We had to add an even hackier hack now to get Boost.Process to
interoperate with Asio's IO contexts. This will be replaced later when
we replace Boost.Process.
With the `ghc::filesystem` dependency from the previous commit. If we
can replace the rest of the Boost.Filesystem dependency then we can get
rid the one nasty runtime dependency we have, and it will make
implementing the chainloading simpler since can reuse more code without
bringing in Boost.
I noticed that there were some realtime adhoc-acceptors running on my
system. That should of course not happen, since these only exist to
catch some sporadic (and likely as a result of a badly behaving plugin)
mutual recursion on the audio thread.
This fixes Waves V13 VST3 plugins crashing when opening the editor. They
will likely still crash later on anyways because they're kinda broken.
Amazing.
This wasn't implemented yet because no plugin tried using the interface
in this way before this, but Surge XT incorporates the host's context
menu items into their own (much more elaborate) context menu. To
accommodate this, we now copy over all of the host's prepopulated
context menu items to the Wine plugin host, and calling the targets
associated with any of those items will cause the target on the
associated context menu item on the host to be called.
This is slightly more complicated than what would otherwise be necessary
because Bitwig does not assign tags to their context menu items and
instead always uses 0.
I'm not a fan of Hungarian notation, but C++ kind of needs it with its
implicit `this`. And of all the common options for this, I find
suffixing members with an underscore the least offensive one.
As reported in #149, the DrumCore 3 plugin would segfault when trying to
drag files from it. This happened because the plugin presumably
underflows somewhere and then reports that it supports 4294967282
different drag-and-drop formats, even though yabridge asked for a
maximum of 16.
They aggressively use the message loop when parts of a plugin's UI
change, sometimes sending as many is 2300 events at once. The old 20
messages per tick limit would cause severe slowdowns in this case.
Since this was never stopped, the `watchdog_handler` thread would also
keep running. Since after e3f0926aef
everything is supposed to exit cleanly, this would cause group host
processes to hang and never exit. Tying the watchdog timer to
`MainContext::run()` is cleaner anyways.
This check was only needed because `operator[]` inserts an empty entry
if the variable doesn't exist. Wine also complains when `WINEPREFIX` is
empty, so we should probably not try to have our own behavior here.
...before terminating it forcefully. Not sure why this
`TerminateProcess()` was here instead of in `group-host.cpp` in the
first place. This way we don't have to duplicate any destructor
behaviour.
This would be useful if you need to do some elaborate namespacing setup
and want all sockets and other temporary files in a single directory
instead of in `${XDG_RUNTIME_DIR:/tmp}`.
This resolves#139.
Apparently these arguments are all passed as a space separated Win32
command line, so we need to do our own quoting. I was about to submit a
patch for this to Wine, but escaping escaped quotes sounds like a pain.
Robbert van der Helm