Age | Commit message (Collapse) | Author |
|
|
|
outside of libardour)
|
|
|
|
|
|
In preparation for archiving files as .flac (fixed point),
normalized with gain factor.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
barriers, even though lock ordering is likely correct now
|
|
::auto_connect_thread_wakeup() because that method is allowed to "fail".
Should fix hangs when closing a session with the ALSA backend (and perhaps others too).
|
|
|
|
gui_move_tempo -> gui_set_tempo_position
gui_move_meter -> gui_set_meter_position
gui_dilate_tempo -> gui_stretch_tempo
|
|
various failures.
|
|
|
|
destroyed.
Otherwise, control protocol ports don't actually get deleted
|
|
|
|
|
|
|
|
during import
|
|
|
|
- affects region relative time conversion when
initial meter is non-zero.
|
|
|
|
MIDI tracks
|
|
MIDI tracks
|
|
|
|
|
|
assigned/unassigned to/from a master
|
|
|
|
'part-of' ctor.
- when drag-copying MIDI regions, a partial region ctor is used
which can lead to _start_beats being off by a sample.
this in turn leads to 7168 happening, but only if magnetic snap
is used and the regions are copied as a group.
this should fix 7168, but the magnetic snap part is not yet addressed.
|
|
read_peaks_with_fpp() already holds _lock, build_peaks_from_scratch()
takes the _lock again.
Depending on glib[mm] and the threading lib it may either result in a
deadlock, or with EDEADLK in undefined behavior when a
non-recursive lock is released twice.
|
|
|
|
|
|
while snapped to grid.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
- don't parse XML into XMLTree
- only read the file, extract relevant elements
- don't read session-template contents, only test file
|
|
|