | | | |
112 | | ._________________. | |
113 | | | | | |
114 | `---| command handler |---' |
115 | |____ ___ ____| |
116 | \ / \ / |
117 | | | |
118 | | | |
119 | | | |
120 `-------------------------' `--------------------------'
121
122
123 Incoming connections arrive at the dispatcher which creates a process
124 dedicated to the connection. Its task is to authenticate the client
125 and to run the command handler which forwards the client request to
126 either the audio file selector or the virtual streaming system. Results
127 (if any) are sent back to the client.
128
129 The audio file selector manages audio files using various database
130 tables. It maintains statistics on the usage of all audio files such as
131 last-played time and the number of times each file was selected. It
132 is also responsible for selecting and loading audio files for
133 streaming. Additional information may be added to the database to allow
134 fine-grained selection based on various properties of the audio file,
135 including information found in (ID3) tags. Simple playlists are also
136 supported. It is possible to store images (album covers) and lyrics
137 in the database and associate these to the corresponding audio files.
138 The section on the [audio file selector](#The.audio.file.selector)
139 discusses this topic in more detail.
140
141 Another component of para_server is the virtual streaming system,
142 which controls the paraslash senders. During streaming it requests
143 small chunks of data (e.g., mp3 frames) from the audio file selector
144 and feeds them to the senders which forward the chunks to connected
145 clients.
146
147 The three senders of para_server correspond to network streaming
148 protocols based on HTTP, DCCP, or UDP. This is explained in the
149 section on [networking](#Networking).
150
151 <h3> para_client </h3>
152
153 The client program to connect to para_server. paraslash commands
154 are sent to para_server and the response is dumped to STDOUT. This
155 can be used by any scripting language to produce user interfaces with
156 little programming effort.
157
158 All connections between para_server and para_client are encrypted
159 with a symmetric session key. For each user of paraslash you must
160 create a public/secret RSA key pair for authentication.
161
162 If para_client is started without non-option arguments, an interactive
163 session (shell) is started. Command history and command completion are
164 supported through libreadline.
165
166 <h3> para_audiod </h3>
167
168 The purpose of para_audiod is to download, decode and play an audio
169 stream received from para_server. A typical setup looks as follows.
170
171
172 .----------------------------.
173 | |
174 | |
175 ._____/ \_____. .___/ \____.
176 | | .----------| |
177 | para_server | | .______| receiver |
178 |_____ ____| | | |___ ____|
179 \ / | | \ /
180 | | | |
181 | | | |
182 | | | |
183 ._____/ \_____. | | .___/ \____.
184 | | | | | |
185 | status task |-----+ | | filter 1 |
186 |_____________| | |___ ____|
187 | \ /
188 | | .____________________.
189 | | | ______ |
190 .____________. | .___/ \____. | .d########b. |
191 | | | | | | .d############b |
192 | dispatcher |----------' | filter 2 | | .d######""####//b. |
193 |_____ ____| |___ ____| | 9######( )######P |
194 \ / \ / | 'b######++######d' |
195 | | | "9############P" |
196 | | | "9a########P" |
197 ._____/ \_____. .___/ \____. | `""""'' |
198 | | | | | ________________ |
199 | para_audioc | | writer |------| |________________| |
200 |_____________| |__________| |____________________|
201
202
203 The status task of para_audiod connects to para_server and runs the
204 "stat" command to retrieve the current server status. If an audio
205 stream is available, para_audiod starts a so-called buffer tree to
206 play the stream.
207
208 The buffer tree consists of a receiver, any number of filters and a
209 writer. The receiver downloads the audio stream from para_server and
210 the filters decode or modify the received data. The writer plays the
211 decoded stream.
212
213 The dispatcher of para_audiod listens on a local socket and runs
214 audiod commands on behalf of para_audioc. For example, para_gui runs
215 para_audioc to obtain status information about para_audiod and the
216 current audio file. Access to the local socket may be restricted by
217 means of Unix socket credentials.
218
219 <h3> para_audioc </h3>
220
221 The client program which talks to para_audiod. Used to control
222 para_audiod, to receive status info, or to grab the stream at any
223 point of the decoding process. Like para_client, para_audioc supports
224 interactive sessions on systems with libreadline.
225
226 <h3> para_recv </h3>
227
228 A command line HTTP/DCCP/UDP stream grabber. The http mode is
229 compatible with arbitrary HTTP streaming sources (e.g. icecast).
230 In addition to the three network streaming modes, para_recv can also
231 operate in local (afh) mode. In this mode it writes the content of
232 an audio file on the local file system in complete chunks to stdout,
233 optionally 'just in time'. This allows to cut an audio file without
234 first decoding it, and it enables third-party software which is unaware
235 of the particular audio format to send complete frames in real time.
236
237 <h3> para_filter </h3>
238
239 A filter program that reads from STDIN and writes to STDOUT.
240 Like para_recv, this is an atomic building block which can be used to
241 assemble higher-level audio receiving facilities. It combines several
242 different functionalities in one tool: decoders for multiple audio
243 formats and a number of processing filters, among these a normalizer
244 for audio volume.
245
246 <h3> para_afh </h3>
247
248 A small stand-alone program that prints tech info about the given
249 audio file to STDOUT. It can be instructed to print a "chunk table",
250 an array of offsets within the audio file.
251
252 <h3> para_write </h3>
253
254 A modular audio stream writer. It supports a simple file writer
255 output plug-in and optional WAV/raw players for ALSA (Linux) and OSS.
256 para_write can also be used as a stand-alone WAV or raw audio player.
257
258 <h3> para_play </h3>
259
260 A command line audio player which supports the same audio formats as
261 para_server. It differs from other players in that it has an insert
262 and a command mode, like the vi editor. Line editing is based on
263 libreadline, and tab completion and command history are supported.
264
265 <h3> para_gui </h3>
266
267 Curses-based gui that presents status information obtained in a curses
268 window. Appearance can be customized via themes. para_gui provides
269 key-bindings for the most common server commands and new key-bindings
270 can be added easily.
271
272 <h3> para_mixer </h3>
273
274 An alarm clock and volume-fader for OSS and ALSA.
275
276 ===========
277 Quick start
278 ===========
279
280 This chapter lists the [necessary software](#Requirements)
281 that must be installed to compile the paraslash package, describes
282 how to [compile and install](#Installation) the paraslash
283 source code and the steps that have to be performed in order to
284 [set up](#Configuration) a typical server and client.
285
286 Requirements
287 ------------
288 <h3> For the impatient </h3>
289
290 git clone git://git.tuebingen.mpg.de/lopsub
291 cd lopsub && make && sudo make install
292 git clone git://git.tuebingen.mpg.de/osl
293 cd osl && make && sudo make install && sudo ldconfig
294 sudo apt-get install autoconf libssl-dev m4 \
295 libmad0-dev libid3tag0-dev libasound2-dev libvorbis-dev \
296 libfaad-dev libspeex-dev libFLAC-dev libsamplerate-dev realpath \
297 libasound2-dev libao-dev libreadline-dev libncurses-dev \
298 libopus-dev
299
300 <h3> Detailed description </h3>
301
302 In any case you will need
303
304 - [lopsub](http://people.tuebingen.mpg.de/maan/lopsub/). The long
305 option parser for subcommands generates the command line and config
306 file parsers for all paraslash executables. Clone the source code
307 repository with
308
309 git clone git://git.tuebingen.mpg.de/lopsub
310
311 - [gcc](ftp://ftp.gnu.org/pub/gnu/gcc) or
312 [clang](http://clang.llvm.org). All gcc versions >= 4.2 are currently
313 supported. Clang version 1.1 or newer should work as well.
314
315 - [gnu make](ftp://ftp.gnu.org/pub/gnu/make) is also shipped with the
316 disto. On BSD systems the gnu make executable is often called gmake.
317
318 - [bash](ftp://ftp.gnu.org/pub/gnu/bash). Some scripts which run
319 during compilation require the _Bourne again shell_. It is most
320 likely already installed.
321
322 - [m4](ftp://ftp.gnu.org/pub/gnu/m4/). Some source files are generated
323 from templates by the m4 macro processor.
324
325 Optional:
326
327 - [libosl](http://people.tuebingen.mpg.de/maan/osl/). The _object
328 storage layer_ library is used by para_server. To clone the source
329 code repository, execute
330
331 git clone git://git.tuebingen.mpg.de/osl
332
333 - [openssl](https://www.openssl.org/) or
334 [libgcrypt](ftp://ftp.gnupg.org/gcrypt/libgcrypt/). At least one
335 of these two libraries is needed as the backend for cryptographic
336 routines on both the server and the client side. Both openssl and
337 libgcrypt are usually shipped with the distro, but you might have
338 to install the development package (`libssl-dev` or `libgcrypt-dev`
339 on debian systems) as well.
340
341 - [flex](https://github.com/westes/flex) and
342 [bison](https://www.gnu.org/software/bison) are needed to build the
343 mood parser of para_server. The build system will skip para_server
344 if these tools are not installed.
345
346 - [libmad](http://www.underbit.com/products/mad/). To compile in MP3
347 support for paraslash, the development package must be installed. It
348 is called `libmad0-dev` on debian-based systems. Note that libmad is
349 not necessary on the server side, i.e., for sending MP3 files.
350
351 - [libid3tag](http://www.underbit.com/products/mad/). For version-2
352 ID3 tag support, you will need the libid3tag development package
353 `libid3tag0-dev`. Without libid3tag, only version-1 tags are
354 recognized. The mp3 tagger also needs this library for modifying
355 (id3v1 and id3v2) tags.
356
357 - [ogg vorbis](https://www.xiph.org/downloads/). For ogg vorbis streams
358 you need libogg, libvorbis, libvorbisfile. The corresponding Debian
359 packages are called `libogg-dev` and `libvorbis-dev`.
360
361 - [libfaad and mp4ff](http://www.audiocoding.com/). For aac files
362 (m4a) you need libfaad and libmp4ff (package: `libfaad-dev`). Note
363 that for some distributions, e.g. Ubuntu, mp4ff is not part of the
364 libfaad package. Install the faad library from sources (available
365 through the above link) to get the mp4ff library and header files.
366
367 - [speex](https://www.speex.org/). In order to stream or decode speex
368 files, libspeex (`libspeex-dev`) is required.
369
370 - [flac](https://xiph.org/flac/). To stream or decode files
371 encoded with the _Free Lossless Audio Codec_, libFLAC (`libFLAC-dev`)
372 must be installed.
373
374 - [libsamplerate](http://www.mega-nerd.com/SRC/index.html). The
375 resample filter will only be compiled if this library is
376 installed. Debian package: `libsamplerate-dev`.
377
378 - [alsa-lib](ftp://ftp.alsa-project.org/pub/lib/). On Linux, you will
379 need to have the ALSA development package `libasound2-dev` installed.
380
381 - [libao](https://ftp.osuosl.org/pub/xiph/releases/ao/). Needed to build
382 the ao writer (ESD, PulseAudio,...). Debian package: `libao-dev`.
383
384 - [curses](ftp://ftp.gnu.org/pub/gnu/ncurses). Needed for
385 para_gui. Debian package: `libncurses-dev`.
386
387 - [GNU
388 Readline](http://cnswww.cns.cwru.edu/php/chet/readline/rltop.html). If
389 this library (`libreadline-dev`) is installed, para_client, para_audioc
390 and para_play support interactive sessions.
391
392 Installation
393 ------------
394 To build the sources from a tarball, execute
395
396 ./configure && make
397
398 To build from git or a gitweb snapshot, run this command instead:
399
400 ./autogen.sh
401
402 There should be no errors but probably some warnings about missing
403 packages which usually implies that not all audio formats will be
404 supported. If headers or libs are installed at unusual locations you
405 might need to tell the configure script where to find them. Try
406
407 ./configure --help
408
409 to see a list of options. If the paraslash package was compiled
410 successfully, execute (optionally)
411
412 make test
413
414 to run the paraslash test suite. If all tests pass, execute as root
415
416 make install
417
418 to install executables under /usr/local/bin and the man pages under
419 /usr/local/man.
420
421 Configuration
422 -------------
423
424 <h3> Create a paraslash user </h3>
425
426 In order to control para_server at runtime you must create a paraslash
427 user. As authentication is based on the RSA crypto system you'll have
428 to create an RSA key pair. If you already have a user and an RSA key
429 pair, you may skip this step.
430
431 In this section we'll assume a typical setup: You would like to run
432 para_server on some host called server_host as user foo, and you want
433 to connect to para_server from another machine called client_host as
434 user bar.
435
436 As foo@server_host, create ~/.paraslash/server.users by typing the
437 following commands:
438
439 user=bar
440 target=~/.paraslash/server.users
441 key=~/.paraslash/id_rsa.pub.$user
442 perms=AFS_READ,AFS_WRITE,VSS_READ,VSS_WRITE
443 mkdir -p ~/.paraslash
444 echo "user $user $key $perms" >> $target
445
446 Next, change to the "bar" account on client_host and generate the
447 key pair with the commands
448
449 ssh-keygen -q -t rsa -b 2048 -N '' -f $key
450
451 This generates the two files id_rsa and id_rsa.pub in ~/.ssh. Note
452 that para_server won't accept keys shorter than 2048 bits. Moreover,
453 para_client rejects private keys which are world-readable.
454
455 para_server only needs to know the public key of the key pair just
456 created. Copy this public key to server_host:
457
458 src=~/.ssh/id_rsa.pub
459 dest=.paraslash/id_rsa.pub.$LOGNAME
460 scp $src foo@server_host:$dest
461
462 Finally, tell para_client to connect to server_host:
463
464 conf=~/.paraslash/client.conf
465 echo 'hostname server_host' > $conf
466
467
468 <h3> Start para_server </h3>
469
470 For this first try, we'll use the info loglevel to make the output
471 of para_server more verbose.
472
473 para_server -l info
474
475 Now you can use para_client to connect to the server and issue
476 commands. Open a new shell as bar@client_host and try
477
478 para_client help
479 para_client si
480
481 to retrieve the list of available commands and some server info.
482 Don't proceed if this doesn't work.
483
484 <h3> Create and populate the database </h3>
485
486 An empty database is created with
487
488 para_client init
489
490 This initializes a couple of empty tables under
491 ~/.paraslash/afs_database-0.4. You normally don't need to look at these
492 tables, but it's good to know that you can start from scratch with
493
494 rm -rf ~/.paraslash/afs_database-0.4
495
496 in case something went wrong.
497
498 Next, you need to add some audio files to that database so that
499 para_server knows about them. Choose an absolute path to a directory
500 containing some audio files and add them to the audio file table:
501
502 para_client add /my/mp3/dir
503
504 This might take a while, so it is a good idea to start with a directory
505 containing not too many files. Note that the table only contains data
506 about the audio files found, not the files themselves.
507
508 You may print the list of all known audio files with
509
510 para_client ls
511
512 <h3> Configure para_audiod </h3>
513
514 We will have to tell para_audiod that it should receive the audio
515 stream from server_host via http:
516
517 para_audiod -l info -r '.:http -i server_host'
518
519 You should now be able to listen to the audio stream once para_server
520 starts streaming. To activate streaming, execute
521
522 para_client play
523
524 Since no playlist has been specified yet, the "dummy" mode which
525 selects all known audio files is activated automatically. See the
526 section on the [audio file selector](#The.audio.file.selector) for how
527 to use playlists and moods to specify which files should be streamed
528 in which order.
529
530 Troubleshooting
531 ---------------
532
533 To identify streaming problems try to receive, decode and play the
534 stream manually using para_recv, para_filter and para_write as follows.
535 For simplicity we assume that you're running Linux/ALSA and that only
536 MP3 files have been added to the database.
537
538 para_recv -r 'http -i server_host' > file.mp3
539 # (interrupt with CTRL+C after a few seconds)
540 ls -l file.mp3 # should not be empty
541 para_filter -f mp3dec -f wav < file.mp3 > file.wav
542 ls -l file.wav # should be much bigger than file.mp3
543 para_write -w alsa < file.wav
544
545 Double check what is logged by para_server and use the --loglevel
546 option of para_recv, para_filter and para_write to increase verbosity.
547
548 ===============
549 User management
550 ===============
551
552 para_server uses a challenge-response mechanism to authenticate
553 requests from incoming connections, similar to ssh's public key
554 authentication method. Authenticated connections are encrypted using
555 the AES stream cipher in integer counter mode.
556
557 In this chapter we briefly describe RSA and AES, and sketch the
558 [authentication handshake](#Client-server.authentication)
559 between para_client and para_server. User management is discussed
560 in the section on [the user_list file](#The.user_list.file).
561 These sections are all about communication between the client and the
562 server. Connecting para_audiod is a different matter and is described
563 in a [separate section](#Connecting.para_audiod).
564
565 RSA and AES
566 -----------
567
568 A block cipher is a transformation which operates on fixed-length
569 blocks. For symmetric block ciphers the transformation is determined
570 by a single key for both encryption and decryption. For asymmetric
571 block ciphers, on the other hand, the key consists of two parts,
572 called the public key and the private key. A message can be encrypted
573 with either key and only the counterpart of that key can decrypt the
574 message. Asymmetric block ciphers can be used for both signing and
575 encrypting a message.
576
577 RSA is an asymmetric block cipher which is used in many applications,
578 including ssh and gpg. The RSA public key encryption and signatures
579 algorithms are defined in detail in RFC 2437. Paraslash relies on
580 RSA for authentication.
581
582 Stream ciphers XOR the input with a pseudo-random key stream to produce
583 the output. Decryption uses the same function calls as encryption.
584 Any block cipher can be turned into a stream cipher by generating the
585 pseudo-random key stream by encrypting successive values of a counter
586 (counter mode).
587
588 AES, the advanced encryption standard, is a well-known symmetric block
589 cipher. Paraslash employs AES in counter mode as described above to
590 encrypt communications. Since a stream cipher key must not be used
591 twice, a random key is generated for every new connection.
592
593 Client-server authentication
594 ----------------------------
595
596 The authentication handshake between para_client and para_server goes
597 as follows:
598
599 - para_client connects to para_server and sends an authentication
600 request for a user. It does so by connecting to TCP port 2990 of the
601 server host. This port is called the para_server _control port_.
602
603 - para_server accepts the connection and forks a child process which
604 handles the incoming request. The parent process keeps listening on the
605 control port while the child process (also called para_server below)
606 continues as follows.
607
608 - para_server loads the RSA public key of that user, fills a
609 fixed-length buffer with random bytes, encrypts that buffer using the
610 public key and sends the encrypted buffer to the client. The first
611 part of the buffer is the challenge which is used for authentication
612 while the second part is the session key.
613
614 - para_client receives the encrypted buffer and decrypts it with the
615 user's private key, thereby obtaining the challenge buffer and the
616 session key. It sends the SHA1 hash value of the challenge back to
617 para_server and stores the session key for further use.
618
619 - para_server also computes the SHA1 hash of the challenge and compares
620 it against what was sent back by the client.
621
622 - If the two hashes do not match, the authentication has failed and
623 para_server closes the connection.
624
625 - Otherwise the user is considered authenticated and the client is
626 allowed to proceed by sending a command to be executed. From this
627 point on the communication is encrypted using the stream cipher with
628 the session key known to both peers.
629
630 paraslash relies on the quality of the pseudo-random bytes provided
631 by the crypto library (openssl or libgcrypt), on the security of
632 the implementation of the RSA and AES crypto routines and on the
633 infeasibility to invert the SHA1 function.
634
635 Neither para_server or para_client create RSA keys on their
636 own. This has to be done once for each user as sketched in
637 [Quick start](#Quick.start) and discussed in more detail
638 [below](#The.user_list.file).
639
640 The user_list file
641 ------------------
642
643 At startup para_server reads the user list file which contains one
644 line per user. The default location of the user list file may be
645 changed with the --user-list option.
646
647 There should be at least one user in this file. Each user must have
648 an RSA key pair. The public part of the key is needed by para_server
649 while the private key is needed by para_client. Each line of the
650 user list file must be of the form
651
652 user <username> <key> <perms>
653
654 where _username_ is an arbitrary string (usually the user's login
655 name), _key_ is the full path to that user's public RSA key, and
656 _perms_ is a comma-separated list of zero or more of the following
657 permission bits:
658
659 +---------------------------------------------------------+
660 | AFS_READ | read the contents of the databases |
661 +-----------+---------------------------------------------+
662 | AFS_WRITE | change database contents |
663 +-----------+---------------------------------------------+
664 | VSS_READ | obtain information about the current stream |
665 +-----------+---------------------------------------------+
666 | VSS_WRITE | change the current stream |
667 +---------------------------------------------------------+
668
669 The permission bits specify which commands the user is allowed to
670 execute. The output of
671
672 para_client help
673
674 contains the permissions needed to execute the command.
675
676 It is possible to make para_server reread the user_list file by
677 executing the paraslash "hup" command or by sending SIGHUP to the
678 PID of para_server.
679
680 Connecting para_audiod
681 ----------------------
682
683 para_audiod listens on a Unix domain socket. Those sockets are
684 for local communication only, so only local users can connect to
685 para_audiod. The default is to let any user connect but this can be
686 restricted on platforms that support UNIX socket credentials which
687 allow para_audiod to obtain the Unix credentials of the connecting
688 process.
689
690 Use para_audiod's --user-allow option to allow connections only for
691 a limited set of users.
692
693 =======================
694 The audio file selector
695 =======================
696
697 paraslash comes with a sophisticated audio file selector (AFS),
698 whose main task is to determine which file to stream next, based on
699 information on the audio files stored in a database. It communicates
700 also with para_client via the command handler whenever an AFS command
701 is executed, for example to answer a database query.
702
703 Besides the simple playlists, AFS supports audio file selection
704 based on _moods_ which act as a filter that limits the set of all
705 known audio files to those which satisfy certain criteria. It also
706 maintains tables containing images (e.g. album cover art) and lyrics
707 that can be associated with one or more audio files.
708
709 In this chapter we sketch the setup of the [AFS
710 process](#The.AFS.process) during server startup and proceed with the
711 description of the [layout](#Database.layout) of the various database
712 tables. The section on [playlists and moods](#Playlists.and.moods)
713 explains these two audio file selection mechanisms in detail
714 and contains practical examples. The way [file renames and content
715 changes](#File.renames.and.content.changes) are detected is discussed
716 briefly before the [Troubleshooting](#Troubleshooting) section
717 concludes the chapter.
718
719 The AFS process
720 ---------------
721
722 On startup, para_server forks to create the AFS process which opens
723 the database tables. The AFS process accepts incoming connections
724 which arrive either on a pipe which is shared with para_server,
725 or on the local socket it is listening on. The setup is as follows.
726
727 .___________________. .______________.
728 | | | |
729 | virtual streaming | | audio format |
730 | system | | handler |
731 |_________ _______| |_____ ______|
732 \ / \ /
733 | |
734 .-'""""`-. | | .-'""""`-.
735 ( ) | | ( )
736 |`-.____.-'| .__/ \________________/ \___. |`-.____.-'|
737 | | | | | |
738 | file |----| AFS (audio file selector) |----| OSL |
739 | system | | process | | database |
740 | | |___________________________| | |
741 |. ' "" ` .| | |. ' "" ` .|
742 | | | | |
743 `-.____.-' | `-.____.-'
744 ._______/ \_______.
745 | |
746 | command handler |
747 |_______ _______|
748 \ /
749 |
750 |
751 |
752 ._____/ \_____.
753 | |
754 | para_client |
755 |_____________|
756
757 The virtual streaming system, which is part of the server process,
758 communicates with the AFS process via pipes and shared memory. When
759 the current audio file changes, it sends a notification through the
760 shared pipe. The AFS process queries the database to determine the
761 next audio file, opens it, verifies that it has not been changed since
762 it was added to the database and passes the open file descriptor back
763 to the virtual streaming system, along with audio file meta-data such
764 as file name, duration, audio format and so on. The virtual streaming
765 system then starts to stream the file.
766
767 The command handlers of all AFS server commands use the local socket
768 to query or update the database. For example, the command handler of
769 the add command sends the path of an audio file to the local socket.
770 The AFS process opens the file and tries to find an audio format
771 handler which recognizes the file. If all goes well, a new database
772 entry with metadata obtained from the audio format handler is added
773 to the database.
774
775 Note that AFS employs
776 [libosl](http://people.tuebingen.mpg.de/maan/osl/), the object
777 storage layer library, as the database backend. This library offers
778 functionality similar to a relational database, but is much more
779 lightweight than a full featured database management system.
780
781 Database layout
782 ---------------
783
784 Metadata about the known audio files is stored in an OSL database. This
785 database consists of the following tables:
786
787 - The audio file table contains path, hash and metadata of each
788 known file.
789
790 - The "attributes" table maps each of the 64 possible attributes to a
791 string.
792
793 - The "blob" tables store images, lyrics, moods, playlists. All of
794 these are optional.
795
796 - The "score" table describes the subset of admissible files for the
797 current playlist or mood.
798
799 All tables are described in more detail below.
800
801 <h3> The audio file table </h3>
802
803 This is the most important and usually also the largest table of the
804 AFS database. It contains the information needed to stream each audio
805 file. In particular the following data is stored for each audio file.
806
807 - SHA1 hash value of the audio file contents. This is computed once
808 when the file is added to the database. Whenever AFS selects this
809 audio file for streaming the hash value is recomputed and checked
810 against the value stored in the database to detect content changes.
811
812 - The time when this audio file was last played.
813
814 - The number of times the file has been played so far.
815
816 - The attribute bitmask.
817
818 - The image id which describes the image associated with this audio
819 file.
820
821 - The lyrics id which describes the lyrics associated with this
822 audio file.
823
824 - The audio format id (MP3, OGG, ...).
825
826 - An amplification value that can be used by the amplification filter
827 to pre-amplify the decoded audio stream.
828
829 - The chunk table. It describes the location and the timing of the
830 building blocks of the audio file. This is used by para_server to
831 send chunks of the file at appropriate times.
832
833 - The duration of the audio file.
834
835 - Tag information contained in the audio file (ID3 tags, Vorbis
836 comments, ...).
837
838 - The number of channels
839
840 - The encoding bitrate.
841
842 - The sampling frequency.
843
844 To add or refresh the data contained in the audio file table, the _add_
845 command is used. It takes the full path of either an audio file or a
846 directory. In the latter case, the directory is traversed recursively
847 and all files which are recognized as valid audio files are added to
848 the database.
849
850 <h3> The attribute table </h3>
851
852 The attribute table contains two columns, _name_ and _bitnum_. An
853 attribute is simply a name for a certain bit number in the attribute
854 bitmask of the audio file table.
855
856 Each of the 64 bits of the attribute bitmask can be set for each
857 audio file individually. Hence up to 64 different attributes may be
858 defined. For example, "pop", "rock", "blues", "jazz", "instrumental",
859 "german_lyrics", "speech", whatever. You are free to choose as
860 many attributes as you like and there are no naming restrictions
861 for attributes.
862
863 A new attribute "test" is created by
864
865 para_client addatt test
866 and
867 para_client lsatt
868
869 lists all available attributes. You can set the "test" attribute for
870 an audio file by executing
871
872 para_client setatt test+ /path/to/the/audio/file
873
874 Similarly, the "test" bit can be removed from an audio file with
875
876 para_client setatt test- /path/to/the/audio/file
877
878 Instead of a path you may use a shell wildcard pattern. The attribute
879 is applied to all audio files matching this pattern:
880
881 para_client setatt test+ '/test/directory/*'
882
883 The command
884
885 para_client -- ls -l=v
886
887 gives you a verbose listing of your audio files also showing which
888 attributes are set.
889
890 In case you wonder why the double-dash in the above command is needed:
891 It tells para_client to not interpret the options after the dashes. If
892 you find this annoying, just say
893
894 alias para='para_client --'
895
896 and be happy. In what follows we shall use this alias.
897
898 The "test" attribute can be dropped from the database with
899
900 para rmatt test
901
902 Read the output of
903
904 para help ls
905 para help setatt
906
907 for more information and a complete list of command line options to
908 these commands.
909
910 <h3> Blob tables </h3>
911
912 The image, lyrics, moods and playlists tables are all blob tables.
913 Blob tables consist of three columns each: The identifier which is
914 a positive number that is auto-incremented, the name (an arbitrary
915 string) and the content (the blob).
916
917 All blob tables support the same set of actions: cat, ls, mv, rm
918 and add. Of course, _add_ is used for adding new blobs to the table
919 while the other actions have the same meaning as the corresponding
920 Unix commands. The paraslash commands to perform these actions are
921 constructed as the concatenation of the table name and the action. For
922 example addimg, catimg, lsimg, mvimg, rmimg are the commands that
923 manipulate or query the image table.
924
925 The add variant of these commands is special as these commands read
926 the blob contents from stdin. To add an image to the image table the
927 command
928
929 para addimg image_name < file.jpg
930
931 can be used.
932
933 Note that the images and lyrics are not interpreted at all, and also
934 the playlist and the mood blobs are only investigated when the mood
935 or playlist is activated with the select command.
936
937 <h3> The score table </h3>
938
939 The score table describes those audio files which are admissible for
940 the current mood or playlist (see below). The table has two columns:
941 a pointer to a row of the audio file table and a score value.
942
943 Unlike all other tables of the database, the score table remains in
944 memory and is never stored on disk. It is initialized at startup and
945 recomputed when the select command loads a new mood or playlist.
946
947 When the audio file selector is asked to open the next audio file,
948 it picks the row with the highest score, opens the corresponding
949 file and passes the file descriptor to the virtual streaming system.
950 At this point the last_played and the num_played fields of the selected
951 file are updated and the score is recomputed.
952
953 Playlists and moods
954 -------------------
955
956 Playlists and moods offer two different ways of specifying the set of
957 admissible files. A playlist in itself describes a set of admissible
958 files. A mood, in contrast, describes the set of admissible files in
959 terms of attributes and other type of information available in the
960 audio file table. As an example, a mood can define a filename pattern,
961 which is then matched against the names of audio files in the table.
962
963 <h3> Playlists </h3>
964
965 Playlists are accommodated in the playlist table of the afs database,
966 using the aforementioned blob format for tables. A new playlist is
967 created with the addpl command by specifying the full (absolute)
968 paths of all desired audio files, separated by newlines. Example:
969
970 find /my/mp3/dir -name "*.mp3" | para addpl my_playlist
971
972 If _my_playlist_ already exists it is overwritten. To activate the
973 new playlist, execute
974
975 para select p/my_playlist
976
977 The audio file selector will assign scores to each entry of the list,
978 in descending order so that files will be selected in order. If a
979 file could not be opened for streaming, its entry is removed from
980 the score table (but not from the playlist).
981
982 <h3> Moods </h3>
983
984 A mood consists of a unique name and a definition. The definition
985 is an expression which describes which audio files are considered
986 admissible. At any time at most one mood can be active, meaning
987 that para_server will only stream files which are admissible for the
988 active mood.
989
990 The expression may refer to attributes and other metadata stored in
991 the database. Expressions may be combined by means of logical and
992 arithmetical operators in a natural way. Moreover, string matching
993 based on regular expression or wildcard patterns is supported.
994
995 The set of admissible files is determined by applying the expression
996 to each audio file in turn. For a mood definition to be valid, its
997 expression must evaluate to a number, a string or a boolean value
998 ("true" or "false"). For numbers, any value other than zero means the
999 file is admissible. For strings, any non-empty string indicates an
1000 admissible file. For boolean values, true means admissible and false
1001 means not admissible. As a special case, the empty expression treats
1002 all files as admissible.
1003
1004 <h3> Mood grammar </h3>
1005
1006 Expressions are based on a context-free grammar which distinguishes
1007 between several types for syntactic units or groupings. The grammar
1008 defines a set of keywords which have a type and a corresponding
1009 semantic value, as shown in the following table.
1010
1011 Keyword | Type | Semantic value
1012 :--------------------|--------:|:----------------------------------
1013 `path` | string | Full path of the current audio file
1014 `artist` | string | Content of the artist meta tag
1015 `title` | string | Content of the title meta tag
1016 `album` | string | Content of the album meta tag
1017 `comment` | string | Content of the somment meta tag
1018 `num_attributes_set` | integer | Number of attributes which are set
1019 `year` | integer | Content of the year meta tag [\*]
1020 `num_played` | integer | How many times the file has been streamed
1021 `image_id` | integer | The identifier of the (cover art) image
1022 `lyrics_id` | integer | The identifier of the lyrics blob
1023 `bitrate` | integer | The average bitrate
1024 `frequency` | integer | The output sample rate
1025 `channels` | integer | The number of channels
1026 `is_set("foo")` | boolean | True if attribute "foo" is set.
1027
1028 [\*] For most audio formats, the year tag is stored as a string. It
1029 is converted to an integer by the mood parser. If the audio file
1030 has no year tag or the content of the year tag is not a number, the
1031 semantic value is zero. A special convention applies if the year tag
1032 is a one-digit or a two-digit number. In this case 1900 is added to
1033 the tag value.
1034
1035 Expressions may be grouped using parentheses, logical and
1036 arithmetical operators or string matching operators. The following
1037 table lists the available operators.
1038
1039 Token | Meaning
1040 :------|:-------
1041 `\|\|` | Logical Or
1042 `&&` | Logical And
1043 `!` | Logical Not
1044 `==` | Equal (can be applied to all types)
1045 `!=` | Not equal. Likewise
1046 `<` | Less than
1047 `<=` | Less or equal
1048 `>=` | Greater or equal
1049 `+` | Arithmetical minus
1050 `-` | Binary/unary minus
1051 `*` | Multiplication
1052 `/` | Division
1053 `=~` | Regular expression match
1054 `=\|` | Filename match
1055
1056 Besides integers, strings and booleans there is an additional type
1057 which describes regular expression or wildcard patterns. Patterns
1058 are not just strings because they also include a list of flags which
1059 modify matching behaviour.
1060
1061 Regular expression patterns are of the form `/pattern/[flags]`. That
1062 is, the pattern is delimited by slashes, and is followed by zero or
1063 more characters, each specifying a flag according to the following
1064 table
1065
1066 Flag | POSIX name | Meaning
1067 :----|--------------:|--------
1068 `i` | `REG_ICASE` | Ignore case in match
1069 `n` | `REG_NEWLINE` | Treat newline as an ordinary character
1070
1071 Note that only extended regular expression patterns are supported. See
1072 regex(3) for details.
1073
1074 Wildcard patterns are similar, but the pattern must be delimited by
1075 `'|'` characters rather than slashes. For wildcard patterns different
1076 flags exist, as shown below.
1077
1078 Flag | POSIX name | Meaning
1079 :----|-----------------------:|--------
1080 `n` | `FNM_NOESCAPE` | Treat backslash as an ordinary character
1081 `p` | `FNM_PATHNAME` | Match a slash only with a slash in pattern
1082 `P` | `FNM_PERIOD` | Leading period has to be matched exactly
1083 `l` | `FNM_LEADING_DIR` [\*] | Ignore "/\*" rest after successful matching
1084 `i` | `FNM_CASEFOLD` [\*] | Ignore case in match
1085 `e` | `FNM_EXTMATCH` [\*\*] | Enable extended pattern matching
1086
1087 [\*] Not in POSIX, but both FreeBSD and NetBSD have it.
1088
1089 [\*\*] GNU extension, silently ignored on non GNU systems.
1090
1091 See fnmatch(3) for details.
1092
1093 Mood definitions may contain arbitrary whitespace and comments.
1094 A comment is a word beginning with #. This word and all remaining
1095 characters of the line are ignored.
1096
1097 <h3> Example moods </h3>
1098
1099 * Files with no/invalid year tag: `year == 0`
1100
1101 * Only oldies: `year != 0 && year < 1980`
1102
1103 * Only 80's Rock or Metal: `(year >= 1980 && year < 1990) &&
1104 (is_set("rock") || is_set("metal"))`
1105
1106 * Files with incomplete tags: `artist == "" || title == "" || album =
1107 "" || comment == "" || year == 0`
1108
1109 * Files with no attributes defined so far: `num_attributes_set == 0`
1110
1111 * Only newly added files: `num_played == 0`
1112
1113 * Only poor quality files: `bitrate < 96`
1114
1115 * Cope with different spellings of Motörhead: `artist =~ /mot(ö|oe{0,1})rhead/i`
1116
1117 * The same with extended wildcard patterns: `artist =| |mot+(o\|oe\|ö)rhead|ie`
1118
1119 <h3> Mood usage </h3>
1120
1121 To create a new mood called "my_mood", write its definition into
1122 some temporary file, say "tmpfile", and add it to the mood table
1123 by executing
1124
1125 para addmood my_mood < tmpfile
1126
1127 If the mood definition is really short, you may just pipe it to the
1128 client instead of using temporary files. Like this:
1129
1130 echo "$MOOD_DEFINITION" | para addmood my_mood
1131
1132 There is no need to keep the temporary file since you can always use
1133 the catmood command to get it back:
1134
1135 para catmood my_mood
1136
1137 A mood can be activated by executing
1138
1139 para select m/my_mood
1140
1141 Once active, the list of admissible files is shown by the ls command
1142 if the "-a" switch is given:
1143
1144 para ls -a
1145
1146 File renames and content changes
1147 --------------------------------
1148
1149 Since the audio file selector knows the SHA1 of each audio file that
1150 has been added to the afs database, it recognizes if the content of
1151 a file has changed, e.g. because an ID3 tag was added or modified.
1152 Also, if a file has been renamed or moved to a different location,
1153 afs will detect that an entry with the same hash value already exists
1154 in the audio file table.
1155
1156 In both cases it is enough to just re-add the new file. In the
1157 first case (file content changed), the audio table is updated, while
1158 metadata such as the num_played and last_played fields, as well as
1159 the attributes, remain unchanged. In the other case, when the file
1160 is moved or renamed, only the path information is updated, all other
1161 data remains as before.
1162
1163 It is possible to change the behaviour of the add command by using the
1164 "-l" (lazy add) or the "-f" (force add) option.
1165
1166 Troubleshooting
1167 ---------------
1168
1169 Use the debug loglevel (-l debug) to show debugging info. All paraslash
1170 executables have a brief online help which is displayed when -h is
1171 given. The --detailed-help option prints the full help text.
1172
1173 If para_server crashed or was killed by SIGKILL (signal 9), it
1174 may refuse to start again because of "dirty osl tables". In this
1175 case you'll have to run the oslfsck program of libosl to fix your
1176 database:
1177
1178 oslfsck -fd ~/.paraslash/afs_database-0.4
1179
1180 However, make sure para_server isn't running before executing oslfsck.
1181
1182 If you don't mind to recreate your database you can start
1183 from scratch by removing the entire database directory, i.e.
1184
1185 rm -rf ~/.paraslash/afs_database-0.4
1186
1187 Be aware that this removes all attribute definitions, all playlists
1188 and all mood definitions and requires to re-initialize the tables.
1189
1190 Although oslfsck fixes inconsistencies in database tables it doesn't
1191 care about the table contents. To check for invalid table contents, use
1192
1193 para_client check
1194
1195 This prints out references to missing audio files as well as invalid
1196 playlists and mood definitions.
1197
1198 Similarly, para_audiod refuses to start if its socket file exists, since
1199 this indicates that another instance of para_audiod is running. After
1200 a crash a stale socket file might remain and you must run
1201
1202 para_audiod --force
1203
1204 once to fix it up.
1205
1206 =======================================
1207 Audio formats and audio format handlers
1208 =======================================
1209
1210 Audio formats
1211 -------------
1212
1213 The following audio formats are supported by paraslash:
1214
1215 <h3> MP3 </h3>
1216
1217 Mp3, MPEG-1 Audio Layer 3, is a common audio format for audio storage,
1218 designed as part of its MPEG-1 standard. An MP3 file is made up of
1219 multiple MP3 frames, which consist of a header and a data block. The
1220 size of an MP3 frame depends on the bit rate and on the number
1221 of channels. For a typical CD-audio file (sample rate of 44.1 kHz
1222 stereo), encoded with a bit rate of 128 kbit, an MP3 frame is about
1223 400 bytes large.
1224
1225 <h3> OGG/Vorbis </h3>
1226
1227 OGG is a standardized audio container format, while Vorbis is an
1228 open source codec for lossy audio compression. Since Vorbis is most
1229 commonly made available via the OGG container format, it is often
1230 referred to as OGG/Vorbis. The OGG container format divides data into
1231 chunks called OGG pages. A typical OGG page is about 4KB large. The
1232 Vorbis codec creates variable-bitrate (VBR) data, where the bitrate
1233 may vary considerably.
1234
1235 <h3> OGG/Speex </h3>
1236
1237 Speex is an open-source speech codec that is based on CELP (Code
1238 Excited Linear Prediction) coding. It is designed for voice
1239 over IP applications, has modest complexity and a small memory
1240 footprint. Wideband and narrowband (telephone quality) speech are
1241 supported. As for Vorbis audio, Speex bit-streams are often stored
1242 in OGG files. As of 2012 this codec is considered obsolete since the
1243 Oppus codec, described below, surpasses its performance in all areas.
1244
1245 <h3> OGG/Opus </h3>
1246
1247 Opus is a lossy audio compression format standardized through RFC
1248 6716 in 2012. It combines the speech-oriented SILK codec and the
1249 low-latency CELT (Constrained Energy Lapped Transform) codec. Like
1250 OGG/Vorbis and OGG/Speex, Opus data is usually encapsulated in OGG
1251 containers. All known software patents which cover Opus are licensed
1252 under royalty-free terms.
1253
1254 <h3> AAC </h3>
1255
1256 Advanced Audio Coding (AAC) is a standardized, lossy compression
1257 and encoding scheme for digital audio which is the default audio
1258 format for Apple's iPhone, iPod, iTunes. Usually MPEG-4 is used as
1259 the container format and audio files encoded with AAC have the .m4a
1260 extension. A typical AAC frame is about 700 bytes large.
1261
1262 <h3> WMA </h3>
1263
1264 Windows Media Audio (WMA) is an audio data compression technology
1265 developed by Microsoft. A WMA file is usually encapsulated in the
1266 Advanced Systems Format (ASF) container format, which also specifies
1267 how meta data about the file is to be encoded. The bit stream of WMA
1268 is composed of superframes, each containing one or more frames of
1269 2048 samples. For 16 bit stereo a WMA superframe is about 8K large.
1270
1271 <h3> FLAC </h3>
1272
1273 The Free Lossless Audio Codec (FLAC) compresses audio without quality
1274 loss. It gives better compression ratios than a general purpose
1275 compressor like zip or bzip2 because FLAC is designed specifically
1276 for audio. A FLAC-encoded file consists of frames of varying size, up
1277 to 16K. Each frame starts with a header that contains all information
1278 necessary to decode the frame.
1279
1280 Meta data
1281 ---------
1282
1283 Unfortunately, each audio format has its own conventions how meta
1284 data is added as tags to the audio file.
1285
1286 For MP3 files, ID3, version 1 and 2 are widely used. ID3 version 1
1287 is rather simple but also very limited as it supports only artist,
1288 title, album, year and comment tags. Each of these can only be at most
1289 32 characters long. ID3, version 2 is much more flexible but requires
1290 a separate library being installed for paraslash to support it.
1291
1292 Ogg vorbis, ogg speex and flac files contain meta data as Vorbis
1293 comments, which are typically implemented as strings of the form
1294 "[TAG]=[VALUE]". Unlike ID3 version 1 tags, one may use whichever
1295 tags are appropriate for the content.
1296
1297 AAC files usually use the MPEG-4 container format for storing meta
1298 data while WMA files wrap meta data as special objects within the
1299 ASF container format.
1300
1301 paraslash only tracks the most common tags that are supported by
1302 all tag variants: artist, title, year, album, comment. When a file
1303 is added to the AFS database, the meta data of the file is extracted
1304 and stored in the audio file table.
1305
1306 Chunks and chunk tables
1307 -----------------------
1308
1309 paraslash uses the word "chunk" as common term for the building blocks
1310 of an audio file. For MP3 files, a chunk is the same as an MP3 frame,
1311 while for OGG files a chunk is an OGG page, etc. Therefore the chunk
1312 size varies considerably between audio formats, from a few hundred
1313 bytes (MP3) up to 16K (FLAC).
1314
1315 The chunk table contains the offsets within the audio file that
1316 correspond to the chunk boundaries of the file. Like the meta data,
1317 the chunk table is computed and stored in the database whenever an
1318 audio file is added.
1319
1320 The paraslash senders (see below) always send complete chunks. The
1321 granularity for seeking is therefore determined by the chunk size.
1322
1323 Audio format handlers
1324 ---------------------
1325
1326 For each audio format paraslash contains an audio format handler whose
1327 first task is to tell whether a given file is a valid audio file of
1328 this type. If so, the audio file handler extracts some technical data
1329 (duration, sampling rate, number of channels etc.), computes the
1330 chunk table and reads the meta data.
1331
1332 The audio format handler code is linked into para_server and executed
1333 via the _add_ command. The same code is also available as a stand-alone
1334 tool, para_afh, which prints the technical data, the chunk table
1335 and the meta data of a file. Moreover, all audio format handlers are
1336 combined in the afh receiver which is part of para_recv and para_play.
1337
1338 ==========
1339 Networking
1340 ==========
1341
1342 Paraslash uses different network connections for control and data.
1343 para_client communicates with para_server over a dedicated TCP control
1344 connection. To transport audio data, separate data connections are
1345 used. For these data connections, a variety of transports (UDP, DCCP,
1346 HTTP) can be chosen.
1347
1348 The chapter starts with the [control
1349 service](#The.paraslash.control.service), followed by a section
1350 on the various [streaming protocols](#Streaming.protocols)
1351 in which the data connections are described. The way
1352 audio file headers are embedded into the stream is discussed
1353 [briefly](#Streams.with.headers.and.headerless.streams) before the
1354 [example section](#Networking.examples) which illustrates typical
1355 commands for real-life scenarios.
1356
1357 Both IPv4 and IPv6 are supported.
1358
1359 The paraslash control service
1360 -----------------------------
1361
1362 para_server is controlled at runtime via the paraslash control
1363 connection. This connection is used for server commands (play, stop,
1364 ...) as well as for afs commands (ls, select, ...).
1365
1366 The server listens on a TCP port and accepts connections from clients
1367 that connect the open port. Each connection causes the server to fork
1368 off a client process which inherits the connection and deals with that
1369 client only. In this classical accept/fork approach the server process
1370 is unaffected if the child dies or goes crazy for whatever reason. In
1371 fact, the child process can not change address space of server process.
1372
1373 The section on [client-server
1374 authentication](#Client-server.authentication) above described the
1375 early connection establishment from the crypto point of view. Here
1376 it is described what happens after the connection (including crypto
1377 setup) has been established. There are four processes involved during
1378 command dispatch as sketched in the following diagram.
1379
1380 server_host client_host
1381 ~~~~~~~~~~~ ~~~~~~~~~~~
1382
1383 +-----------+ connect +-----------+
1384 |para_server|<------------------------------ |para_client|
1385 +-----------+ +-----------+
1386 | ^
1387 | fork +---+ |
1388 +----------> |AFS| |
1389 | +---+ |
1390 | ^ |
1391 | | |
1392 | | connect (cookie) |
1393 | | |
1394 | | |
1395 | fork +-----+ inherited connection |
1396 +---------->|child|<--------------------------+
1397 +-----+
1398
1399 Note that the child process is not a child of the afs process,
1400 so communication of these two processes has to happen via local
1401 sockets. In order to avoid abuse of the local socket by unrelated
1402 processes, a magic cookie is created once at server startup time just
1403 before the server process forks off the AFS process. This cookie is
1404 known to the server, AFS and the child, but not to unrelated processes.
1405
1406 There are two different kinds of commands: First there are commands
1407 that cause the server to respond with some answer such as the list
1408 of all audio files. All but the addblob commands (addimg, addlyr,
1409 addpl, addmood) are of this kind. The addblob commands add contents
1410 to the database, so they need to transfer data the other way round,
1411 from the client to the server.
1412
1413 There is no knowledge about the server commands built into para_client,
1414 so it does not know about addblob commands. Instead, the server sends
1415 a special "awaiting data" packet for these commands. If the client
1416 receives this packet, it sends STDIN to the server, otherwise it
1417 dumps data from the server to STDOUT.
1418
1419 Streaming protocols
1420 -------------------
1421
1422 A network (audio) stream usually consists of one streaming source,
1423 the _sender_, and one or more _receivers_ which read data over the
1424 network from the streaming source.
1425
1426 Senders are thus part of para_server while receivers are part of
1427 para_audiod. Moreover, there is the stand-alone tool para_recv which
1428 can be used to manually download a stream, either from para_server
1429 or from a web-based audio streaming service.
1430
1431 The following three streaming protocols are supported by paraslash:
1432
1433 - HTTP. Recommended for public streams that can be played by any
1434 player like mpg123, xmms, itunes, winamp, etc. The HTTP sender is
1435 supported on all operating systems and all platforms.
1436
1437 - DCCP. Recommended for LAN streaming. DCCP is currently available
1438 only for Linux.
1439
1440 - UDP. Recommended for multicast LAN streaming.
1441
1442 See the Appendix on [network protocols](#Network.protocols)
1443 for brief descriptions of the various protocols relevant for network
1444 audio streaming with paraslash.
1445
1446 It is possible to activate more than one sender simultaneously.
1447 Senders can be controlled at run time and via config file and command
1448 line options.
1449
1450 Note that audio connections are _not_ encrypted. Transport or Internet
1451 layer encryption should be used if encrypted data connections are
1452 needed.
1453
1454 Since DCCP and TCP are both connection-oriented protocols, connection
1455 establishment/teardown and access control are very similar between
1456 these two streaming protocols. UDP is the most lightweight option,
1457 since in contrast to TCP/DCCP it is connectionless. It is also the
1458 only protocol supporting IP multicast.
1459
1460 The HTTP and the DCCP sender listen on a (TCP/DCCP) port waiting for
1461 clients to connect and establish a connection via some protocol-defined
1462 handshake mechanism. Both senders maintain two linked lists each:
1463 The list of all clients which are currently connected, and the list
1464 of access control entries which determines who is allowed to connect.
1465 IP-based access control may be configured through config file and
1466 command line options and via the "allow" and "deny" sender subcommands.
1467
1468 Upon receiving a GET request from the client, the HTTP sender sends
1469 back a status line and a message. The body of this message is the
1470 audio stream. This is common practice and is supported by many popular
1471 clients which can thus be used to play a stream offered by para_server.
1472 For DCCP things are a bit simpler: No messages are exchanged between
1473 the receiver and sender. The client simply connects and the sender
1474 starts to stream.
1475
1476 DCCP is an experimental protocol which offers a number of new features
1477 not available for TCP. Both ends can negotiate these features using
1478 a built-in negotiation mechanism. In contrast to TCP/HTTP, DCCP is
1479 datagram-based (no retransmissions) and thus should not be used over
1480 lossy media (e.g. WiFi networks). One useful feature offered by DCCP
1481 is access to a variety of different congestion-control mechanisms
1482 called CCIDs. Two different CCIDs are available per default on Linux:
1483
1484
1485 - _CCID 2_. A Congestion Control mechanism similar to that of TCP. The
1486 sender maintains a congestion window and halves this window in response
1487 to congestion.
1488
1489
1490 - _CCID-3_. Designed to be fair when competing for bandwidth.
1491 It has lower variation of throughput over time compared with TCP,
1492 which makes it suitable for streaming media.
1493
1494 Unlike the HTTP and DCCP senders, the UDP sender maintains only a
1495 single list, the _target list_. This list describes the set of clients
1496 to which the stream is sent. There is no list for access control and
1497 no "allow" and "deny" commands for the UDP sender. Instead, the "add"
1498 and "delete" commands can be used to modify the target list.
1499
1500 Since both UDP and DCCP offer an unreliable datagram-based transport,
1501 additional measures are necessary to guard against disruptions over
1502 networks that are lossy or which may be subject to interference (as
1503 is for instance the case with WiFi). Paraslash uses FEC (Forward
1504 Error Correction) to guard against packet losses and reordering. The
1505 stream is FEC-encoded before it is sent through the UDP socket and
1506 must be decoded accordingly on the receiver side.
1507
1508 The packet size and the amount of redundancy introduced by FEC can
1509 be configured via the FEC parameters which are dictated by server
1510 and may also be configured through the "sender" command. The FEC
1511 parameters are encoded in the header of each network packet, so no
1512 configuration is necessary on the receiver side. See the section on
1513 [FEC](#Forward.error.correction) below.
1514
1515 Streams with headers and headerless streams
1516 -------------------------------------------
1517
1518 For OGG/Vorbis, OGG/Speex and wma streams, some of the information
1519 needed to decode the stream is only contained in the audio file
1520 header of the container format but not in each data chunk. Clients
1521 must be able to obtain this information in case streaming starts in
1522 the middle of the file or if para_audiod is started while para_server
1523 is already sending a stream.
1524
1525 This is accomplished in different ways, depending on the streaming
1526 protocol. For connection-oriented streams (HTTP, DCCP) the audio file
1527 header is sent prior to audio file data. This technique however does
1528 not work for the connectionless UDP transport. Hence the audio file
1529 header is periodically being embedded into the UDP audio data stream.
1530 By default, the header is resent after five seconds. The receiver has
1531 to wait until the next header arrives before it can start decoding
1532 the stream.
1533
1534 Networking examples
1535 -------------------
1536
1537 The "si" (server info) command lists some information about the
1538 currently running server process.
1539
1540 -> Show PIDs, number of connected clients, uptime, and more:
1541
1542 para_client si
1543
1544 By default para_server activates both the HTTP and th DCCP sender on
1545 startup. This can be changed via command line options or para_server's
1546 config file.
1547
1548 -> List config file options for senders:
1549
1550 para_server -h
1551
1552 -> Receive a DCCP stream using CCID2 and write the output into a file:
1553
1554 host=foo.org; ccid=2; filename=bar
1555 para_recv --receiver "dccp --host $host --ccid $ccid" > $filename
1556
1557 Note the quotes around the arguments for the dccp receiver. Each
1558 receiver has its own set of command line options and its own command
1559 line parser, so arguments for the dccp receiver must be protected
1560 from being interpreted by para_recv.
1561
1562 -> Receive FEC-encoded multicast stream and write the output into a file:
1563
1564 filename=foo
1565 para_recv -r udp > $filename
1566
1567 -> Receive this (FEC-encoded) unicast stream:
1568
1569 filename=foo
1570 para_recv -r 'udp -i 0.0.0.0' > $filename
1571
1572 -> Create a minimal config for para_audiod for HTTP streams:
1573
1574 c=$HOME/.paraslash/audiod.conf.min; s=server.foo.com
1575 echo receiver \".:http -i $s\" > $c
1576 para_audiod --config $c
1577
1578 =======
1579 Filters
1580 =======
1581
1582 A paraslash filter is a module which transforms an input stream into
1583 an output stream. Filters are included in the para_audiod executable
1584 and in the stand-alone tool para_filter which usually contains the
1585 same modules.
1586
1587 While para_filter reads its input stream from STDIN and writes
1588 the output to STDOUT, the filter modules of para_audiod are always
1589 connected to a receiver which produces the input stream and a writer
1590 which absorbs the output stream.
1591
1592 Some filters depend on a specific library and are not compiled in
1593 if this library was not found at compile time. To see the list of
1594 supported filters, run para_filter and para_audiod with the --help
1595 option. The output looks similar to the following:
1596
1597 Available filters:
1598 compress wav amp fecdec wmadec prebuffer oggdec aacdec mp3dec
1599
1600 Out of these filter modules, a chain of filters can be constructed,
1601 much in the way Unix pipes can be chained, and analogous to the use
1602 of modules in gstreamer: The output of the first filter becomes the
1603 input of the second filter. There is no limitation on the number of
1604 filters and the same filter may occur more than once.
1605
1606 Like receivers, each filter has its own command line options which
1607 must be quoted to protect them from the command line options of
1608 the driving application (para_audiod or para_filter). Example:
1609
1610 para_filter -f 'mp3dec --ignore-crc' -f 'compress --damp 1'
1611
1612 For para_audiod, each audio format has its own set of filters. The
1613 name of the audio format for which the filter should be applied can
1614 be used as the prefix for the filter option. Example:
1615
1616 para_audiod -f 'mp3:prebuffer --duration 300'
1617
1618 The "mp3" prefix above is actually interpreted as a POSIX extended
1619 regular expression. Therefore
1620
1621 para_audiod -f '.:prebuffer --duration 300'
1622
1623 activates the prebuffer filter for all supported audio formats (because
1624 "." matches all audio formats) while
1625
1626 para_audiod -f 'wma|ogg:prebuffer --duration 300'
1627
1628 activates it only for wma and ogg streams.
1629
1630 Decoders
1631 --------
1632
1633 For each supported audio format there is a corresponding filter
1634 which decodes audio data in this format to 16 bit PCM data which
1635 can be directly sent to the sound device or any other software that
1636 operates on undecoded PCM data (visualizers, equalizers etc.). Such
1637 filters are called _decoders_ in general, and xxxdec is the name of
1638 the paraslash decoder for the audio format xxx. For example, the mp3
1639 decoder is called mp3dec.
1640
1641 Note that the output of the decoder is about 10 times larger than
1642 its input. This means that filters that operate on the decoded audio
1643 stream have to deal with much more data than filters that transform
1644 the audio stream before it is fed to the decoder.
1645
1646 Paraslash relies on external libraries for most decoders, so these
1647 libraries must be installed for the decoder to be included in the
1648 executables. For example, the mp3dec filter depends on the mad library.
1649
1650 Forward error correction
1651 ------------------------
1652
1653 As already mentioned [earlier](#Streaming.protocols), paraslash
1654 uses forward error correction (FEC) for the unreliable UDP and
1655 DCCP transports. FEC is a technique which was invented already in
1656 1960 by Reed and Solomon and which is widely used for the parity
1657 calculations of storage devices (RAID arrays). It is based on the
1658 algebraic concept of finite fields, today called Galois fields, in
1659 honour of the mathematician Galois (1811-1832). The FEC implementation
1660 of paraslash is based on code by Luigi Rizzo.
1661
1662 Although the details require a sound knowledge of the underlying
1663 mathematics, the basic idea is not hard to understand: For positive
1664 integers k and n with k < n it is possible to compute for any k given
1665 data bytes d_1, ..., d_k the corresponding r := n -k parity bytes p_1,
1666 ..., p_r such that all data bytes can be reconstructed from *any*
1667 k bytes of the set
1668
1669 {d_1, ..., d_k, p_1, ..., p_r}.
1670
1671 FEC-encoding for unreliable network transports boils down to slicing
1672 the audio stream into groups of k suitably sized pieces called _slices_
1673 and computing the r corresponding parity slices. This step is performed
1674 in para_server which then sends both the data and the parity slices
1675 over the unreliable network connection. If the client was able
1676 to receive at least k of the n = k + r slices, it can reconstruct
1677 (FEC-decode) the original audio stream.
1678
1679 From these observations it is clear that there are three different
1680 FEC parameters: The slice size, the number of data slices k, and the
1681 total number of slices n. It is crucial to choose the slice size
1682 such that no fragmentation of network packets takes place because
1683 FEC only guards against losses and reordering but fails if slices are
1684 received partially.
1685
1686 FEC decoding in paralash is performed through the fecdec filter which
1687 usually is the first filter (there can be other filters before fecdec
1688 if these do not alter the audio stream).
1689
1690 Volume adjustment (amp and compress)
1691 ------------------------------------
1692
1693 The amp and the compress filter both adjust the volume of the audio
1694 stream. These filters operate on uncompressed audio samples. Hence
1695 they are usually placed directly after the decoding filter. Each
1696 sample is multiplied with a scaling factor (>= 1) which makes amp
1697 and compress quite expensive in terms of computing power.
1698
1699 <h3> amp </h3>
1700
1701 The amp filter amplifies the audio stream by a fixed scaling factor
1702 that must be known in advance. For para_audiod this factor is derived
1703 from the amplification field of the audio file's entry in the audio
1704 file table while para_filter uses the value given at the command line.
1705
1706 The optimal scaling factor F for an audio file is the largest real
1707 number F >= 1 such that after multiplication with F all samples still
1708 fit into the sample interval [-32768, 32767]. One can use para_filter
1709 in combination with the sox utility to compute F:
1710
1711 para_filter -f mp3dec -f wav < file.mp3 | sox -t wav - -e stat -v
1712
1713 The amplification value V which is stored in the audio file table,
1714 however, is an integer between 0 and 255 which is connected to F
1715 through the formula
1716
1717 V = (F - 1) * 64.
1718
1719 To store V in the audio file table, the command
1720
1721 para_client -- touch -a=V file.mp3
1722
1723 is used. The reader is encouraged to write a script that performs
1724 these computations :)
1725
1726 <h3> compress </h3>
1727
1728 Unlike the amplification filter, the compress filter adjusts the volume
1729 of the audio stream dynamically without prior knowledge about the peak
1730 value. It maintains the maximal volume of the last n samples of the
1731 audio stream and computes a suitable amplification factor based on that
1732 value and the various configuration options. It tries to chose this
1733 factor such that the adjusted volume meets the desired target level.
1734
1735 Note that it makes sense to combine amp and compress.
1736
1737 Misc filters (wav and prebuffer)
1738 --------------------------------
1739
1740 These filters are rather simple and do not modify the audio stream at
1741 all. The wav filter is only useful with para_filter and in connection
1742 with a decoder. It asks the decoder for the number of channels and the
1743 sample rate of the stream and adds a Microsoft wave header containing
1744 this information at the beginning. This allows to write wav files
1745 rather than raw PCM files (which do not contain any information about
1746 the number of channels and the sample rate).
1747
1748 The prebuffer filter simply delays the output until the given time has
1749 passed (starting from the time the first byte was available in its
1750 input queue) or until the given amount of data has accumulated. It
1751 is mainly useful for para_audiod if the standard parameters result
1752 in buffer underruns.
1753
1754 Both filters require almost no additional computing time, even when
1755 operating on uncompressed audio streams, since data buffers are simply
1756 "pushed down" rather than copied.
1757
1758 ======
1759 Output
1760 ======
1761
1762 Once an audio stream has been received and decoded to PCM format,
1763 it can be sent to a sound device for playback. This part is performed
1764 by paraslash _writers_ which are described in this chapter.
1765
1766 Writers
1767 -------
1768
1769 A paraslash writer acts as a data sink that consumes but does not
1770 produce audio data. Paraslash writers operate on the client side and
1771 are contained in para_audiod and in the stand-alone tool para_write.
1772
1773 The para_write program reads uncompressed audio data from STDIN. If
1774 this data starts with a wav header, sample rate, sample format and
1775 channel count are read from the header. Otherwise CD audio (44.1KHz
1776 16 bit little endian, stereo) is assumed but this can be overridden
1777 by command line options. para_audiod, on the other hand, obtains
1778 the sample rate and the number of channels from the decoder.
1779
1780 Like receivers and filters, each writer has an individual set of
1781 command line options, and for para_audiod writers can be configured
1782 per audio format separately. It is possible to activate more than
1783 one writer for the same stream simultaneously.
1784
1785 OS-dependent APIs
1786 -----------------
1787
1788 Unfortunately, the various flavours of Unix on which paraslash
1789 runs on have different APIs for opening a sound device and starting
1790 playback. Hence for each such API there is a paraslash writer that
1791 can play the audio stream via this API.
1792
1793 - *ALSA*. The _Advanced Linux Sound Architecture_ is only available on
1794 Linux systems. Although there are several mid-layer APIs in use by
1795 the various Linux distributions (ESD, Jack, PulseAudio), paraslash
1796 currently supports only the low-level ALSA API which is not supposed
1797 to be change. ALSA is very feature-rich, in particular it supports
1798 software mixing via its DMIX plugin. ALSA is the default writer on
1799 Linux systems.
1800
1801 - *OSS*. The _Open Sound System_ is the only API on \*BSD Unixes and
1802 is also available on Linux systems, usually provided by ALSA as an
1803 emulation for backwards compatibility. This API is rather simple but
1804 also limited. For example only one application can open the device
1805 at any time. The OSS writer is activated by default on BSD Systems.
1806
1807 - *FILE*. The file writer allows to capture the audio stream and
1808 write the PCM data to a file on the file system rather than playing
1809 it through a sound device. It is supported on all platforms and is
1810 always compiled in.
1811
1812 - *AO*. _Libao_ is a cross-platform audio library which supports a wide
1813 variety of platforms including PulseAudio (gnome), ESD (Enlightened
1814 Sound Daemon), AIX, Solaris and IRIX. The ao writer plays audio
1815 through an output plugin of libao.
1816
1817 Examples
1818 --------
1819
1820 -> Use the OSS writer to play a wav file:
1821
1822 para_write --writer oss < file.wav
1823
1824 -> Enable ALSA software mixing for mp3 streams:
1825
1826 para_audiod --writer 'mp3:alsa -d plug:swmix'
1827
1828
1829 ===
1830 Gui
1831 ===
1832
1833 para_gui executes an arbitrary command which is supposed to print
1834 status information to STDOUT. It then displays this information in
1835 a curses window. By default the command
1836
1837 para_audioc -- stat -p
1838
1839 is executed, but this can be customized via the --stat-cmd option. In
1840 particular it possible to use
1841
1842 para_client -- stat -p
1843
1844 to make para_gui work on systems on which para_audiod is not running.
1845
1846 Key bindings
1847 ------------
1848
1849 It is possible to bind keys to arbitrary commands via custom
1850 key-bindings. Besides the internal keys which can not be changed (help,
1851 quit, loglevel, version...), the following flavours of key-bindings
1852 are supported:
1853
1854 - external: Shutdown curses before launching the given command.
1855 Useful for starting other ncurses programs from within para_gui,
1856 e.g. aumix or dialog scripts. Or, use the mbox output format to write
1857 a mailbox containing one mail for each (admissible) file the audio
1858 file selector knows about. Then start mutt from within para_gui to
1859 browse your collection!
1860
1861 - display: Launch the command and display its stdout in para_gui's
1862 bottom window.
1863
1864 - para: Like display, but start "para_client <specified command>"
1865 instead of "<specified command>".
1866
1867 The general form of a key binding is
1868
1869 key_map k:m:c
1870
1871 which maps key k to command c using mode m. Mode may be x, d or p
1872 for external, display and paraslash commands, respectively.
1873
1874 Themes
1875 ------
1876
1877 Currently there are only two themes for para_gui. It is easy, however,
1878 to add more themes. To create a new theme one has to define the
1879 position, color and geometry for for each status item that should be
1880 shown by this theme. See gui_theme.c for examples.
1881
1882 The "." and "," keys are used to switch between themes.
1883
1884 Examples
1885 --------
1886
1887 -> Show server info:
1888
1889 key_map "i:p:si"
1890
1891 -> Jump to the middle of the current audio file by pressing F5:
1892
1893 key_map "<F5>:p:jmp 50"
1894
1895 -> vi-like bindings for jumping around:
1896
1897 key_map "l:p:ff 10"
1898 key_map "h:p:ff 10-"
1899 key_map "w:p:ff 60"
1900 key_map "b:p:ff 60-"
1901
1902 -> Print the current date and time:
1903
1904 key_map "D:d:date"
1905
1906 -> Call other curses programs:
1907
1908 key_map "U:x:aumix"
1909 key_map "!:x:/bin/bash"
1910 key_map "^E:x:/bin/sh -c 'vi ~/.paraslash/gui.conf'"
1911
1912 ===========
1913 Development
1914 ===========
1915
1916 Contributing
1917 ------------
1918
1919 Paraslash is an open source project and contributions are
1920 welcome. Here's a list of things you can do to help the project:
1921
1922 - Report problems with building, installing or running the software.
1923 In particular, test the experimental git branches ("next" and "pu").
1924 This helps to identify and fix problems before the code gets merged
1925 and thus keeps the master branch as stable as possible.
1926 - Proofread the documentation (manual, web pages, man pages, source
1927 code documentation) and point out unclear or poorly written parts. If
1928 you are a native English speaker you will easily find a lot of text
1929 that could be improved.
1930 - Run analysis tools (coverity, afl, sparse, etc.) and report issues
1931 found by those tools.
1932 - Suggest new features you would like to see implemented.
1933 - Compile and test on your favorite architecture or operating
1934 system. The code is tested only on a limited set of systems, so you
1935 will probably encounter problems when building on different systems.
1936 - Post about paraslash on your blog or on social networks.
1937 - Build and maintain Debian/RPM packages for your favorite distribution.
1938
1939 Note that there is no mailing list, no bug tracker and no discussion
1940 forum for paraslash. If you'd like to contribute, or have questions
1941 about contributing, send email to Andre Noll <maan@tuebingen.mpg.de>.
1942 New releases are announced by email. If you would like to receive
1943 these announcements, contact the author through the above address.
1944
1945 Tools
1946 -----
1947
1948 In order to compile the sources from the git repository (rather than
1949 from tar balls) and for contributing non-trivial changes to the
1950 paraslash project, some additional tools should be installed on a
1951 developer machine.
1952
1953 - [git](http://git.or.cz/). As described in more detail
1954 [below](#Git.branches), the git source code management tool is used for
1955 paraslash development. It is necessary for cloning the git repository
1956 and for getting updates.
1957
1958 - [autoconf](ftp://ftp.gnu.org/pub/gnu/autoconf/) GNU autoconf creates
1959 the configure file which is shipped in the tarballs but has to be
1960 generated when compiling from git.
1961
1962 - [discount](http://www.pell.portland.or.us/~orc/Code/discount/). The
1963 HTML version of this manual and some of the paraslash web pages are
1964 written in the Markdown markup language and are translated into html
1965 with the converter of the *Discount* package.
1966
1967 - [doxygen](http://www.stack.nl/~dimitri/doxygen/). The documentation
1968 of paraslash's C sources uses the doxygen documentation system. The
1969 conventions for documenting the source code is described in the
1970 [Doxygen section](#Doxygen).
1971
1972 - [global](ftp://ftp.gnu.org/pub/gnu/global). This is used to generate
1973 browsable HTML from the C sources. It is needed by doxygen.
1974
1975 Git branches
1976 ------------
1977
1978 Paraslash has been developed using the git source code management
1979 tool since 2006. Development is organized roughly in the same spirit
1980 as the git development itself, as described below.
1981
1982 The following text passage is based on "A note from the maintainer",
1983 written by Junio C Hamano, the maintainer of git.
1984
1985 There are four branches in the paraslash repository that track the
1986 source tree: "master", "maint", "next", and "pu".
1987
1988 The "master" branch is meant to contain what is well tested and
1989 ready to be used in a production setting. There could occasionally be
1990 minor breakages or brown paper bag bugs but they are not expected to
1991 be anything major, and more importantly quickly and easily fixable.
1992 Every now and then, a "feature release" is cut from the tip of this
1993 branch, named with three dotted decimal digits, like 0.4.2.
1994
1995 Whenever changes are about to be included that will eventually lead to
1996 a new major release (e.g. 0.5.0), a "maint" branch is forked off from
1997 "master" at that point. Obvious, safe and urgent fixes after the major
1998 release are applied to this branch and maintenance releases are cut
1999 from it. New features never go to this branch. This branch is also
2000 merged into "master" to propagate the fixes forward.
2001
2002 A trivial and safe enhancement goes directly on top of "master".
2003 New development does not usually happen on "master", however.
2004 Instead, a separate topic branch is forked from the tip of "master",
2005 and it first is tested in isolation; Usually there are a handful such
2006 topic branches that are running ahead of "master". The tip of these
2007 branches is not published in the public repository to keep the number
2008 of branches that downstream developers need to worry about low.
2009
2010 The quality of topic branches varies widely. Some of them start out as
2011 "good idea but obviously is broken in some areas" and then with some
2012 more work become "more or less done and can now be tested by wider
2013 audience". Luckily, most of them start out in the latter, better shape.
2014
2015 The "next" branch is to merge and test topic branches in the latter
2016 category. In general, this branch always contains the tip of "master".
2017 It might not be quite rock-solid production ready, but is expected to
2018 work more or less without major breakage. The maintainer usually uses
2019 the "next" version of paraslash for his own pleasure, so it cannot
2020 be _that_ broken. The "next" branch is where new and exciting things
2021 take place.
2022
2023 The two branches "master" and "maint" are never rewound, and "next"
2024 usually will not be either (this automatically means the topics that
2025 have been merged into "next" are usually not rebased, and you can find
2026 the tip of topic branches you are interested in from the output of
2027 "git log next"). You should be able to safely build on top of them.
2028
2029 However, at times "next" will be rebuilt from the tip of "master" to
2030 get rid of merge commits that will never be in "master". The commit
2031 that replaces "next" will usually have the identical tree, but it
2032 will have different ancestry from the tip of "master".
2033
2034 The "pu" (proposed updates) branch bundles the remainder of the
2035 topic branches. The "pu" branch, and topic branches that are only in
2036 "pu", are subject to rebasing in general. By the above definition
2037 of how "next" works, you can tell that this branch will contain quite
2038 experimental and obviously broken stuff.
2039
2040 When a topic that was in "pu" proves to be in testable shape, it
2041 graduates to "next". This is done with
2042
2043 git checkout next
2044 git merge that-topic-branch
2045
2046 Sometimes, an idea that looked promising turns out to be not so good
2047 and the topic can be dropped from "pu" in such a case.
2048
2049 A topic that is in "next" is expected to be polished to perfection
2050 before it is merged to "master". Similar to the above, this is
2051 done with
2052
2053 git checkout master
2054 git merge that-topic-branch
2055 git branch -d that-topic-branch
2056
2057 Note that being in "next" is not a guarantee to appear in the next
2058 release (being in "master" is such a guarantee, unless it is later
2059 found seriously broken and reverted), nor even in any future release.
2060
2061 Coding Style
2062 ------------
2063
2064 The preferred coding style for paraslash coincides more or less
2065 with the style of the Linux kernel. So rather than repeating what is
2066 written [there](https://www.kernel.org/doc/Documentation/process/coding-style.rst),
2067 here are the most important points.
2068
2069 - Burn the GNU coding standards.
2070 - Never use spaces for indentation.
2071 - Tabs are 8 characters, and thus indentations are also 8 characters.
2072 - Don't put multiple assignments on a single line.
2073 - Avoid tricky expressions.
2074 - Don't leave whitespace at the end of lines.
2075 - The limit on the length of lines is 80 columns.
2076 - Use K&R style for placing braces and spaces:
2077
2078 if (x is true) {
2079 we do y
2080 }
2081
2082 - Use a space after (most) keywords.
2083 - Do not add spaces around (inside) parenthesized expressions.
2084 - Use one space around (on each side of) most binary and ternary operators.
2085 - Do not use cute names like ThisVariableIsATemporaryCounter, call it tmp.
2086 - Mixed-case names are frowned upon.
2087 - Descriptive names for global variables are a must.
2088 - Avoid typedefs.
2089 - Functions should be short and sweet, and do just one thing.
2090 - The number of local variables shouldn't exceed 10.
2091 - Gotos are fine if they improve readability and reduce nesting.
2092 - Don't use C99-style "// ..." comments.
2093 - Names of macros defining constants and labels in enums are capitalized.
2094 - Enums are preferred when defining several related constants.
2095 - Always use the paraslash wrappers for allocating memory.
2096 - If the name of a function is an action or an imperative.
2097 command, the function should return an error-code integer
2098 (<0 means error, >=0 means success). If the name is a
2099 predicate, the function should return a "succeeded" boolean.
2100
2101 Doxygen
2102 -------
2103
2104 Doxygen is a documentation system for various programming
2105 languages. The API reference on the paraslash web page is generated
2106 by doxygen.
2107
2108 It is more illustrative to look at the source code for examples than
2109 to describe the conventions in this manual, so we only describe which
2110 parts of the code need doxygen comments, but leave out details on
2111 documentation conventions.
2112
2113 As a rule, only the public part of the C source is documented with
2114 Doxygen. This includes structures, defines and enumerations in header
2115 files as well as public (non-static) C functions. These should be
2116 documented completely. For example, each parameter and the return
2117 value of a public function should get a descriptive doxygen comment.
2118
2119 No doxygen comments are necessary for static functions and for
2120 structures and enumerations in C files (which are used only within
2121 this file). This does not mean, however, that those entities need
2122 no documentation at all. Instead, common sense should be applied to
2123 document what is not obvious from reading the code.
2124
2125 ========
2126 Appendix
2127 ========
2128
2129 Network protocols
2130 -----------------
2131
2132 <h3> IP </h3>
2133
2134 The _Internet Protocol_ is the primary networking protocol used for
2135 the Internet. All protocols described below use IP as the underlying
2136 layer. Both the prevalent IPv4 and the next-generation IPv6 variant
2137 are being deployed actively worldwide.
2138
2139 <h3> Connection-oriented and connectionless protocols </h3>
2140
2141 Connectionless protocols differ from connection-oriented ones in
2142 that state associated with the sending/receiving endpoints is treated
2143 implicitly. Connectionless protocols maintain no internal knowledge
2144 about the state of the connection. Hence they are not capable of
2145 reacting to state changes, such as sudden loss or congestion on the
2146 connection medium. Connection-oriented protocols, in contrast, make
2147 this knowledge explicit. The connection is established only after
2148 a bidirectional handshake which requires both endpoints to agree
2149 on the state of the connection, and may also involve negotiating
2150 specific parameters for the particular connection. Maintaining an
2151 up-to-date internal state of the connection also in general means
2152 that the sending endpoints perform congestion control, adapting to
2153 qualitative changes of the connection medium.
2154
2155 <h3> Reliability </h3>
2156
2157 In IP networking, packets can be lost, duplicated, or delivered
2158 out of order, and different network protocols handle these
2159 problems in different ways. We call a transport-layer protocol
2160 _reliable_, if it turns the unreliable IP delivery into an ordered,
2161 duplicate- and loss-free delivery of packets. Sequence numbers
2162 are used to discard duplicates and re-arrange packets delivered
2163 out-of-order. Retransmission is used to guarantee loss-free
2164 delivery. Unreliable protocols, in contrast, do not guarantee ordering
2165 or data integrity.
2166
2167 <h3> Classification </h3>
2168
2169 With these definitions the protocols which are used by paraslash for
2170 steaming audio data may be classified as follows.
2171
2172 - HTTP/TCP: connection-oriented, reliable,
2173 - UDP: connectionless, unreliable,
2174 - DCCP: connection-oriented, unreliable.
2175
2176 Below we give a short descriptions of these protocols.
2177
2178 <h3> TCP </h3>
2179
2180 The _Transmission Control Protocol_ provides reliable, ordered delivery
2181 of a stream and a classic window-based congestion control. In contrast
2182 to UDP and DCCP (see below), TCP does not have record-oriented or
2183 datagram-based syntax, i.e. it provides a stream which is unaware
2184 and independent of any record (packet) boundaries. TCP is used
2185 extensively by many application layers. Besides HTTP (the Hypertext
2186 Transfer Protocol), also FTP (the File Transfer protocol), SMTP (Simple
2187 Mail Transfer Protocol), SSH (Secure Shell) all sit on top of TCP.
2188
2189 <h3> UDP </h3>
2190
2191 The _User Datagram Protocol_ is the simplest transport-layer protocol,
2192 built as a thin layer directly on top of IP. For this reason, it offers
2193 the same best-effort service as IP itself, i.e. there is no detection
2194 of duplicate or reordered packets. Being a connectionless protocol,
2195 only minimal internal state about the connection is maintained, which
2196 means that there is no protection against packet loss or network
2197 congestion. Error checking and correction (if at all) are performed
2198 in the application.
2199
2200 <h3> DCCP </h3>
2201
2202 The _Datagram Congestion Control Protocol_ combines the
2203 connection-oriented state maintenance known from TCP with the
2204 unreliable, datagram-based transport of UDP. This means that it
2205 is capable of reacting to changes in the connection by performing
2206 congestion control, offering multiple alternative approaches. But it
2207 is bound to datagram boundaries (the maximum packet size supported
2208 by a medium), and like UDP it lacks retransmission to protect
2209 against loss. Due to the use of sequence numbers, it is however
2210 able to react to loss (interpreted as a congestion indication) and
2211 to ignore out-of-order and duplicate packets. Unlike TCP it allows
2212 to negotiate specific, binding features for a connection, such as
2213 the choice of congestion control: classic, window-based congestion
2214 control known from TCP is available as CCID-2, rate-based, "smooth"
2215 congestion control is offered as CCID-3.
2216
2217 <h3> HTTP </h3>
2218
2219 The _Hypertext Transfer Protocol_ is an application layer protocol
2220 on top of TCP. It is spoken by web servers and is most often used
2221 for web services. However, as can be seen by the many Internet radio
2222 stations and YouTube/Flash videos, http is by far not limited to the
2223 delivery of web pages only. Being a simple request/response based
2224 protocol, the semantics of the protocol also allow the delivery of
2225 multimedia content, such as audio over http.
2226
2227 <h3> Multicast </h3>
2228
2229 IP multicast is not really a protocol but a technique for one-to-many
2230 communication over an IP network. The challenge is to deliver
2231 information to a group of destinations simultaneously using the
2232 most efficient strategy to send the messages over each link of the
2233 network only once. This has benefits for streaming multimedia: the
2234 standard one-to-one unicast offered by TCP/DCCP means that n clients
2235 listening to the same stream also consume n-times the resources,
2236 whereas multicast requires to send the stream just once, irrespective
2237 of the number of receivers. Since it would be costly to maintain state
2238 for each listening receiver, multicast often implies connectionless
2239 transport, which is the reason that it is currently only available
2240 via UDP.
2241
2242 Abstract socket namespace
2243 -------------------------
2244 UNIX domain sockets are a traditional way to communicate between
2245 processes on the same machine. They are always reliable (see above)
2246 and don't reorder datagrams. Unlike TCP and UDP, UNIX domain sockets
2247 support passing open file descriptors or process credentials to
2248 other processes.
2249
2250 The usual way to set up a UNIX domain socket (as obtained from
2251 socket(2)) for listening is to first bind the socket to a file system
2252 pathname and then call listen(2), then accept(2). Such sockets are
2253 called _pathname sockets_ because bind(2) creates a special socket
2254 file at the specified path. Pathname sockets allow unrelated processes
2255 to communicate with the listening process by binding to the same path
2256 and calling connect(2).
2257
2258 There are two problems with pathname sockets:
2259
2260 * The listing process must be able to (safely) create the
2261 socket special in a directory which is also accessible to
2262 the connecting process.
2263
2264 * After an unclean shutdown of the listening process, a stale
2265 socket special may reside on the file system.
2266
2267 The abstract socket namespace is a non-portable Linux feature which
2268 avoids these problems. Abstract sockets are still bound to a name,
2269 but the name has no connection with file system pathnames.
2270
2271 License
2272 -------
2273
2274 Paraslash is licensed under the GPL, version 2. Most of the code
2275 base has been written from scratch, and those parts are GPL V2
2276 throughout. Notable exceptions are FEC and the WMA decoder. See the
2277 corresponding source files for licencing details for these parts. Some
2278 code sniplets of several other third party software packages have
2279 been incorporated into the paraslash sources, for example log message
2280 coloring was taken from the git sources. These third party software
2281 packages are all published under the GPL or some other license
2282 compatible to the GPL.
2283
2284 Acknowledgements
2285 ----------------
2286
2287 Many thanks to Gerrit Renker who read an early draft of this manual
2288 and contributed significant improvements.
2289
2290 ==========
2291 References
2292 ==========
2293
2294 Articles
2295 --------
2296 - [Polynomial Codes over Certain Finite
2297 Fields](http://kom.aau.dk/~heb/kurser/NOTER/KOFA01.PDF) by Reed, Irving
2298 S.; Solomon, Gustave (1960), Journal of the Society for Industrial
2299 and Applied Mathematics (SIAM) 8 (2): 300-304, doi:10.1137/0108018)
2300
2301 RFCs
2302 ----
2303
2304 - [RFC 768](http://www.ietf.org/rfc/rfc768.txt) (1980): User Datagram
2305 Protocol
2306
2307 - [RFC 791](http://www.ietf.org/rfc/rfc791.txt) (1981): Internet
2308 Protocol
2309
2310 - [RFC 2437](http://www.ietf.org/rfc/rfc2437.txt) (1998): RSA
2311 Cryptography Specifications
2312
2313 - [RFC 4340](http://www.ietf.org/rfc/rfc4340.txt) (2006): Datagram
2314 Congestion Control Protocol (DCCP)
2315
2316 - [RFC 4341](http://www.ietf.org/rfc/rfc4341.txt) (2006): Congestion
2317 Control ID 2: TCP-like Congestion Control
2318
2319 - [RFC 4342](http://www.ietf.org/rfc/rfc4342.txt) (2006): Congestion
2320 Control ID 3: TCP-Friendly Rate Control (TFRC)
2321
2322 - [RFC 6716](http://www.ietf.org/rfc/rfc6716.txt) (2012): Definition
2323 of the Opus Audio Codec
2324
2325 Application web pages
2326 ---------------------
2327
2328 - [paraslash](http://people.tuebingen.mpg.de/maan/paraslash/)
2329 - [xmms](https://xmms2.org/wiki/Main_Page)
2330 - [mpg123](http://www.mpg123.de/)
2331 - [gstreamer](https://gstreamer.freedesktop.org/)
2332 - [icecast](http://www.icecast.org/)
2333 - [Audio Compress](https://beesbuzz.biz/code/audiocompress.php)
2334
2335 External documentation
2336 ----------------------
2337
2338 - [The mathematics of
2339 Raid6](https://www.kernel.org/pub/linux/kernel/people/hpa/raid6.pdf)
2340 by H. Peter Anvin
2341
2342 - [Effective Erasure Codes for reliable Computer Communication
2343 Protocols](http://info.iet.unipi.it/~luigi/fec_ccr.ps.gz) by Luigi
2344 Rizzo
2345
2346 Code
2347 ----
2348 - [Original FEC
2349 implementation](http://info.iet.unipi.it/~luigi/vdm.tar.gz) by
2350 Luigi Rizzo)