gcc-8.2 -O3 optimization enables -finline-functions.
For reasons yet unknown this causes issues with Glib::Threads::Private
when loading sessions (creating new ones is fine) and likely some
other operations.
Without this change the following happens (5.12 as well as current git)
when loading an existing session:
---
[Switching to Thread 6092.0x12fc]
Breakpoint 1, ARDOUR::SessionEvent::create_per_thread_pool (name=..., nitems=4096)
at ../libs/ardour/session_events.cc:60
60 in ../libs/ardour/session_events.cc
$6 = (const std::__cxx11::string &) @0x210dfc60: {static npos = 18446744073709551615,
_M_dataplus = {<std::allocator<char>> = {<__gnu_cxx::new_allocator<char>> = {<No data fields>}, <No data fields>},
_M_p = 0x210dfc70 "GUI"}, _M_string_length = 3, {_M_local_buf = "GUI\000\000\000\000\000<E0>P<DA>:\000\000\000",
_M_allocated_capacity = 4805959}}
#0 ARDOUR::SessionEvent::create_per_thread_pool (name=..., nitems=4096) at ../libs/ardour/session_events.cc:60
#1 0x0000000000434e77 in ?? ()
[..]
Thread 1 (Thread 6092.0x12fc):
#0 0x0000000061f152ef in PBD::RingBuffer<void*>::write (this=this@entry=0x50, src=src@entry=0x210dcdc8,
cnt=cnt@entry=1) at ../libs/pbd/pbd/ringbuffer.h:170
#1 0x0000000061ee954d in push (t=<optimized out>, this=0x0) at ../libs/pbd/pool.cc:290
#2 CrossThreadPool::flush_pending_with_ev (this=0x0, ptr=<optimized out>) at ../libs/pbd/pool.cc:254
#3 0x00000000218d5b01 in operator() (a0=0x210dcdc8, this=0x48)
at /home/ardour/win-stack-w64/include/boost/function/function_template.hpp:682
#4 operator()<boost::function<void(ARDOUR::SessionEvent*)>, boost::_bi::list0> (a=<synthetic pointer>, f=...,
this=0x68) at /home/ardour/win-stack-w64/include/boost/bind/bind.hpp:259
#5 operator() (this=0x48) at /home/ardour/win-stack-w64/include/boost/bind/bind.hpp:1294
#6 boost::detail::function::void_function_obj_invoker0<boost::_bi::bind_t<boost::_bi::unspecified, boost::function<void (ARDOUR::SessionEvent*)>, boost::_bi::list1<boost::_bi::value<ARDOUR::SessionEvent*> > >, void>::invoke(boost::detail::function::function_buffer&) (function_obj_ptr=...)
at /home/ardour/win-stack-w64/include/boost/function/function_template.hpp:159
#7 0x0000000063e2dc39 in operator() (this=0x210dcf70)
at /home/ardour/win-stack-w64/include/boost/function/function_template.hpp:682
#8 AbstractUI<Gtkmm2ext::UIRequest>::call_slot(PBD::EventLoop::InvalidationRecord*, boost::function<void ()> const&)
(this=0x3ada50e0, invalidation=0x0, f=...) at ../libs/pbd/pbd/abstract_ui.cc:434
#9 0x00000000216cbb96 in ARDOUR::SessionEventManager::merge_event (this=<optimized out>, ev=<optimized out>)
at /home/ardour/win-stack-w64/include/boost/function/function_template.hpp:507
#10 0x00000000216e3861 in non-virtual thunk to ARDOUR::Session::queue_event(ARDOUR::SessionEvent*) ()
at ../../win-stack-w64/include/glibmm-2.4/glibmm/threads.h:709
#11 0x00000000216c91bf in ARDOUR::SessionEventManager::clear_events(ARDOUR::SessionEvent::Type, boost::function<void ()>) (this=this@entry=0x4256cc08, type=type@entry=ARDOUR::SessionEvent::Skip, after=...)
at ../libs/ardour/session_events.cc:156
#12 0x000000002168874a in ARDOUR::Session::sync_locations_to_skips (this=this@entry=0x4256ca90)
at /usr/lib/gcc/x86_64-w64-mingw32/8.3-posix/include/c++/new:169
#13 0x0000000021688802 in ARDOUR::Session::update_skips (this=this@entry=0x4256ca90, loc=loc@entry=0x0,
consolidate=consolidate@entry=false) at ../libs/ardour/session.cc:1683
#14 0x0000000021699649 in _locations_changed (locations=..., this=0x4256ca90) at ../libs/ardour/session.cc:1859
#15 apply<ARDOUR::Session> (method=
(void (ARDOUR::Session::*)(ARDOUR::Session * const, const ARDOUR::Locations::LocationList &)) 0x216994f0 <ARDOUR::Session::_locations_changed(std::__cxx11::list<ARDOUR::Location*, std::allocator<ARDOUR::Location*> > const&)>,
obj=..., this=0x40ff4c60) at ../libs/ardour/ardour/location.h:240
#16 ARDOUR::Session::locations_changed (this=this@entry=0x4256ca90) at ../libs/ardour/session.cc:1840
#17 0x0000000021700abd in ARDOUR::Session::set_state (this=0x4256ca90, node=..., version=6000)
at ../libs/ardour/session_state.cc:1657
#18 0x0000000021713456 in ARDOUR::Session::post_engine_init (this=this@entry=0x4256ca90) at ../libs/pbd/pbd/xml++.h:81
#19 0x00000000216ad5cd in ARDOUR::Session::Session (this=0x4256ca90, eng=..., fullpath=..., snapshot_name=...,
bus_profile=<optimized out>, mix_template=...) at ../libs/ardour/session.cc:383
#20 0x000000000046db8b in ?? ()
Backtrace stopped: previous frame inner to this frame (corrupt stack?)
---
Note that in the trace starting at #19 the source-file references
are offset :post_engine_init() isn't in xml++.h
(likely debug symbols are incorrect due to inline functions)
The application crashes due to #2 ::flush_pending_with_ev (this=0x0 ..)
But the earlier breakpoint above shows that the ThreadPool was
created, and this #11 ::clear_events() should not assign NULL.
(This happens on a single-core CPU, too)
Locating the start of the loop range at transport start does not count as "have looped". This was preventing
pre-roll when starting to loop, which in turn caused data/event loss because the start/end sample ranges
would be incorrect inside process
This reverts commit 8702ff2189,
and b10d9cf09b.
There was a misconception on the iterator (port vs message
in sequence), besides Ardour's mixer-strip is preferably used
with a single MIDI port. Most plugin-standards also only support
one port.
If need be LuaDSP run_map() can be used to handle multiple
MIDI I/O ports already.
We accomplish this by explicitly setting the global_obs to NULL after
free'ing it. The crash that led to this fix can be replicated as such
$ sendosc localhost 3819 /set_surface i 8 i 159 i 8
$ sendosc localhost 3819 /set_surface i 0 i 0 i 0
$ sendosc localhost 3819 /set_surface i 0 i 0 i 0
In this example the observer is created, then free'd, and then free'd
*again* because the new observer was never made (sur->feedback[x]
checks fail)
gcc, and recent clang-10 can construct new objects
using references as arguments.
However clang-8 (and MSVC?) do not:
"error: no matching function for call to 'operator new'"
The compiler apparently does not expand the template
class A <-> `A*` vs. `A const&`
for different cases.
Mixbus v6 uses AutomationType BusSendLevel like other Sends,
however with different min/max range compared to default Ardour
Aux sends. Control surfaces should use interface/internal API.
Bug was introduced in 128a45954c, declick-amp gain was
overridden, but declick not cleared. For some reason this did
not affect audio-only exports nor all session exports.
This currently only handles initial session load by dbl-click on
a session file. It also does not work if the user or new or
subject to a pre-release warning. These last two restrictions
may remain in place, although if a person's first exposure to
the program is "copy my session, download program, dbl-click
on the session file", then that's not so nice
This is a dumb stack allocator using LIFO allocation/free, with
a fallback to new/delete. This works well for small STL containers
in particular std::vector, but it's also suitable for std::map<>,
in particular copy constructing small POD maps (plugin pin mappings).
Eventually this could be combined with TLSF for a flexible
memory arena management. This is however not currently needed
for any the planned use-cases.
This code is ANSI C++98 compatible, and yet also works with
modern C++11, C++14
This use-case is better served via stem-export.
This export mode never worked correctly when latent plugins
are present on the track. or when realtime export was chosen.
This used to call
track.export_stuff() -> bounce_process()
periodically in small chunks from freewheeling callback,
which is no longer functional.
Ardour's playback is aligned to master-out:
"When the playback clock reads 01:00:00:00, the sample corresponding
to 01:00:00:00 is audible at the speaker(s)"
When exporting, and grabbing data from output ports, the signal
is offset by the master-bus physical playback latency. This was
compensated for, but lead to initial silence in the exported file.
New approach is to start capturing export data during pre-roll,
at the time when playback is written to the output buffers.
To also shaves off a common offset to make this work with
realtime export. Effectively this emulates processing with
disconnected master-output port, while still keeping any
latency of effects on the master-bus itself.
Last but not least: jack updates latencies when freewheeling,
(setting HW latency to zero). The callback arrives asynchronously
some time after enabling freewheeling, but after Export
Ports have been configured. Those callbacks are ignored.
After exporting a time-span, the next time-span was started
directly from the rt-callback. This had various issues.
In particular with realtime export.
Post-processing of a realtime-export enables freewheeling
and is driven by freewheel callbacks. Freewheeling needs to be
safely disabled for an upcoming realtime export.
A similar issues existed when mixing realtime and non-realtime exports.
