Playlist::_split_region copies the region and BBT_Time member is default
constructed so position is bar 1 beat 1.
The regions sample position is then initialized to the new position using the
property list.
Playlist::add_region_internal is then used to add new regions to the playlist
which calls region->set_position() but as the region position has already been
initialized it does not recompute the _bbt_time member based on the new sample
position.
Then when a tempo change occurs and Region::update_after_tempo_map_change is
called the default initialized _bbt_time member is used to incorrectly
determine the new sample position.
So the change removes the initialization of the new region position using the
property list initialization method and just lets the playlist set the position
of the region in Playlist::add_region_internal so that the region _bbt_time
member is recomputed in Region::set_position_internal.
Taking a readlock after a writelock in the same thread should result
in a deadlock, yet pthread on Linux returns EDEADLK and continues.
glib-2.42.0 ignores EDEADLK and assumes the lock was taken. Releasing
the lock later causes issues: "Calling g_rw_lock_writer_unlock() on a
lock that is not held by the current thread leads to undefined behaviour."
The issue at hand:
AudioStreamView::redisplay_track()
-> foreach_region() #<< WriteLock
-> add_region_view()
...
-> AudioRegionView::create_one_wave()
-> RegionView::update_coverage_frames
-> Playlist::top_unmuted_region_at() #<< ReadLock
All current users of Playlist::foreach_region() are in the GUI
and AFAICT read-only (display regions, update visuals)
Currently we don't do anything special on layering prefs change
(relayer only occurs when each playlist is first edited). The idea here is
that "undo" is still available to restore previous layering in case of any
surprises.
Comments in various call sites of Evoral::coverage() marking things I think
are dubious (with XXX). Also straightened up the alignment of some ASCII
art in libs/ardour/diskstream.cc
This will be used when you paste with Ripple; it maintains the whitespace that you chose in the selection.
Also fix default setting of Dim contol by storing the default(normal) value for MPControl Controllables.
Add a value for Ripple to EditMode enum.
Add Ripple edit mode to edit mode dropdown, by adding it to the
Editor::build_edit_mode_menu() helper function, and remove the old code that
added items to the (now unused) Editor::edit_mode_strings.
Add the regions that should be affected by the drag to RegionDrag::_views so
that the drag carries them along automatically.
Use a copy of the RegionList in Playlist::core_ripple(), since bad things
happen when iterating over regions and they get moved around in the list.
Handle rippling in removal of regions from playlist.
When dragging in ripple mode, exclude all regions that lie before the
original start position of the selected regions being dragged from
rippling: this is what Mixbus does.
Make editor dragging respect snap-to settings, by using the existing
compute_x_delta() function, which did almost the right thing. Move setting
of _last_frame_position out of that function so all ripple-dragged regions
can move.
Ripple when dragging from region list: even though Mixbus doesn't do this, it
seems like a good idea.
Prevent multi-track selection being dragged across tracks, by making
RegionMotionDrag::y_movement_allowed() virtual, and overriding it in
RegionRippleDrag to forbid dragging of selections containing regions on more
than one track to dofferent tracks in ripple mode.
Remember which TimeAxisView a ripple-mode drag that's allowed cross-track
drags started from, so that the effect of rippling regions after any region
that's dragged off that track can be undone.
When adding regions in splice mode, exclude the region being added from
possibly being shuffled. I don't know whether this might have some other
adverse effects, but it fixes an obvious defect in 'Splice' mode where
newly-recorded regions 'jump' to the playhead position when recording
stops, and since splice mode is pretty much broken anyway, I don't think it
can make matters any worse.
the regions on the compound list are arranged the same
as on the original.
git-svn-id: svn://localhost/ardour2/branches/3.0@12643 d708f5d6-7413-0410-9779-e7cbd77b26cf
It's slightly possible that this causes trivial build failures on different
configurations, but otherwise shouldn't cause any problems (i.e. no actual
changes other than include/naming/namespace stuff). I deliberately avoided
removing libardour-config.h since this can mysteriously break things, though a
few of those do seem to be unnecessary.
This commit only targets includes of ardour/*.h. There is also a very large
number of unnecessary includes of stuff in gtk2_ardour; tackling that should
also give a big improvement in build time when things are modified.
git-svn-id: svn://localhost/ardour2/branches/3.0@12420 d708f5d6-7413-0410-9779-e7cbd77b26cf
they are now done with region fades, rather than separate objects.
After this commit, Ardour will try to convert your session files
to the new crossfade format, but will make a backup in your
session folder first.
If you have works in progress using Ardour 3 it is
***STRONGLY RECOMMENDED*** that you back up session files before
updating to this commit.
git-svn-id: svn://localhost/ardour2/branches/3.0@11986 d708f5d6-7413-0410-9779-e7cbd77b26cf
for the undo history, so that changes to regions' layering_index
get stored in the undo record. Make Playlist::update use add_region_internal
so that undone regions don't have their layering_index corrupted.
Setup layering indices on relayer() so that deletion of regions
causes an update.
git-svn-id: svn://localhost/ardour2/branches/3.0@11123 d708f5d6-7413-0410-9779-e7cbd77b26cf
works properly. Stop the layer being a stateful property, as it is
always derived from layering_index, unambigiously, by relayer().
git-svn-id: svn://localhost/ardour2/branches/3.0@11120 d708f5d6-7413-0410-9779-e7cbd77b26cf