stray line where I forgot to change "data" to "d" that should have been part of bf11107b

With the latest nw.js versions (0.17.x and later), this attribute prevents the
anti-aliasing of patch cords, selection rectangle etc., which looks really
bad. Commenting out the corresponding lines in pdgui.js reenables the
anti-aliasing.

fix bugs that were keeping the dmg package from building (leaving out the license for now because it relies on outdated Carbon libs)

just ship ladspa.h with the external since that header hasn't changed in forever. This makes it easier to build on OSX where we'd have to fetch it with wget/curl anyway

comsetic change... also, the previous commit may have technically been a bugfix. I don't think the bracket was supposed to be there without a leading forward slash

- Make the Ctrl+/ keybinding work on German keyboards. This actually adds a
  binding for the 7 key which has the / symbol on it as a shifted key on
  German keyboards. For some reason that key isn't properly mapped by nw.js,
  so we do it manually. We do this unconditionally, as there doesn't seem to
  be a portable way to check which keyboard is actually active.

- Add a secondary, more international-keyboard-friendly, Ctrl+, keybinding for
  the "dsp on" function. The comma works much better for this function on some
  European keyboards which have the / key on the first row of the keyboard, as
  it is next to the period on most European keyboards (the Latin ones at
  least). The intention here is to make it easy on a wider range of keyboards
  to quickly turn off dsp and then turn it on again using keyboard commands.
  The Ctrl+, keybinding should work well for that purpose with almost any
  Latin keyboard. But note that the Ctrl+/ keybinding remains to be available
  as well for backward compatibility.

NOTE: Right now these only work in canvas windows, we still need to add
something similar to the console windows to make it work there.

fix "stale pointer" bug when re-opening a subpatch on a scalar that has an event bound to it. The bug was that the scalar symbol wasn't getting unbound because scalar_vis wasn't getting called.

try removing the artifacts to see if it caused a regression with the Ubuntu runners hanging on exit (I can't debug it through gitlab because of its arbitrary limit on build log size... :(