Thanks for the feedback guys! I share the concerns brought up here and agree that this is no suitable feature for the fluidsynth library. I.e. the current proposal to make it a FluidSetting and expose a find_soundfont() function via the API is not suitable. IMO this should be implemented as exclusive feature for the executable. Preferably **not** as compile time option. A normal user who acutally benefits from this feature should not dependent on any compile flags a packager has or hasn't specified when compiling fluidsynth.
Keep in mind that synth.default-soundfont approaches a similar problem using a cmake flag. However I couldn't find any distribution (ubuntu, megeia, fedora, openSUSE) making use of this. I propose to keep it simple and implement this using an environment variable: If a soundfont specified on the commandline isn't found in the current working directory, the fluidsynth exec. looks through the $PATH variable and takes the first one it finds. The user only needs to setup up this env. var. in his ~/.bash.rc or similar. Ofc. we could also use a custom env. variable. I like $PATH because it's convenient on *nix and Windows. Also when implementing this, I suggest to remove synth.default-soundfont and possibly also the DEFAULT_SOUNDFONT cmake flag. After all, the current implementation only uses the default soundfont when the user plays back a MIDI file. Tom _______________________________________________ fluid-dev mailing list fluid-dev@nongnu.org https://lists.nongnu.org/mailman/listinfo/fluid-dev
