Use poll() with a finite timeout so we can quit the thread properly
without manually writing stuff into the pipe.
Check for pending events at the beginning of the loop so we can't end up
in an endless loop.
Furthermore check return value of snd_seq_event_input() and get out of
loop if it indicates an error.
Use QString::isEmpty() rather than comparing against an empty string.
Furthermore use QStrin::toUtf8() instead of QString::toAscii() for
retrieving name of a MidiPort.
Suppressed a message box that would pop up in command
line interface causing a crash and made lmms act properly
when a rendering plugin fails. Need to figure out how to
terminate program in project_renderer line 222.
The mp3 decoder was using an uncomfortably large buffer,
laughably because of a negative overflow. Fixed this and
added a failsafe check to prevent buffer overflows.
Additionally changed printfs to qWarnings and fixed copyright
notice in audio_file_mp3.
The audio_jack driver currently treats the jack transport status as a
flag for whether or not to output audio. This is incorrect and
unintended usage of jack transport. The attached patch simply removes
all references to jack transport from the driver, so that LMMS will
always be able to produce audio.
Closes#2816029.
Per request, make the Controller-rack vertically sizable. This is also
part of our effort to slowly deprecate old fixed-positioning code in
favor of layouts.
Switching to HQ mode in standard workflow doesn't make much sense and
is likely to cause problems. Therefore remove the HQ mode button for
for time being. One can still export projects in high(er) quality.
* cleaned up code and improved coding style
* removed old SINGLE_SOURCE_COMPILE macro relicts
* use QString::toUtf8() instead of QString::toAscii() wherever possible
The macro logic for defining the SPINLOCK_PAUSE macro tested against
LMMS_HOST_X86_64 twice instead of testing against LMMS_HOST_X86_64 and
LMMS_HOST_X86. This caused the SPINLOCK_PAUSE macro not being set on
x86.
Added macros PREFETCH_READ() and PREFETCH_WRITE() which insert proper
prefetch instructions. At least on my Intel Atom CPU this improves
performany slightly. Probably needs some further tuning.
Signed-off-by: Tobias Doerffel <tobias.doerffel@gmail.com>
The new CPU framework replaces the old BasicOps framework. It is more
flexible and the build process isn't such a mess anymore (pre-compiled
assembler files etc.). It will hopefully see some improvements and
extensions soon.
Signed-off-by: Tobias Doerffel <tobias.doerffel@gmail.com>
Removed myself from list of copyright holders as I didn't help out
with LAME support at all.
Signed-off-by: Tobias Doerffel <tobias.doerffel@gmail.com>
Properly handle HTTP status code 302 which indicates a moved site,
i.e. try to fetch data from the new location. This finally makes
WebResources work again with LSP being moved to http://lmms.info.
Furthermore cleaned up code a bit. There's no need for a global
QHttp object or a global QBuffer object. Instead allocate them on
the stack in the according methods.
Furthermore removed unused/incomplete support for non-blocking download.
We should introduce support for a ProgressDialog solution (also allowing
to cancel operations) later.
Signed-off-by: Tobias Doerffel <tobias.doerffel@gmail.com>
<superjoe> why is that code even there?
<pgib> until the next mixer-period, then the model's value is reset
to the actual automation track
<pgib> I have no idea. I think it should be removed honestly. again,
there is no comment explainign the true intent
* superjoe experiments
<pgib> it was probably a poor attempt to make it so when you adjust
the only automation point for a track, that the value of the model
will change immediately'
<pgib> I think it is safe to move the newTime==0 chunk
<pgib> if we want to make the automationEditor immediately change the
value, there needs to be more logic
<pgib> and it would belong in automationEditor -- not automationPattern
<pgib> but this fixes the superAutomation why-in-the-hell bug
When you press down or up while browsing in the resource browser,
it plays the sample.
FIXED: When you click on the keyboard in the resource browser, and then
click on the treeview, it didn't transfer focus.
exportProjectDialog - select the output file format based on
what the user has already chosen in the export dialog. Prevents
the user from having to choose twice.
(also fixed whitespace and added modelines where applicable)
Refactor supported file extensions from AudioFileProcessor to
sampleBuffer, where the magic really happens. Change the hard-
coded MessageBox explaining supported formats to the real deal.
Also added modelines and switched spaces to tabs in the files
affected.
There is a new folder setting called Lame Library. AudioFileMp3
will look here for libmp3lame.so.0. If it can't find it it will
notify the user how to make it work and abort rendering gracefully.
Explicitely destroy instance of ResourceBrowser in destructor of
MainWindow. This fixes crash at exit due to auto-deletion mechanisms
after the whole engine already has been shutdown.
Signed-off-by: Tobias Doerffel <tobias.doerffel@gmail.com>
Setting time signatures 1/16+ lead to divisions by zero in two places.
Fixed this by adding according qMax() call (closes#2818125).
Signed-off-by: Tobias Doerffel <tobias.doerffel@gmail.com>
MidiPort::outputChannelModel is ranged from 1 to 16 for displaying
user-friendly values. However internally MIDI channels are ranged from
0 to 15. Therefore added realOutputChannel() which returns zero-based
MIDI channel which should be used everywhere except for the GUI.
Fixes MIDI events being sent to VST plugins on channel 2 instead of
channel 1. Makes some more VST plugins actually usable.
Signed-off-by: Tobias Doerffel <tobias.doerffel@gmail.com>
Starting with Qt 4.5 there were minor graphical glitches in
TrackContentWidget. Fix this by not setting Qt::WA_OpaquePaintEvent
attribute.
Signed-off-by: Tobias Doerffel <tobias.doerffel@gmail.com>
When pasing notes via Ctrl+V somehow the edit-tool was left in an
undefined state. Therefore explicitely switch back to draw mode so
user can move pasted notes (closes#2808607).
Signed-off-by: Tobias Doerffel <tobias.doerffel@gmail.com>
Variuos features (native/Qt semaphores/shared memory) are now
configured by macros like USE_QT_SHMEM and USE_QT_SEMAPHORES. This
allows central and individual configuration of features according to
the platform to build for at the beginning of the file.
Finally makes RemotePlugin work on OS X by not using native semaphores
rather than QSystemSemaphore's.
Signed-off-by: Tobias Doerffel <tobias.doerffel@gmail.com>
In InstrumentTrack::play(...) the local variable "bb_track" always
was either undefined or NULL due to a lost call (probably during M/V
architecture switch) to bbTrack::findBBTrack(...). This could lead to
crashes and/or misbehaviour when playing frozen patterns in BB tracks.
Signed-off-by: Tobias Doerffel <tobias.doerffel@gmail.com>
Added new DEBUG_REMOTE_PLUGIN macro. If it is set, the process won't
be launched. Instead important information are printed allowing the
developer to run the process with according parameters in a debugger.
Signed-off-by: Tobias Doerffel <tobias.doerffel@gmail.com>
Commit e3fa8762ef introduced a crash
in destructor of InstrumentTrack when using IPH based instruments.
Therefore explicitely remove all play handles after calling
silenceAllNotes().
Signed-off-by: Tobias Doerffel <tobias.doerffel@gmail.com>
This commit allows finer control of which kind of PlayHandles get
removed in Mixer::removePlayHandles().
This is now used by InstrumentTrack which only removes NotePlayHandles
in InstrumentTrack::silenceAllNotes(). Fixes broken preview of resources
loaded by IPH-based instruments.
Signed-off-by: Tobias Doerffel <tobias.doerffel@gmail.com>
A type name for ResourceItem::PluginSpecificResource was missing in the
s_typeNames map. Therefore the type property of such resources was not
saved properly and had to be guessed each time at startup.
Signed-off-by: Tobias Doerffel <tobias.doerffel@gmail.com>
