53ad2d187f
Move processing thread list from ARDOUR::Graph into AudioBackend implementation
2013-10-04 18:53:37 +10:00
df363a4fb3
Add AudioBackendThread class to support different thread type on windows
2013-10-04 18:51:47 +10:00
82f0f3a9a7
fix up crash caused by changes in object creation order and setting the state of the click IO object
2013-10-02 16:17:22 -04:00
c736d88a5f
don't ping JACK server 4 times to see if it is already up when ardour starts
2013-09-25 22:56:28 -04:00
d6274d5c3e
fix up enum mess caused by switching from JACK to Engine as the name for jack-related transport sync
...
This will allow older versions of Ardour to continue to startup when the user sync preference is "JACK",
because the string used by enum_2_string() will still be "JACK".
Versions of ardour3 from git after the enum change until this commit will leave ardour.rc unloaded by old
versions of Ardour ***if*** the user sync choice was "JACK".
2013-09-25 18:24:32 -04:00
e43a8dac1c
make information extraction from session file options list robust against some old badly formatted session files
...
Fixes crash experienced by GillesM, where a rogue XML entry caused a segfault
2013-09-25 17:59:10 -04:00
f5cd838afc
start explicitly linking against libtimecode (this needs to happen elsewhere too)
2013-09-24 22:22:43 -04:00
4209e8a834
add new static functions to get SR and disk sample format from session XML
2013-09-23 21:35:51 -04:00
20f28c82ec
make session creation from a template work again
2013-09-22 11:19:34 -04:00
f7586f3a0d
likely fix for creating new sessions from templates
2013-09-21 12:50:45 -04:00
3e5c1e3fc1
Revert "actually load new state after creating a session."
...
This reverts commit 914ba475de
2013-09-20 16:16:30 -04:00
eb3cc04eff
Merge branch 'master' of https://github.com/nmains/ardour
2013-09-20 15:21:14 -04:00
b188033bf7
rather hacky fix for the change of the SyncSource > JACK enum to SyncSource > Engine
...
This is needed in order to get ardour started if ardour.rc names "JACK" as the sync source
2013-09-20 14:30:50 -04:00
960a47330a
switch from JACK_Slave to Engine_Slave
2013-09-19 17:34:23 -04:00
57d24608e7
show a different audio/MIDI option screen if JACK is already running. try to fix problem with using AudioBackend::<property-value>() functions before anything has been set.
2013-09-19 15:03:20 -04:00
f51537b71d
substantive changes to allow the audio/MIDI setup dialog to change settings as intended
...
So far only tested with JACK backend, but expected to be more correct for other backends too
2013-09-17 21:09:13 -04:00
nick_m
914ba475de
actually load new state after creating a session.
2013-09-18 02:02:31 +10:00
32452bb3b3
fix crash caused by trying to unregister null ports
2013-09-17 10:01:02 -04:00
710ee36eaa
a few DEBUG_TRACE improvements
2013-09-16 22:02:36 -04:00
39ee88b3f9
more session construction order changes, and removal of n_physical_{inputs,outputs} members which were (a) not initialized early enough (b) not used anywhere except monitor bus connection.
...
Things almost make sense now.
2013-09-16 12:08:19 -04:00
910755b51b
fix missing initialization of two Session members, which fixes a crash during quit-without-finished-session
2013-09-16 10:23:37 -04:00
7cc451d81d
fix language issue
2013-09-16 10:07:52 -04:00
7c8f3adc75
allow waf to handle inconsistent case where cwiid.h is present but not the pkg-config file
2013-09-16 10:07:27 -04:00
31157cb343
fix ordering in session construction so that new sessions and existing ones both work
2013-09-16 09:57:22 -04:00
bde50d5396
really fix build issues after removal of "using" directives
2013-09-14 16:34:08 -04:00
25b9e266eb
fix build issues after removal of "using" directives
2013-09-14 16:24:21 -04:00
f5191e6257
cerr output when JACK halt callback is executed
2013-09-14 16:18:17 -04:00
76a08ad9e9
fix return value from thread creation function when not connected to JACK
2013-09-14 16:17:49 -04:00
6b23417ae0
move creation of Graph till after we have the engine running, since we need access to the backend
2013-09-14 16:17:24 -04:00
dce6b71e82
use visibility control on JACK audio backend, so that we can import this change into windows branch
2013-09-13 12:32:50 -04:00
83a826095d
update JACK backend to use new inheritance structure for AudioBackend
2013-09-13 11:21:43 -04:00
f9d08daa47
change inheritance so that an AudioBackend IS-A PortEngine
...
This allows a derived (concrete) implementation to share information (e.g. sample rate, buffer size)
between the audio backend side of things and the port management side of things.
2013-09-13 11:21:15 -04:00
b085a79938
explanatory comment
2013-09-13 10:19:01 -04:00
ccc81f05be
remove some debug output messages
2013-09-12 17:25:01 -04:00
28d692b490
don't stop metering thread when session is removed; move engine-setup code into its own method. sorry, ardour build-from-source folk :)
2013-09-12 17:25:01 -04:00
7c88670680
lots of changes to auto-start (and stop) the backend for latency measurements, and continuing work on the session construction/engine configuration flow
2013-09-12 16:28:51 -04:00
9af75b2ac9
reinstate unblocking-of-connections in session setup, accidentally dropped in reorganization work
2013-09-12 16:27:58 -04:00
6b3907d57f
change names, add comment, improve return type to avoid extra call to get EngineControl::State*
2013-09-12 12:26:59 -04:00
4df3666738
add AudioBackendNativeThread to serve the same role as jack_native_thread_t
2013-09-12 11:29:47 -04:00
9f2ab81df6
always add a timeout value of 200msec to jack, just like qjackctl does.
...
THIS IS A HACK. LONG TERM GOAL: understand why ardour gets zombified on the way up.
2013-09-11 21:00:21 -04:00
f85b362351
*maybe* fix issues with zombification during session loading
2013-09-11 16:33:40 -04:00
91e2eab01a
make use of measured latency values to set dialog controls, and use actual port latency to compute extra delay
2013-09-11 13:02:32 -04:00
18c4c9656a
remove some configure time tests from libs/ardour regarding JACK which are no longer relevant
2013-09-11 12:44:21 -04:00
7c3535ebf8
make quit work even when latency measurement is in progress
2013-09-11 10:06:02 -04:00
7c719e441a
fix handling of new session
...
defer save state till after everything done in post_engine_init() is complete.
2013-09-11 09:59:56 -04:00
d305dc1e0c
more tweaks to latency measurement
...
don't open ports till absolutely necessary (store names for future use).
tidy up parts of the dialog (tab)
2013-09-10 23:25:15 -04:00
676ff80697
basic functionality for hardware latency measurement
2013-09-10 22:58:33 -04:00
209e4bdcae
many changes relating to session construction and audioengine interaction
...
every session member is now initialized using C++ constructor syntax
session construction reordered to clarify the split(s) between work
where the engine is not relevant and work where is it is. this
split is still not 100% obvious, but is enormously clearer than
previously.
if engine/backend are not running as session is created, and the SR
of the sample rate is known, attempt to force backend to that value.
2013-09-10 15:41:19 -04:00
19d05fa436
start reordering of Session initialization
...
the goal is to allow a clear demarcation of where we need
the audioengine and have existing data parameters (e.g. sample rate)
for the session so that we can coerce the engine (if possible)
to match to session values.
also, to rationalize construction even more and use C++
constructor syntax as much as possible.
2013-09-09 22:40:54 -04:00
bea5434639
fix name of control app for Delta1010
2013-09-09 21:35:47 -04:00
2a89234027
move control app launching back into audio backend to allow ASIO/CoreAudio model to work; push initial state of AMS dialog to backend
...
Issues remain with the basic model of the AMS dialog - when is
newly chosen state pushed into the backend (which can then
modify the control app button sensitivity. This is a special
problem for this button because APIs like ASIO and CoreAudio
probably don't allow us to launch a control app for an arbitrary
device, but only one actually in use. In this sense it is
different from properties like available buffer size etc, where
we can typically query without actually using the device.
2013-09-09 21:23:12 -04:00
29c9d94dbe
fix control app breakage
2013-09-09 18:06:44 -04:00
f06bab1fa3
add envy24 control app name
2013-09-09 17:10:33 -04:00
ecfeeda4b8
launching control app is now responsibility of ardour GUI, not audio backend; use ARDOUR_DEVICE_CONTROL_APP if set in the environment
2013-09-09 16:48:27 -04:00
f06187735d
initial work on adding access to h/w control apps when ardour starts JACK
2013-09-09 15:17:43 -04:00
a228643e40
add can_change_{sample_rate,buffer_size}_while_running() methods to an AudioBackend
...
Allows the GUI and other stuff to know whether or not changing the SR/bufsize is possible while running, which is about to become useful
2013-09-09 15:17:43 -04:00
56465fda10
move engine dialog from a widget to an actual dialog; emit Session::AudioEngineSetupRequired when loading a session if it is necessary
2013-09-09 15:17:42 -04:00
5ec69e2374
move inadvertently moved jack_slave.cc back to its rightful spot
2013-09-07 13:39:36 -04:00
f96652e12a
other changes for separated jack backend
2013-09-07 11:13:07 -04:00
f4cf283f26
move JACK audio backend to its own folder and adjust build system to reflect that (installed version may now work)
2013-09-07 11:03:57 -04:00
ad8d83289f
add support in JACK backend for specifying number of input and output channels to use
2013-09-06 17:21:23 -04:00
0456788663
fix JACK backend setup so that we know about graph, port changes etc.
2013-09-06 13:06:44 -04:00
79fed95c1b
tighten up error handling a little bit during various setup functions for Mackie Control
2013-09-06 10:39:14 -04:00
018465f739
exit early from AsyncMIDIPort::drain() if it will never be used in a process callback again
2013-09-06 10:38:49 -04:00
94efddd240
fix a problem creating and displaying connected status for ports not owned by ardour (e.g. system:....)
...
This was caused by using jack_port_get_connections() which will
not return the correct status for ports owned by another JACK
client. Because of the potential for deadlock by calling
jack_port_get_all_connections(), an extra argument was added
to several PortEngine:: API calls to specify whether the call
is in a process-callback context, which defaults to true.
The only place where false is passed is within the GlobalPortMatrix
when we need to determine whether two non-ardour ports are
connected.
2013-09-05 21:39:43 -04:00
1c49138e00
move MidiPortManager from AudioEngine to Session
...
This makes the responsibilities and ownership of non-Route related MIDI ports
more clear, and removes a few wierd bits of code. It also ensures that open/close/open
on the same session will retain connections for those MIDI ports
2013-09-05 13:22:34 -04:00
35cc3245db
rename interface_ stuff in the engine dialog to device_, since that feels a bit clearer terminology
2013-09-04 12:37:44 -04:00
d074bc586e
Merge branch 'master' into audioengine
2013-09-03 17:41:49 -04:00
9261d5908d
fix error introduced during recent changes to Mackie control support, involving ambiguous enum usage
2013-09-03 08:48:11 -04:00
12ec604140
updated .po files from waf i18n, plus an updated gtk2_ardour/de.po from Edgar Aichinger
2013-09-03 08:00:58 -04:00
d2cdc7da64
add std:: to clarify use of llabs on OS X
2013-09-02 17:45:04 -04:00
66292718a1
improve scroll and drag behaviour of volume-controller knob
2013-09-01 03:19:13 +02:00
dd0a1cd846
fix keyboard interaction with volume knobs (monitor section)
2013-08-31 20:37:02 +02:00
14f5557bec
fix typo in IEC-meter 'weird-float' protection.
2013-08-29 18:46:03 +02:00
2172700144
protect meters against weird float values
2013-08-29 17:36:52 +02:00
03c26762e6
update external plugin UI handling
...
amend to 4cdb0181d972d0
2013-08-29 13:05:11 +02:00
4864833669
fix vari-speed phase (when used for calculation only)
...
and clarify an old comment.
2013-08-27 20:55:15 +02:00
1d972d05c5
add LV2 support for kx:external-ui#
...
the ABI of http://kxstudio.sf.net/ns/lv2ext/external-ui# is identical to
http://lv2plug.in/ns/extensions/ui#external
It just adds a distinction between #Host and #Widget for cosmetic reasons.
Sadly some newer plugins are released with only support for kxstudio URI :(
2013-08-25 16:21:18 +02:00
1421a34e52
pot/po file updates caused by waf i18n
2013-08-24 11:46:45 -04:00
41f6f5574b
fix confusion about whether flip_mode in the Mackie support code is a bool or an enum
2013-08-24 11:46:44 -04:00
Tim Blechmann
a5de2c1067
export: add support for CAF
2013-08-24 14:23:25 +02:00
3247c5a51c
undo the hiding of SR-dependent LV2 ports
2013-08-22 17:46:19 -04:00
868112bece
update czech translations from pavel fric, part 2
2013-08-15 14:55:57 -04:00
7fe7e7ad4b
update czech translations from pavel fric
2013-08-15 14:55:10 -04:00
04bf9d1e95
fix newly-appearing crash-at-close caused by muddled thinking in pbd/pthread_utils
...
threads created with this code can now just return a value as they normally would, and the
infrastructure will ensure cleanup. there is no longer any reason to call pthread_exit_pbd()
and so that has been removed.
2013-08-15 11:44:47 -04:00
588cc3af74
check master sources when determining whether a region uses a source. should fix #5618
2013-08-15 11:43:37 -04:00
6bcdf4f0f1
fix merge issues with master
2013-08-14 13:10:59 -04:00
e711dfae12
remove misleading + incorrect documentation
2013-08-14 12:44:59 -04:00
2de09888b6
some const_cast<> additions for OS X where gcc believes that volatile int* means const int*
2013-08-14 12:40:49 -04:00
8d8770f4cc
remove JACK midi system stuff from jack_utils, because we're not going to use it (for a while, at least)
2013-08-14 12:40:10 -04:00
7b34b6a3f6
provide clock_gettime() implementation for lovely OSX, which doesn't have it
2013-08-14 12:39:34 -04:00
dfac01bdd4
don't bother with old MIDI port data in ardour.rc
...
result: we lose connections to global MIDI ports in older sessions. may fix this in the future
2013-08-13 22:06:31 -04:00
07c63fb2d7
save and restore all ardour-owned MIDI ports
...
still need to check on MCU port status
2013-08-13 21:59:05 -04:00
a5a75d5e0d
fixes to get MTC (and probably MIDI clock) slaving working again
...
incoming MIDI data has to be parsed EVERY process cycle, not just when Slave::speed_and_position() is called.
The private MIDI::Parser owned by the MTC and MClck slaves was irrelevant, since the port has its own.
See comments in midi_port.h on the strangled inheritance heirarchy.
2013-08-13 12:53:28 -04:00
63745512bc
Call setup_libpbd_enums in PBD::init for portability
2013-08-13 20:15:47 +10:00
1db894f886
Call Gio::init from PBD::init instead of ARDOUR::init
...
PBD needs Gio for PBD::copy_file and perhaps others
2013-08-13 20:15:47 +10:00
55c6d4f2be
Call PBD::init in ARDOUR::init and PBD::cleanup in ARDOUR::cleanup
2013-08-13 20:15:47 +10:00
29193c76cc
Add PBD::init and PBD::cleanup
...
A bit of refactoring to move initialization of libpbd into libpbd
rather than rely on "client" code(ui,libardour,tests etc) to do it
2013-08-13 20:15:47 +10:00
9d340af10d
Prevent libardour from being initialized more than once
...
This is not thread safe of course, it is not intended to be.
This was implemented as at one stage ARDOUR::init was being called multiple
times in the testsuite which was causing a subtle bug that took some
time to track down.
2013-08-13 20:15:47 +10:00
d51eba1162
Change ARDOUR::init return type to bool type for success/failure
2013-08-13 20:15:47 +10:00
a5a3f713d5
assorted extra debug output for MTC
2013-08-12 05:40:11 -04:00
4c4376d5dd
fix thinko in 5dee49e19
2013-08-11 17:26:34 +02:00
116694a2cb
fix 5dee49e19 (32/64 bit unsigned int compatibility)
2013-08-11 17:19:04 +02:00
5dee49e194
make LV2 communication buffers independent from jack-midi buffer-size
...
fixes issues with plugin communication that are common with
jack1 use due to its very small midi-buffers.
2013-08-11 15:10:41 +02:00
50bbf3f656
fix output latency option in jack command line
2013-08-10 09:25:18 -04:00
9680c345c4
don't write a bogus jack cmdline ... leave the existing ~/.jackdrc file alone if an error creating the cmdline occurs
2013-08-09 18:35:07 -04:00
862315ec2a
actually write JACK command line out to ~/.jackdrc so that settings take effect
2013-08-09 18:11:04 -04:00
4b9289e17e
don't waste time with sending MTC or MMC after stop if we're disconnected from the engine.
...
Also, add note about threading oddities with send_mmc_locate() being called from ::non_realtime_stop() which gets
called from a *realtime* context.
2013-08-09 18:06:08 -04:00
cec2a3da36
allow ardour to exit after JACK disconnects us
2013-08-09 16:57:23 -04:00
f0fcda2044
Merge branch 'master' into audioengine
2013-08-09 16:29:23 -04:00
8685d8eb47
remove use of jack_get_time() from ARDOUR::get_microseconds(), use clock_gettime() instead since this function is never used in reference to a backend clock
2013-08-09 15:40:46 -04:00
ca76496ef3
more purging of JACK as an explicit name from libardour
2013-08-09 13:56:23 -04:00
d0381b902c
remove more naming that refers to JACK
2013-08-09 13:39:33 -04:00
55e52bcbd5
remove another naming that refers to JACK
2013-08-09 13:38:09 -04:00
8960ebcdf9
remove another naming that refers to JACK
2013-08-09 13:27:05 -04:00
521f8eb776
more sensible ALSA device names now in use
2013-08-09 13:26:40 -04:00
fc040a72e9
use original names for ARDOUR MIDI "persistent" ports (MMC, MTC, etc)
2013-08-09 13:03:57 -04:00
5abeb799cf
try new code from rui/qjackctl for names of ALSA PCM devices
2013-08-09 13:03:21 -04:00
a34d707175
split up session code that uses parts of the JACK API (timebase + session event handling) and connect it directly to the jack audiobackend
...
i've made the audiobackend call the session directly so that only one object (ARDOUR::Session) has a need for the JACK types
and only one .cc file (session_jack.cc) needs jack.h. having ARDOUR::AudioEngine act as an intermediary would be cleaner
conceptually but would end up causing two different ARDOUR objects to have jack types in their own API.
2013-08-09 12:15:37 -04:00
c9f6b0dab5
the JACK audio backend shared object depends on libasound (ALSA) on linux, so it requires explicit linkage
2013-08-09 11:04:04 -04:00
fcb76c8c05
fix freewheeling
2013-08-08 22:12:52 -04:00
3070cd723b
fix crashes, clean up session code warmup and other details in audioengine
2013-08-08 19:32:56 -04:00
111d35c9c7
add methods to queue full MTC and SPP messages, now that they are synchronous
2013-08-08 19:31:10 -04:00
d90e2b4221
rationalize (a bit) engine start/stop/restart so that it is possible to start up, disconnect from JACK and then reconnect
2013-08-08 16:31:08 -04:00
917aed0da0
add new files (code based on the old midi++ JACKMidiPort)
2013-08-08 09:10:11 -04:00
1ab61b8564
major redesign of MIDI port heirarchy and management (part 2)
2013-08-07 22:22:11 -04:00
83a0c30c24
major redesign of MIDI port heirarchy and management
...
basic, very flaky functionality is back. program unstable at present
2013-08-07 22:21:36 -04:00
a1370f83db
fix compiler warning
2013-08-07 15:57:29 +02:00
616f2a0370
fix conflicts after merge with master
2013-08-05 14:22:32 -04:00
499b7fcfa9
document PortEngine API
2013-08-05 13:55:50 -04:00
a5f69910e4
add notion of unavailable devices to ARDOUR::AudioBackend
2013-08-05 13:19:23 -04:00
a66e3859e1
can now start JACK based on config dialog
2013-08-05 12:51:51 -04:00
7218bd91de
start work on the changes to EngineControl (dialog) to integrate with new backend design, and add "requires-driver" concept to AudioBackend to handle JACK specifically
2013-08-04 14:03:19 -04:00
f5c52f500a
Timecode: fix subframe rounding
...
if sub-frames are used, frames are rounded up for the later half of
a sub-frame. e.g. @48KSPS, 25fps, 100 subframes per frame
audio-samples >=1910 already yield 00:00:00:01
2013-08-04 16:22:00 +02:00
976fc86811
fix no-roll LTC encoder, honor locates when transport is stopped
2013-08-04 16:13:43 +02:00
efc46f1553
make master-meter in toolbar optional
2013-08-04 00:06:29 +02:00
333a3c9d02
cleanup more dangling members of AudioEngine
2013-08-03 17:49:19 -04:00
a64908af7f
fix undeclared use of cerr
2013-08-03 17:47:26 -04:00
1d50d041c9
fix problem with session exit caused by dangling unused members in AudioEngine
2013-08-03 16:45:07 -04:00
f2ca3ecf0c
add file
2013-08-03 16:44:13 -04:00
c2e7c32c53
add tim's jack_utils code to rationalize setup of JACK config
2013-08-03 16:37:10 -04:00
91e0e0f9f8
advance track's play-position even if processing is locked
...
fixes
* http://tracker.ardour.org/view.php?id=5628
* http://tracker.ardour.org/view.php?id=5561
2013-08-03 21:35:19 +02:00
1c4d00e8b7
audioengine branch can now load and run at least one test session.
...
currently hard-coded to deal only with the situation where JACK is already running
2013-08-03 11:57:56 -04:00
b14d9a6451
properly handle port removal in route.
...
bug:
- add midi track
- add additional midi port
- remove additional midi port
- press play
-> crash
because..
ARDOUR::MidiTrack::roll's route_buffer still has 2 midi-channels. This is
because processor_max_streams was not updated (no plugins added/removed)
fill_buffers_with_input() eventually asserts when trying to get the 2nd
(nonexistant) midi port via ARDOUR::PortSet::nth_midi_port() in
ARDOUR::PortSet::port().
2013-08-03 13:44:37 +02:00
0eb5b7d1ab
Merge branch 'midiclock' of https://github.com/axetota/ardour
2013-08-03 01:37:25 +02:00
Michael Fisher
e82574de05
Send and Read MIDI Positions the right way
2013-08-02 18:23:36 -05:00
99fb7346f0
on session-load: skip output-change-handler until IOs are restored
...
fixes loading a session with tracks that have more input-ports
than output-ports.
the problem:
input -ports are added first, when added, IO::ensure_port() emits
IOChange::ConfigurationChanged which in turn triggers the route to
allocate corresponding output ports.
Due to this callback the output-ports were created before the actual
output-port-configuration and plugin-configuration from the
session-state was parsed and set.
2013-08-03 00:51:08 +02:00
23eba1cc39
disallow invalid port-removal
...
do not allow port-removal if the port would be re-added immediately
after that again because the main-delivery actually needs it.
As a side effect this prevents this crash:
* create a stereo-track, then remove one output
-> unhandled exception "AudioEngine::PortRegistrationFailure&"
The problem:
- the port is removed from the RCU ports list,
but Port::drop() (which calls jack_port_unregister) is only called
from the Port's destructor at some later time.
(because a reference to the port still exists elsewhere)
- the jack-port is not yet removed.
- meanwhile Delivery::configure_io comes along and notices that
there are more audio-buffers than ports and tries to re-register the port.
- but the port still exists in jack, so it fails and throws an exception
...which is not handled.
2013-08-03 00:51:07 +02:00
94f366190e
fix typos in display of error message about an LV2 plugin preset
2013-08-02 12:50:52 -04:00
Michael Fisher
71c26e2820
Resolve duplicate symbol 'cocoa_open_uri' on OSX builds.
...
cocoa_open_uri.mm was being added to the source list
twice and also compiling twice. This patch enusres it is
only added once.
2013-08-02 12:01:47 -04:00
