Re: [linux-audio-dev] Lock-free gtk/jack interaction

Jens M Andreasen wrote: Indeed. You exactly got what I mean. It then becomes more a design issue : the gui, based on gtk, is "shadow-capable" but I'm not using this ability. Additionally, redesigning things so that the gui properly "shadows" the sequence would oblige me to heavily optimize things. When adding a single track I currently destroy the whole workspace and query the sequence to so that the gui can redraw the whole thing, recreating all widgets. By redesigning the gui, I would keep many widgets as they are and only create some new ones, or destroy a few... This is very important, when you have a big workspace (ie: 1k beats, 16 tracks), because things then get very slow the way I currently do. I did think about this, but it required coding a lot of small functions for a variety specialized routines, and I would also need to change the way files get loaded (this last point is the main reason why I wanted to put the shadow within the sequence itself). Anyway, although it requires some work, I think this is the way to go. I hesitate here : would a such sequence-to-gui notification design be more souple/maintainable than a "shadowing" gui, for future, unplaned features ? I guess the two paradigms could coexist, if I ever need the RT thread to send signals to the gui... My jackd manpage states that the default frames/period setting is 1024, which means that for a frame rate of 44100 Hz, I'm around 44 periods by second, which is less than a monitor refresh. And what if some user needs to set that up to 4096 frames/period or more ? -- og