On Wed, 4 Jun 2025 09:55:21 +0200 Kristian Amlie <krist...@amlie.name> wrote: > >I gave this a quick spin, and here are my findings: > >* It seems to work pretty well in general. As expected, I'd say, nothing >is better than that! > >* In particular, loading a patch from the UI produces the warning, but >loading via Program Change does not, which I believe is correct (you >have no way to click yes from your MIDI controller!)
Indeed! The same applies to Vector control and Midi-learn. The appropriate flags will be set while activating these, but not for the actual incoming MIDI >* There is some new, and quite verbose, log output. I assume this will >be removed. It will be retained but commented out. I'm using it to check that the right flags are being set/cleared. >* One thing to fix though: The warning is triggered when closing the LV2 >UI window. In this case it shouldn't warn, since closing the window is >not a destructive operation, and the host is expected to save the state >anyway. Good point. I'll fix that. There is a bit of a puzzle when looking at the CLI for control. I've not done any deep dives there for a very long time, and see that two heavily used functions, sendDirect (39) and sendNormal (119) have been moved entirely into a header file. I don't understand why this was done. These are quite substantial functions, and it surely massively increases the eventual code size, yet they are not critical for execution time so why was this done? These would be the obvious place for monitoring changes, otherwise it becomes a line-by-line examination :( -- Will J Godfrey _______________________________________________ Yoshimi-devel mailing list Yoshimi-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/yoshimi-devel