[Media_api] Brief PortMidi Question

[cpr at musetrap.com]

back in win98 days (while working at opcode), I ran into this exact problem, and
had to write code to deal with it.. I had hoped it had become a non-issue in
the xp days, and had pulled this code out of my personal midi library.. but,
maybe we still need to think about it..

-cpr 

Quoting Roger Dannenberg <rbd at cs.cmu.edu>:

> Adam,
>     That code is in there because there were some claims and some 
> uncertainty about whether Microsoft's code could possibly issue a 
> callback while another was in progress. The feeling was no, it won't 
> happen, but if there *were* concurrent callbacks on the same device, it 
> would cause problems, so some checks were added. It seems plausible that 
> there could be concurrent callbacks on *different* devices, and it 
> sounds like a bug that this is tripping the assert().
>     I want to look at the code, because I don't think anyone has really 
> considered whether the code is safe for multiple processors concurrently 
> handling MIDI callbacks. In the meantime, you could either read the code 
> too or just take out the entry count and assert() code and see what happens.
>     I'll report back when I've taken a closer look.
>     Thanks for the report -- it's a real issue that needs some attention.
>     -Roger
> 
> _______________________________________________
> media_api mailing list
> media_api at create.ucsb.edu
> http://lists.create.ucsb.edu/mailman/listinfo/media_api
> 




----------------------------------------------------------------
This message was sent using IMP, the Internet Messaging Program.