Andre Noll [Mon, 31 Aug 2009 19:17:21 +0000 (21:17 +0200)]
mp3dec: Improve bad main_data_begin pointer error handling.
These errors from mad_frame_decode() are non-fatal and happen if the
stream is started at the middle of the file, e.g. when para_audiod
is started while para_server is already streaming.
If libmad encounters such an error it throws away the first (and
probably the second) frame which messes up the timing in udp/fec mode,
causing an audible buffer underrun after the remaining frames of the
first fec group have been decoded and fed to the writer.
This patch makes the mp3dec filter keep track of bad main_data_begin
pointer errors that happen at the start of the stream. In this case
decoding is deferred until more data has arrived or 60ms have passed.
Andre Noll [Sat, 29 Aug 2009 20:17:55 +0000 (22:17 +0200)]
Alsa timing improvements.
This moves the computation of the select timeout from
alsa_write_post_select() to alsa_write_pre_select(). The code now
computes when the next buffer underrun would occur and uses that
value to set the timeout for the next select call. This decreases
the number of writes to the alsa handle and therefore also the CPU
usage of para_write/para_audiod.
Andre Noll [Tue, 21 Jul 2009 15:56:02 +0000 (17:56 +0200)]
Rename LIST_HEAD to INITIALIZED_LIST_HEAD.
This fixes the following warning on NetBSD:
In file included from udp_send.c:25:
list.h:37:1: warning: "LIST_HEAD" redefined
In file included from /usr/include/net/if.h:84,
from udp_send.c:14:
/usr/include/sys/queue.h:88:1: warning: this is the location of the previous definition
Andre Noll [Sun, 19 Jul 2009 15:56:22 +0000 (17:56 +0200)]
Micro optimizations for the amp filter.
Use memcpy in the special case amp==0 (no amplification) and optimize the code in
the performance-critical loop. Intrestingly, using the likely()/unlikely() macros made the
code slower.
Results (three runs on identical input data on a 32bit x86 machine under Linux, gcc-4.4.0):
old with --amp 3:
0m0.776s 0m0.790s 0m0.812s, avg: 792
new with --amp 3:
0m0.456s 0m0.492s 0m0.477s, avg: 475
speedup: 1.67
old with --amp 0:
0m0.791s 0m0.808s 0m0.810s, avg: 803
new with --amp 0:
0m0.100s 0m0.103s 0m0.094s, avg: 99
speedup: 8.1
Andre Noll [Sun, 19 Jul 2009 13:07:28 +0000 (15:07 +0200)]
Move functions only needed by audiod from stat.c to audiod_command.c.
This was 4 out of 5 and thus should decrease the size of executables for non-Linux
systems a bit as we don't use -fdata-sections -ffunction-sections and -Wl,--gc-sections
there.
Is also allows to remove some entries from para.h.
Andre Noll [Sun, 19 Jul 2009 01:10:34 +0000 (03:10 +0200)]
client_post_select(): Defer decrypting of server challenge.
The old code worked only by pure luck: We allocated the crypt buffer on the stack as
an ordinary automatic variable. This buffer was filled right after we received the challenge
from the server, but it was used in a _subsequent_ call to client_post_select(). There's
no guarantee that the content of the crypt buffer stays the same between these two calls.
So defer the decryption until the client status is CL_RECEIVED_CHALLENGE, i.e. until
the socket file descriptor is known to be ready for sending back the SHA1 of the decrypted
challenge.
Andre Noll [Sat, 11 Jul 2009 19:01:27 +0000 (21:01 +0200)]
Fix com_init() in case arguments are given.
As the ->name field of struct afs_table was only initialized in the afs process,
com_init(), which gets forked from the server process, did not see the table
names and would segfault due to a NULL pointer dereference if table names
were given as arguments.
Fix it by initializing the ->name fields in the definition of the afs_tables array.
This bug was introduced in commit 53d503ce back in 2007...
Andre Noll [Sat, 11 Jul 2009 15:13:04 +0000 (17:13 +0200)]
Scoring performance enhancements.
The old scoring code computed the afsi, afhi, and the path of the audio file for each
item of the mood, which is unnecessary. This patch moves these computations from
get_item_score() into compute_mood_score() so that afsi, afhi and path are only
computed once per audio file.
As a result, get_item_score() can no longer fail, so we may skip the checks for negative
return values.
Andre Noll [Sat, 11 Jul 2009 13:55:04 +0000 (15:55 +0200)]
Move mood methods to a separate file.
As more mood methods will be added in subsequent patches, it's good to separate
moods from mood methods.
This also removes the played_rarely mood method, since it never really worked: Whether
or not a file was played rarely depends on the afs statistics which have not been
computed at the time the mood scrore function is executed.
Andre Noll [Thu, 9 Jul 2009 18:11:55 +0000 (20:11 +0200)]
Add the new "year" mood method.
This includes a Y2K heuristic for year tags that contain only two digits. If that two-digit
number N is less or equal than the current year minus 2000, assume that it means 2000 + N.
Otherwise, assume 1900 + N.
Andre Noll [Wed, 8 Jul 2009 21:10:49 +0000 (23:10 +0200)]
Introduce para_regcomp.
A wrapper for regcomp() that logs an error message if the regcomp() failed
and uses a return value according to the paraslash rules. Currently there is
only one user of regular expressions, is_v4_dot_quad(), which is converted
to para_regex(), but new callers will be added soon.
Unfortunately, this change made it necessary to include regex.h in all .c files
that use string.h. Clean up the order in which headers are included a bit while
we're at it.
Maybe I should rethink the rule "Only .c files shall include header files"...
Andre Noll [Wed, 8 Jul 2009 19:59:39 +0000 (21:59 +0200)]
Pass full argument list to mood parsers.
Currently, for each mood line only the first word after the mood method is passed
to the mood parser. Passing all arguments allows for mood parsers that take more
than one argument.l
Andre Noll [Sun, 5 Jul 2009 16:11:32 +0000 (18:11 +0200)]
Replace split_args() by create_argv().
The latter function is superior as it honors quotes and special characters like '\n'.
This allows for example to use whitespace characters in mood methods.
Fixup and simplify all callers accordingly.
The patch broke the grab command of para_audiod which is deactivated ATM and
will be fixed in a subsequent patch.
Andre Noll [Sun, 5 Jul 2009 08:40:55 +0000 (10:40 +0200)]
Fix a bug in the output of para_audioc.
This bites In case of short writes to stdout. The bug was literally there since day one
but never showed up as short writes usually don't happen with the rather small amount
of data that is written by para_audioc.
Andre Noll [Sun, 5 Jul 2009 08:53:06 +0000 (10:53 +0200)]
gui: Avoid unnecessary strdup's and replace newlines by spaces in curses output.
align_str() made a copy of the given buffer because one of its callers, print_in_bar(),
passes a string literal and align_str() needs to truncate large strings.However, all other
callers pass a writable string anyway, so do the strdup() in print_in_bar(). This avoids the
unnecessary copies for each status item.
The areas for the status items was designed to hold only a part of a single line, so status
item text should not contain newlines. The patch replaces newlines by spaces.
Andre Noll [Sun, 5 Jul 2009 08:44:23 +0000 (10:44 +0200)]
gui: Use variable-sized input buffer for the audiod pipe.
The previous fixed size of 8192 might be too small if the audio file contains large
tags. OTOH, 8192 is much too large for the bulk of all audio files. So use a small
buffer that gets increased on demand.
Andre Noll [Sun, 5 Jul 2009 08:40:55 +0000 (10:40 +0200)]
Fix a bug in the output of para_audioc.
This bites In case of short writes to stdout. The bug was literally there since day one
but never showed up as short writes usually don't happen with the rather small amount
of data that is written by para_audioc.
Andre Noll [Sat, 4 Jul 2009 19:45:48 +0000 (21:45 +0200)]
Add a check for libosl to configure.ac.
If libosl is not found, configure will exit with an error message describing how
to download and install libosl. We could make it optional and build only the
client side of paraslash if libosl is not available, but that seems not be worth the
trouble as everyone likely wants to build para_server anyway.