The connection was stored by the Port where it was initiated from:
1. Connect A to B
2. Disconnect B from A
(1) Port A remembers the connection to B
(2) Port B does not know about the connection from A
Since disconnect is initiated on port B, port A still retained
the connection information.
When restarting the engine, Port::reconnect() reestablished the
connection.
This is only relevant when libardour's own connection information is
used. e.g. the session is closed without engine, or when re-starting
the engine while the session is open.
Changes include workarounds for weird Rubberband API (pad, drop, truncate requirements),
and a shift to push logic after a slot is finished back up into the TriggerBox.
Not finished yet, some easily encountered bugs remain
::snap_to() was intended to round a Beats value to the nearest multiple
of another Beats value. It did not do that, but instead rounded down.
Worse, it used Beats::operator/ which in turn uses int_div_round(),
which is incorrect for a situation where we need integer truncation.
The changes fix the actual arithmetic and add 2 variant functions so that the
API includes round down, round up and round to nearest.
max_samplepos and max_samplecnt and both INT64_MAX which is (a) too large to fit into a signed 62 bit
integer and (b) definitely too large to be represented in a signed 62 bit superclock value.
Move the constructors that use samplepos_t into the .cc file, and treat these two values as special
cases that mean "as large/late/huge/long as possible".
This works around a race-condition, calling d'tors from
two threads concurrently.
The GUI thread destroys ctrl surfaces. ~~MIDIControllable()
calls ::drop_external_control() -> ::midi_forget()
This unsubscribes from signals (notably MIDI::Parser events)
by calling ScopedConnection::disconnect(), Connection::disconnect().
At the same time auto_connect_thread can call
PortManager::clear_pending_port_deletions() which removes
the MIDI port and destorys the MIDI::Parser.
~Parser() calls Connection::signal_going_away() to invalidate
connected signals.
This can deadlock if it is called concurrently with
Connection::disconnect() on the same signal.
see also
https://discourse.ardour.org/t/ardour-session-close-hangs/106523/10
Instead of just restarting it inside the same ::run() call, mark it as waiting,
so that it will restart at the next quantization point.
This isn't a final or even fully correct solution, but it's a useful test of
the idea.
The old code used the instantaneous tempo at T0 to compute where the next
quarter note would be. This is incorrect, since the tempo is
changing (continuously, for now) during the time represented by that quarter
note. Instead, we need to add a quarter note (or technically, whatever the
tempo note type is) to get a new position in beats, then compute the superclock
time at that location (which will use our equations for tempo, including the
use of omega, the ramp factor).
These comments should correct an impression left in the commit message for
6e9e28343bc3695d that there may be some sort of problem with synchronization
of TempoMap changes. The actual problem is that TempoMap edits are done using
RCU, so the modifications are performaned using a copy of the map, but with
map elements taken from the pre-copy version.
