Commit Graph

1770 Commits

Author SHA1 Message Date
Robbert van der Helm a56e4b337f Rename the VST3 mutual recursion functions 2021-04-29 14:03:15 +02:00
Robbert van der Helm 45d83ad9a1 Revert "Separate mutual recursion on GUI and other threads"
This reverts commit a495f1a67f.

This ended up not being an issue. What we _do_ have to do, sadly, is to
have a mutual recursion context stack per plugin. Otherwise multiple
plugin instances can deadlock eachother.
2021-04-29 13:54:48 +02:00
Robbert van der Helm a2e3e691d2 Change wording in the changelog 2021-04-29 12:30:19 +02:00
Robbert van der Helm a0abfed689 Merge the two prefetching changelog entries 2021-04-29 12:25:54 +02:00
Robbert van der Helm a495f1a67f Separate mutual recursion on GUI and other threads
I wasn't able to get this to clash, but this way we can be 100% sure
that there aren't any weird issues.
2021-04-29 12:07:11 +02:00
Robbert van der Helm 8b168b310c Fix mutual recursion with latency in Ardour/Mixbus
This would cause Ardour and Mixbus to freeze when inserting a latency
introducing (JUCE based) VST3 plugin. As mentioned in #98.
2021-04-29 03:13:53 +02:00
Robbert van der Helm 22f94dd22f Prevent unnecessary copies in ScopedValueCache 2021-04-29 02:58:05 +02:00
Robbert van der Helm ee24969539 Mention SWAM Cello among the fixed plugins
Since this now no longer requires an option to get good performance.
2021-04-29 02:19:17 +02:00
Robbert van der Helm fb04524b1a Update the tested Mixbus version 2021-04-29 01:22:50 +02:00
Robbert van der Helm 6f6e6c5b94 Reword caching to prefetching
Since that makes it much clearer what we're actually doing. With old
`cache_time_info` was actually caching the response, but now we're
querying it before the plugin has even requested the information.
2021-04-29 01:07:14 +02:00
Robbert van der Helm c82eb35243 Also cache audioMasterGetCurrentProcessLevel()
Melda plugins seem to call this during every processing cycle.
2021-04-29 01:02:10 +02:00
Robbert van der Helm 383b0c3f10 Remove the cache_time_info option
The new time info caching behaviour supersedes this by getting rid of
callbacks altogether.
2021-04-29 00:46:43 +02:00
Robbert van der Helm 2a9840828f Add a todo for also caching the process level 2021-04-29 00:42:44 +02:00
Robbert van der Helm b8f8dd7236 Log cached audioMasterGetTime() calls
Just like we log cached VST3 function calls.
2021-04-29 00:41:56 +02:00
Robbert van der Helm 1deb4cf664 Send the VST2 transport info along with processing
And cache it during the processing cycle. This greatly reduces the
overhead of bridging VST2 plugins.
2021-04-29 00:32:25 +02:00
Robbert van der Helm 6a3c726acf Get rid of spurious accessibility label 2021-04-29 00:31:11 +02:00
Robbert van der Helm 6cd1b664d3 Rearrange fields in Vst2Bridge 2021-04-29 00:31:11 +02:00
Robbert van der Helm 70adf5493e Add a logger to the Wine VST2 bridge 2021-04-29 00:31:11 +02:00
Robbert van der Helm 32db921b9b Move generic logger on the Wine side to common.h
We'll also need this for VST2 plugins when we start caching the time
info.
2021-04-29 00:31:11 +02:00
Robbert van der Helm dee964ec89 Add a helper for scope-bound caching of values 2021-04-29 00:31:11 +02:00
Robbert van der Helm 676f995f83 Remove the changelog entry for force_ftz
Since this behaviour is now always enabled.
2021-04-28 15:36:14 +02:00
Robbert van der Helm 14b0ba0c62 Clarify the definition of empty_arrangement
This is the same thing of course, but it's a bitfield so just a single 0
may seem a bit odd if you don't already know that.
2021-04-28 13:41:24 +02:00
Robbert van der Helm 216c6bc4f2 Always provide at least some speaker arrangement
This works around a bug in the VST3 version of W. A. Production
Imperfect as mentioned in #97. Even if it's a synth and numInputs is 0,
the plugin will still try to read the input arrangement.
2021-04-28 13:35:15 +02:00
Robbert van der Helm 23f94b35d0 Fix move semantics in Win32Thread and Win32Timer
I was sure that moving an `std::optional<T>` would reset the object you
moved from, but apparently not.
2021-04-28 12:56:49 +02:00
Robbert van der Helm 160c6acb85 Properly handle move semantics in the scoped FTZ 2021-04-28 12:55:58 +02:00
Robbert van der Helm 93f089eca7 Always set FTZ instead of gating it behind a flag
After a quick round of testing it seems like REAPER doesn't always
enable this on the audio thread, but Bitwig, Ardour, Carla and Renoise
do. So it should be safe to just get rid of the option and to leave this
enabled all the time.
2021-04-28 12:07:14 +02:00
Robbert van der Helm d75805820d Mark the options that are not yet in 3.1.0 2021-04-28 00:53:31 +02:00
Robbert van der Helm a12e0dcc27 Enable SSE2 instructions
We need this to enable FTZ in 3aac8e3483.
SSSE2 is a default on x86_64, but we need to explicitly enable it when
compiling the 32-bit bitbridge.
2021-04-27 23:50:28 +02:00
Robbert van der Helm 2be41da9b6 Add compatibility option to force flushing to zero
This prevents Kush Audio REDDI from taking down the DAW when the host
passes it denormalized audio to process. I've discovered that the issue
with this plugin had to do with denormals in the issue linked below, but
I didn't realize that we can just enable the FTZ flag for plugins that
don't already do so.

https://github.com/osxmidi/LinVst/issues/174
2021-04-27 23:33:50 +02:00
Robbert van der Helm 3aac8e3483 Add a RAII wrapper for enabling flush-to-zero 2021-04-27 23:00:41 +02:00
Robbert van der Helm 73d9f5f9b8 Remove unnecessary include 2021-04-27 20:46:08 +02:00
Robbert van der Helm ab56115c87 Add Bitwig+Chromaphone 3 as a hide_daw candidate
Under Bitwig, the plugin will crash when trying to enter text unless you
hide the DAW the plugin is being hosted under.
2021-04-27 18:57:09 +02:00
Robbert van der Helm ee50044062 Also allow hiding the host for VST3 plugins 2021-04-27 18:50:47 +02:00
Robbert van der Helm 3a73fb9689 Allow hiding the DAW name for VST2 plugins
This can be useful when plugins have (broken) host-specific behaviour
that you want to avoid. I'll later add a list of host/plugin
combinations where this may be useful to the readme.
2021-04-27 18:41:36 +02:00
Robbert van der Helm 38abdbee50 Rearrange the configuration options 2021-04-27 17:54:46 +02:00
Robbert van der Helm b52362e698 Fix time info cache being applied to other opcodes
I don't know how I've never noticed this, but we should of course only
be handling `audioMasterGetTime()` this way. This also explains why
enabling this permanently in the past broke some plugins.
2021-04-27 17:32:32 +02:00
Robbert van der Helm 5a68b8b39d Flush the reparent from 0c7dbe8
This should add a root reparent to prevent that X11 error, but it should
still prevent flickering by doing it just before closing the window.
2021-04-27 03:43:45 +02:00
Robbert van der Helm 0d1d368e4f Fix spelling in the changelog 2021-04-26 23:41:26 +02:00
Robbert van der Helm dfc9ad5d64 Add changelog entry for the version detection fix
For the WINELOADER version detection fix from #95.
2021-04-26 19:29:15 +02:00
Robbert van der Helm 93df3fa1da Fix mismatching types in 2aadf52 #95
`wine_path` was a string, and `boost::process::search_path()` returns an
`boost::filesystem::path`. We should just be using paths here anyways.
2021-04-26 19:25:28 +02:00
Kira Bruneau 2aadf5256b Fix printing wine version with custom WINELOADER
- Don't call bp::search_path when using WINELOADER. It will return an
  empty string for an absolute path.

- To match the behaviour of the exe wrapper scripts, only print the
  wine version from WINELOADER if the path is executable.
2021-04-26 19:20:10 +02:00
Robbert van der Helm 0c7dbe8a4a Reparent to the root window before deferring close
We did this before implementing the deferred close in yabridge 3.0.0. It
didn't seem necessary anymore so we got rid of it, but without this
closing an iZotope Rx plugin's editor in Renoise was guaranteed to
trigger an X11 error and crash Renoise. Doing this reparent doesn't seem
to cause any slowdown but it does at least fix the specific combination
of iZotope Rx and Renoise.
2021-04-26 18:47:58 +02:00
Robbert van der Helm 0d53f03b28 Get rid of an incorrect comment
This probably ended up here during a refactor.
2021-04-26 18:32:50 +02:00
Robbert van der Helm f6cf1a7dd2 Use explicit narrowing for SetWindowLongPtr() 2021-04-26 18:32:50 +02:00
Robbert van der Helm 661ea5f573 Update the tested REAPER version 2021-04-26 15:27:45 +02:00
Robbert van der Helm 10b4c0929d Mention corefonts in the runtime dependencies 2021-04-26 15:07:40 +02:00
Robbert van der Helm 75892794f8 Add a suggestion for disabling Serum's tooltips
These won't disappear once they're shown. I though that I'd already
listed this somewhere, but apparently not.
2021-04-26 13:48:08 +02:00
Robbert van der Helm a42ee0400f Print warning explicit COM init was necessary
So far only PSPaudioware InfiniStrip needed this. but it may be a good
idea to make this visible since it's probably an issue with the
plugin (even if most Windows hosts will have COM initialized).
2021-04-26 00:42:17 +02:00
Robbert van der Helm 68c846324c Also explicitly list simple dispatcher opcodes 2021-04-25 23:07:24 +02:00
Robbert van der Helm 3e71fc19b8 Explicitly handle VST2 callbacks with nullptrs
This is probably not needed, but we'll do it just in case another plugin
does something similar.
2021-04-25 22:44:15 +02:00