b2dc5a52b7
volatile means hardware-volatile.
2016-02-22 15:25:28 -05:00
52cb196695
Revert "Fix compilation warnings. (gtk2_ardour part)"
...
These fixes are wrong, not on their own, but because GtkPiano* should not be
using volatile. I'll fix that separately.
This reverts commit ad35ab78eb
2016-02-22 15:21:50 -05:00
Adrian Knoth
6fa88273aa
Spelling correction patch from Debian
...
Patch taken (and forward-ported to HEAD) from
<https://anonscm.debian.org/cgit/pkg-multimedia/ardour.git/plain/debian/patches/0100-fix-typos.patch >
2016-02-22 15:13:01 -05:00
Nils Philippsen
53f1981eff
don't trip over missing nm tool
2016-02-22 15:12:11 -05:00
André Nusser
6054d72c1b
Revert and additional change due to --strict warnings.
2016-02-22 15:11:02 -05:00
André Nusser
ad35ab78eb
Fix compilation warnings. (gtk2_ardour part)
2016-02-22 15:11:02 -05:00
André Nusser
1afbdaff58
Fix compilation warnings. (libs part)
2016-02-22 15:11:02 -05:00
86d27a736c
better system for managing kbd focus after cmd-w closing a dialog.
...
Key, Button and Focus In events in the editor+mixer windows cause the respective window to be noted
as the window to which focus should return after a dialog is closed with cmd/ctrl-w
2016-02-22 14:03:55 -05:00
60c4b30b7e
manually revert most of previous commit related to dialog/main window focus management
2016-02-22 13:55:31 -05:00
3a08ac5e1f
os x: an initial attempt at fixing loss of keyboard focus after cmd-w to close dialog
2016-02-22 11:49:17 -05:00
100e156727
Fix 6783, Using "Esc" key doesn't cancel creating new Track/Bus Group
...
Also fixes it for other sources of delete events like closing the dialog.
2016-02-22 15:22:56 +10:00
30497db823
prepare for gcc6 rint() precision degradation
...
gcc6 returns a float for "rint ((float) val)"
2016-02-22 00:07:21 +01:00
f6fe87d63e
prevent some mess when automating control ports on midi synths
2016-02-21 22:19:08 +01:00
3cc4ee8029
alsa: allow environment variable ARDOUR_ALSA_DEVICE to identify a single, arbitrary ALSA PCM name as the only available/allowed device
2016-02-21 12:44:54 -05:00
2d18283379
make threaded waveview rendering a configure time option (enabled by default)
2016-02-21 12:34:10 -05:00
344cf05540
OS X: fix the handling of SUPER,HYPER,META (see code comment for more)
2016-02-21 10:48:35 -05:00
13552140a2
prevent crash is VST does not implement effEditGetRect
2016-02-21 13:24:20 +01:00
43d3b4656a
add an OOM check.
2016-02-20 20:47:13 +01:00
67d99a3c5b
help clang static analyzer
2016-02-20 20:36:58 +01:00
f519edbe8e
fix segfault in libardour unit-tests (possible crash in mackie sf)
2016-02-20 20:26:34 +01:00
ff427be0fa
add processor lookup by PBD::ID via session
...
In preparation for Lua bindings this completes the basic set of
Session object lookup: route, source, controllable, region, processor
2016-02-20 20:26:18 +01:00
f0d40fb34a
fix linux bundle installer
2016-02-20 12:08:55 +01:00
318f3850de
third time lucky
2016-02-19 23:25:15 +01:00
452bca79eb
and now without typo
2016-02-19 23:19:19 +01:00
9633ea3b01
fix Windows OSC support
2016-02-19 23:16:38 +01:00
b0518fa2be
Revert "disable threaded waveform rendering for 4.7"
...
This reverts commit b0032adb44
2016-02-19 21:13:29 +01:00
02aa1be06c
possible fix for AVLinux VST bundles.
2016-02-19 21:11:55 +01:00
d84a8222f2
add MCP device file for Nucleus-via-platform MIDI
2016-02-18 20:28:56 -05:00
b0032adb44
disable threaded waveform rendering for 4.7
2016-02-18 18:56:05 -05:00
c92e822332
remove HYPER,SUPER,META modifier masks from Keyboard::RelevantModifierMask so that modifier_state_equals() works on OS X
...
GTK on OS X adds META when Command is pressed even though it also uses MOD2 to indicate COMMAND. So Cmd-w was actually META-MOD2-w and modifier_state_equals(MOD2)
would fail. This should have nothing but good side effects
2016-02-18 09:44:50 -05:00
dd7649a409
add a static utility function (should really be in utils.cc and public) to show modifier state in english; use said function
2016-02-18 09:43:03 -05:00
801a7ea7c4
Add Pavel Potocek to authors
2016-02-18 10:52:35 +10:00
Pavel Potocek
8d408d000d
Change the tap tempo estimator to least-squares regression
2016-02-18 10:32:49 +10:00
ec708926fa
now with return value
2016-02-17 17:03:14 +01:00
23e96f7f9b
Fix Post Export "Open Folder" button for OS X
2016-02-17 16:29:58 +01:00
5e50a7d6cf
allot to open OSX Finder with arbitrary paths
2016-02-17 16:29:25 +01:00
90efd9f010
translatable missing plugin message
2016-02-17 14:44:22 +01:00
9c5bdc6cf5
amend missing plugin message to hint that a plugin may be blacklisted
2016-02-17 14:29:53 +01:00
58f7a11287
Make sure that MSVC knows which version of 'ceil()' we want
...
(it doesn't have a version that understands 'framecnt_t')
2016-02-17 12:17:05 +00:00
f8bc9c66a5
OSC: remove "/session/loaded" message altogether, for testing on Win.
2016-02-16 16:06:39 -06:00
86edc1ccfe
Fill some more VST transport flags.
2016-02-16 15:33:12 -06:00
f30d1d6073
fix normalize + trim silence at end.
2016-02-16 22:24:01 +01:00
0ce4113dbb
prepare threshold for silence trim
2016-02-16 21:25:54 +01:00
31442cd63d
work-around case where metric is NULL in split cycles.
...
This really needs to be fixed on a higher level.
e.g Pluin::connect_and_run() needs to be passed the current time identical to
PluginInsert::connect_and_run()
2016-02-16 18:17:13 +01:00
cbdae78511
try to catch https://github.com/rncbc/synthv1/issues/1#issuecomment-183041991
...
(n_samples = 64 but Ardour sends midi-buffer with an event at 960)
some guesses as to why:
- split cycle for looping (nominal: 1024, cycle split:64)
- plugin uses _session.transport_frame() directly :(
(not latency compensated offset or looped position)
- "offset" is not taken into account for midi buffers
- tempo/metric change (metric iterator is wrong after loop)
2016-02-16 15:36:43 +01:00
c1910351a5
Revert "proper debug prints for Analyser"
...
This reverts commit 9a281963e1
2016-02-16 15:13:12 +01:00
240daf95f5
re-bin analysis results to take silence striping into account.
2016-02-16 15:09:31 +01:00
a664fb703e
chunk data before analysis, prefer 8K
...
* consistent spectrum display for all reports (same binsize)
* improved performance (power of two)
2016-02-16 14:25:49 +01:00
c52854fd3d
Analysis: 1 sample slack (for resampling rounding)
2016-02-16 14:24:08 +01:00
a929b8b872
fix Export SRC buffer report:
...
the resampler is always fed with N channels, it won't output half channels
2016-02-16 14:00:31 +01:00
