]>
git.tue.mpg.de Git - paraslash.git/log
summary |
shortlog | log |
commit |
commitdiff |
tree
first ⋅ prev ⋅ next
Andre [Mon, 13 Mar 2006 04:53:07 +0000 (05:53 +0100)]
update FEATURES
Mention the new dccp sender.
Andre [Mon, 13 Mar 2006 02:49:05 +0000 (03:49 +0100)]
update to libortp-0.9.1
glib was dropped from libortp, and paraslash does not need it
either. So kill the check for glib in configure.ac and no longer
link para_server, para_recv, and para_audiod gainst glib. Moreover,
change guint32 to uint32_t in ortp_recv.c and ortp_send.c.
Andre [Mon, 13 Mar 2006 00:50:16 +0000 (01:50 +0100)]
make install: do not create an unused directory
$(VARDIR)/fifo is unused for ages.
Andre [Mon, 13 Mar 2006 00:43:58 +0000 (01:43 +0100)]
update NEWS
Andre [Mon, 13 Mar 2006 00:39:41 +0000 (01:39 +0100)]
Fix help text of --list_receivers.
It still used the old syntax with colons as separators.
Andre [Mon, 13 Mar 2006 00:35:25 +0000 (01:35 +0100)]
speed up the compress filter
Complete rewrite of the core loop of the compress filter. The gprof
output of para_audiod (see below) indicates that CPU usage goes down
to about 40%.
new:
~~~~
Each sample counts as 0.01 seconds.
% cumulative self self total
time seconds seconds calls s/call s/call name
35.94 1.47 1.47 7414 0.00 0.00 compress
22.74 2.40 0.93 57319 0.00 0.00 mp3dec
16.87 3.09 0.69 1 0.69 4.04 audiod_mainloop
34.59 1.47 1.47 7414 0.00 0.00 compress
22.59 2.43 0.96 1 0.96 4.20 audiod_mainloop
17.53 3.17 0.74 61285 0.00 0.00 mp3dec
35.41 1.48 1.48 7414 0.00 0.00 compress
20.10 2.32 0.84 1 0.84 4.15 audiod_mainloop
17.82 3.06 0.74 58675 0.00 0.00 mp3dec
Average of self-seconds (3rd column):
compress 1.47
audiod_mainloop 0.83
mp3dec 0.80
0.2.11:
~~~~~~~
Each sample counts as 0.01 seconds.
% cumulative self self total
time seconds seconds calls s/call s/call name
56.38 3.31 3.31 7414 0.00 0.00 compress
12.93 4.08 0.76 57566 0.00 0.00 mp3dec
11.39 4.75 0.67 1 0.67 5.86 audiod_mainloop
57.59 3.38 3.38 7414 0.00 0.00 compress
14.08 4.20 0.82 63095 0.00 0.00 mp3dec
11.26 4.86 0.66 1 0.66 5.82 audiod_mainloop
57.79 4.12 4.12 7414 0.00 0.00 compress
13.62 5.08 0.97 1 0.97 7.07 audiod_mainloop
8.08 5.66 0.57 57196 0.00 0.00 mp3dec
Average of self-seconds (3rd column):
compress 3.60
audiod_mainloop 0.76
mp3dec 0.71
Andre [Sat, 11 Mar 2006 16:45:20 +0000 (17:45 +0100)]
Merge branch 'dccp'
Fix up merge conflicts by hand.
Conflicts:
configure.ac
error.h
server.ggo
Andre [Sat, 11 Mar 2006 15:29:29 +0000 (16:29 +0100)]
new codename, reset version to git
Andre [Sat, 11 Mar 2006 15:26:38 +0000 (16:26 +0100)]
the paraslash-0.2.11 release tarball
Andre [Sat, 11 Mar 2006 15:23:15 +0000 (16:23 +0100)]
paraslash 0.2.11
Andre [Sat, 11 Mar 2006 03:40:55 +0000 (04:40 +0100)]
whitespace cleanup
Andre [Fri, 10 Mar 2006 22:44:15 +0000 (23:44 +0100)]
update NEWS
Andre [Fri, 10 Mar 2006 22:01:33 +0000 (23:01 +0100)]
deleted: index.html
It crept in during the initial commit, and it is long outdated.
Andre [Fri, 10 Mar 2006 21:54:36 +0000 (22:54 +0100)]
mysql_selector.c: rename get_dbinfo() to get_selector_info()
Andre [Fri, 10 Mar 2006 21:49:33 +0000 (22:49 +0100)]
mysql_selector.c: rename refresh_mmd_dbinfo() to refresh_selector_info()
Andre [Fri, 10 Mar 2006 21:47:16 +0000 (22:47 +0100)]
kill a macro which is only used once.
Insert the text directly instead.
Andre [Fri, 10 Mar 2006 21:41:50 +0000 (22:41 +0100)]
mmd: rename dbt_change to selector_change
Andre [Fri, 10 Mar 2006 21:32:33 +0000 (22:32 +0100)]
mmd: rename dbt_num to selector_num
Andre [Fri, 10 Mar 2006 21:28:01 +0000 (22:28 +0100)]
mmd: rename dbinfo to selector_info
Andre [Fri, 10 Mar 2006 21:20:31 +0000 (22:20 +0100)]
Some more (trivial) "database tool" -> "audio file selector" conversions
Andre [Fri, 10 Mar 2006 21:00:57 +0000 (22:00 +0100)]
dbtool -> selector update for web documentation web page
Andre [Fri, 10 Mar 2006 20:57:25 +0000 (21:57 +0100)]
More dbtool -> selector updates
Andre [Fri, 10 Mar 2006 20:53:05 +0000 (21:53 +0100)]
rename mysql.c to mysql_selector.c
Just to be consistent.
Andre [Fri, 10 Mar 2006 20:48:25 +0000 (21:48 +0100)]
README.mysql: cdt -> chs update
Andre [Fri, 10 Mar 2006 20:13:49 +0000 (21:13 +0100)]
Rename E_BAD_DBTOOL to E_BAD_SELECTOR
Andre [Fri, 10 Mar 2006 20:11:36 +0000 (21:11 +0100)]
Rename random_dbtool.c to random_selector.c
Andre [Fri, 10 Mar 2006 20:01:50 +0000 (21:01 +0100)]
rename plm_dbtool.c to playlist_selector.c
Andre [Fri, 10 Mar 2006 19:41:25 +0000 (20:41 +0100)]
Rename "dbtool" status item to "selector"
Andre [Fri, 10 Mar 2006 19:35:11 +0000 (20:35 +0100)]
Change doxygen file description from dbtool to selector
Andre [Fri, 10 Mar 2006 19:30:43 +0000 (20:30 +0100)]
Another trivial rename
Andre [Fri, 10 Mar 2006 19:29:03 +0000 (20:29 +0100)]
gui,sdl_gui: change the prefix from "dbtool" to "selector"
Andre [Fri, 10 Mar 2006 19:24:52 +0000 (20:24 +0100)]
trivial dbtool->selector renames
Andre [Fri, 10 Mar 2006 19:18:08 +0000 (20:18 +0100)]
Rename some local variables in command.c
Andre [Fri, 10 Mar 2006 19:14:46 +0000 (20:14 +0100)]
rename struct dbtool to audio_file_selector
Andre [Fri, 10 Mar 2006 19:04:25 +0000 (20:04 +0100)]
db.h: Kill an unused enum
Andre [Fri, 10 Mar 2006 18:57:15 +0000 (19:57 +0100)]
Rename com_cdt to com_chs
chs: change (audio file) selector.
Andre [Fri, 10 Mar 2006 18:49:49 +0000 (19:49 +0100)]
dbtool -> audio file selector renaming
This updates only Documentation and log messages.
Andre [Fri, 10 Mar 2006 17:58:17 +0000 (18:58 +0100)]
Rename the --dbtool option to --selector
As two out of the three "database tools" of paraslash don't use a
database at all, let's call them "audio file selectors" instead.
Andre [Fri, 10 Mar 2006 02:12:23 +0000 (03:12 +0100)]
ogg.c: don't use ogg_int64_t
ssize_t should do equally well and avoids gcc warnings.
Andre [Fri, 10 Mar 2006 01:47:26 +0000 (02:47 +0100)]
whitespace fixes
Andre [Fri, 10 Mar 2006 01:32:12 +0000 (02:32 +0100)]
audiod: fix a memory leak in ringbuffer handling
Found by valgrind.
Andre [Thu, 9 Mar 2006 15:47:31 +0000 (16:47 +0100)]
web/header2.html: update the changes menu item to point to gitweb.
From within the doxygen API reference that menu item still pointed
to the old cvs changelog which is no longer present.
Andre [Thu, 9 Mar 2006 15:36:56 +0000 (16:36 +0100)]
Merge branch 'no_colon_separators'
Andre [Thu, 9 Mar 2006 15:34:51 +0000 (16:34 +0100)]
Convert demo-script to the new syntax without colons.
As this won't work with paraslash-0.2.10, change the script
to download the git tarball.
Andre [Thu, 9 Mar 2006 15:28:18 +0000 (16:28 +0100)]
Add a missing header file to Makefile.in
ipc.h was not included in the tarball. With this fix, the nightly git
snapshot compiles again.
Andre [Wed, 8 Mar 2006 23:52:03 +0000 (00:52 +0100)]
maintainer-clean: also remove skencil/*.ps skencil/*.pdf and doc
Andre [Wed, 8 Mar 2006 23:40:44 +0000 (00:40 +0100)]
header2.html is needed by doxygen
Andre [Wed, 8 Mar 2006 23:17:35 +0000 (00:17 +0100)]
Add Ian McDonald to CREDITS
Also, add a copyright notice at the top of dccp.c dccp.h dccp_recv.c and dccp_send.c
Andre [Wed, 8 Mar 2006 23:08:01 +0000 (00:08 +0100)]
Add GPL headers for dccp.c dccp.h, dccp_recv.c, and dccp_send.c
Andre [Wed, 8 Mar 2006 23:04:15 +0000 (00:04 +0100)]
First draft of the dccp sender/receiver pair.
Andre [Wed, 8 Mar 2006 17:45:01 +0000 (18:45 +0100)]
kill a pointless "optind = 1".
Andre [Tue, 7 Mar 2006 13:20:09 +0000 (14:20 +0100)]
Merge branch 'xf_64_compile_warning_fixes'
Andre [Tue, 7 Mar 2006 00:47:28 +0000 (01:47 +0100)]
README: Mention the name of some debian packages
libasound2-dev is needed for para_play, ogg vorbis support requires
libogg-dev and libvorbis-dev. Thanks to Steffen Klassert for this info.
Andre [Tue, 7 Mar 2006 00:13:12 +0000 (01:13 +0100)]
ogg.c: use %z modifier
Andre [Tue, 7 Mar 2006 00:09:39 +0000 (01:09 +0100)]
oggdec.c: %d -> %zd fixes
Andre [Mon, 6 Mar 2006 20:57:27 +0000 (21:57 +0100)]
another %d -> %zd fix
Noticed by Steffen Klassert
Andre [Mon, 6 Mar 2006 20:51:19 +0000 (21:51 +0100)]
net.h: use proper type for the addrlen parameter of para_accept() also in prototype
An earlier patch changed that parameter from size_t to socklen_t
but missed to update also the corresponding prototype in net.h. Thanks to
Steffen Klassert for pointing this out.
Andre [Sun, 5 Mar 2006 13:18:24 +0000 (14:18 +0100)]
Simplify for_each_line()
It is always called with the num parameter being zero. So nuke that
parameter and update the docs.
Andre [Sun, 5 Mar 2006 12:59:39 +0000 (13:59 +0100)]
Kill an outdated comment..
..which stated that filter command line options should be separated by colons.
Andre [Sun, 5 Mar 2006 12:49:11 +0000 (13:49 +0100)]
para_filter: Update --help
The help text still used the old syntax with colons as separators. Fix that.
Andre [Sun, 5 Mar 2006 12:40:43 +0000 (13:40 +0100)]
para_filter: Fix and improve help text for --list_filters
The old help text still used colons as separators. Mark those filters
which have their own command line parser with a star in the filter
list. Some trivial whitespace fixes also.
Andre [Sun, 5 Mar 2006 12:14:09 +0000 (13:14 +0100)]
use %zd instead of %d for size_t
As of version 2.1 of glibc there is the %z modifier, so use it.
This avoids many warnings of the type
warning: int format, different type arg
on x86_64.
Andre [Sun, 5 Mar 2006 12:08:35 +0000 (13:08 +0100)]
para_accept(): Use socklen_t instead of size_t
Fixes the following warning on x86_64:
net.c:76: warning: int format, different type arg (arg 4)
Andre [Thu, 2 Mar 2006 21:33:14 +0000 (22:33 +0100)]
(no log message)
Andre [Thu, 2 Mar 2006 17:10:11 +0000 (18:10 +0100)]
kill another superfluous '#include "gcc-compat.h"'
Commit
4b4c68de8d7390f966e46d9402d5499d2e8ee227 missed one of these.
Andre [Thu, 2 Mar 2006 14:08:38 +0000 (15:08 +0100)]
fix off-by-one error
Also, make "bufsize" const instead of static.
Andre [Mon, 27 Feb 2006 09:19:39 +0000 (10:19 +0100)]
no need to include gcc-compat.h twice
It gets already included from para.h.
Andre [Sun, 26 Feb 2006 17:03:54 +0000 (18:03 +0100)]
make header_len an integer
This fixes the following warning:
ortp_send.c:169: warning: pointer targets in passing argument 1 of 'af->get_header_info' differ in signedness
Andre [Sun, 26 Feb 2006 16:17:05 +0000 (17:17 +0100)]
simplify open_audiod_pipe(), remove a trailing newline which caused
the top window to flicker if para_audiod is not running.
Andre [Sat, 25 Feb 2006 09:16:22 +0000 (10:16 +0100)]
Merge branch 'no_colon_separators'
Andre [Sat, 25 Feb 2006 09:15:54 +0000 (10:15 +0100)]
Merge branch 'para_play_fixes'
Andre [Fri, 24 Feb 2006 15:13:57 +0000 (16:13 +0100)]
Use whitespace instead of colons to separate receiver/filter args
This is much nicer but the receiver/filter argument must now be quoted if any
options are present.
Andre [Fri, 24 Feb 2006 13:09:42 +0000 (14:09 +0100)]
another typo in comment
ipc.c:44: Warning: expected whitespace after param command
Andre [Fri, 24 Feb 2006 11:21:33 +0000 (12:21 +0100)]
configurable input buffer size and prebuffering
Andre [Thu, 23 Feb 2006 17:34:59 +0000 (18:34 +0100)]
play.c: pcm_write() aborts on errors. No need to check the return value.
Andre [Thu, 23 Feb 2006 17:29:45 +0000 (18:29 +0100)]
Fix --device bug
Duuuh, the --device option never worked because "plug:swmix" was harcoded
in play.c instead of using the value from the configuration. Even worse,
the default in play.ggo was _also_ set to "plug:swmix". Replace that by
"plughw:0,0".
This patch also simplifies and optimizes play_pcm().
Anyway, software mixing is still recommended. Here's an aprropriate
/etc/asound.conf:
---------------------------------------------------- /etc/asound.conf
pcm.swmix {
type dmix
# any unique number here
ipc_key 313
ipc_perm 0666
slave {
pcm "hw:0,0"
# these settings may require tweaking for different sound
# cards; this is for the Powerbook's built-in snd-powermac
# probably not required at all for well-behaved cards...
period_time 0
period_size 1024
buffer_size 8192
# mentioning rate fixes wrong speed/pitch in native ALSA stuff
# rate 44100
}
}
pcm.dsp0 {
type plug
slave.pcm "swmix"
}
ctl.mixer0 {
type hw
card 0
}
pcm.!default {
type plug
slave.pcm "swmix"
}
---------------------------------------------------------------------
a Use the
: Use a large input
buffer and For one, the
Andre [Wed, 22 Feb 2006 12:06:14 +0000 (13:06 +0100)]
fix another typo in comment
Andre [Wed, 22 Feb 2006 11:48:07 +0000 (12:48 +0100)]
announce gitweb
Andre [Wed, 22 Feb 2006 11:21:09 +0000 (12:21 +0100)]
add more ipc documentation
Also, add a GPL header and replace "shared memory area" by "shared memory segment".
Andre [Wed, 22 Feb 2006 07:39:17 +0000 (08:39 +0100)]
typo
Andre [Wed, 22 Feb 2006 07:07:15 +0000 (08:07 +0100)]
replace link to the changelog by link to gitweb
Andre [Tue, 21 Feb 2006 23:25:48 +0000 (00:25 +0100)]
add documentation of struct plm_client_data
Andre [Tue, 21 Feb 2006 21:25:32 +0000 (22:25 +0100)]
replace cvs checkout instructions by git clone command
Andre [Tue, 21 Feb 2006 18:30:30 +0000 (19:30 +0100)]
for those of us who are interested in history..
Andre [Tue, 21 Feb 2006 18:22:40 +0000 (19:22 +0100)]
Typo
Andre [Tue, 21 Feb 2006 18:07:37 +0000 (19:07 +0100)]
replace link to cvs snapshot by git snapshot
Andre [Tue, 21 Feb 2006 17:38:02 +0000 (18:38 +0100)]
cvs->git update
Andre [Tue, 21 Feb 2006 17:04:43 +0000 (18:04 +0100)]
Avoid busy loop if someone nasty removes the semaphores currently in use.
Just die if semop failed 500 times.
Andre [Tue, 21 Feb 2006 16:52:23 +0000 (17:52 +0100)]
change version.
The shiny new git tree was still called "cvs".
Andre [Tue, 21 Feb 2006 16:48:54 +0000 (17:48 +0100)]
mention plm and update the mysql description.
Also, selector is now the prefered word, rather than dbtool.
Andre [Tue, 21 Feb 2006 16:07:56 +0000 (17:07 +0100)]
use the random dbtool as default
That simplifies the code quite a bit.
Andre [Tue, 21 Feb 2006 15:42:12 +0000 (16:42 +0100)]
make para_server use the new ipc subsystem
Straight-forward conversion.
This BTW gets rid of the ancient /dev/zero hack which worked pretty
well for all the years but is not general enough for the plm dbtool.
Another 42 also disappeared ;)
Andre [Tue, 21 Feb 2006 14:52:45 +0000 (15:52 +0100)]
fix the plm database tool
(Re)loading the playlist uses the new mutex and
shm helpers and is hopefully a race-free implementation.
It works as follows:
The plm init function (parent, server context) reserves
a shared memory area (client data) as well as two unlocked
mutexes. The first mutex serializes access for racing clients
that want to replace the playlist. The second mutex serializes
the access between parent and child (com_lpl()).
com_lpl() loads the playlist from the client into a tmp buffer.
It then creates a shm area (client_shm) and copies that buffer
to the area. Further, another mutex (shm) is created. This
mutex is initially locked.
Next, it grabs the client lock and the server lock, and updates
client_data with the id of client_shm. Then the parent is
woken up (SIGUSR1) and the server lock is released. com_lpl()
finally goes to sleep by acquiring the shm lock.
In this situation, only the parent may run. It grabs the
server lock, reads client_data to attach client_shm and
reloads the playlist with data from client_shm. When ready,
the parent detaches client_shm, and drops the shm lock and
the server lock.
This wakes up com_lpl() which destroys client_shm, the
shm mutex and the tmp buffer.
Andre [Tue, 21 Feb 2006 08:22:36 +0000 (09:22 +0100)]
Merge branch 'dbtool_preselect'
Andre [Tue, 21 Feb 2006 08:20:30 +0000 (09:20 +0100)]
Merge branch 'ipc'
Conflicts:
configure.ac
error.h
Fix these conflicts.
Andre [Tue, 21 Feb 2006 08:09:36 +0000 (09:09 +0100)]
add dbtool hooks
pre_select/post_select. Also, add a private_dbtool pointer for
each audio format to mmd.
Andre [Tue, 21 Feb 2006 07:12:46 +0000 (08:12 +0100)]
The new ipc subsystem
Contains functions for dealing with mutexes and shared memory.
Andre [Tue, 21 Feb 2006 06:06:17 +0000 (07:06 +0100)]
the new plm database tool
does not work yet as the ipc problems are not solved.
Andre [Mon, 20 Feb 2006 06:25:59 +0000 (07:25 +0100)]
rename dopey to random
It isn't so dopey anymore and the upcoming plm database tool
won't be much smarter, so..
Andre [Mon, 20 Feb 2006 06:03:36 +0000 (07:03 +0100)]
initial git commit
Let's try if this works out.