|
From: | David O'Toole |
Subject: | Re: Octal / Dynamic BGM / Indrema |
Date: | Tue Nov 28 16:21:02 2000 |
User-agent: | Mozilla/5.0 (X11; U; Linux 2.4.0-test7 i586; en-US; m18) Gecko/20001010 |
At first I thought, "a seperate daemon, that's a pain," but when I tried to pinpoint specifically why it wouldn't be as good as having the player inline, I realized I'd probably be forking the player _anyhow_. So that's quite alright.
Yes... the "core thread" of Octal needs to be pretty autonomous. The current dev version has the GTK gui in one thread (don't want pixmap engine causing audio dropouts :-) and the audio core in a high-priority thread, communicating thru pipes and shared memory. If I really wanted to get down and dirty with semaphores and locks I could, but again the idea is that to prevent audio dropouts the core thread can't block for anyone but the audio chip
Machines shouldn't pose a serious problem, since they'd be running inline with Octal, and not the game. At the very least, writing a machine is far simpler than writing a mod player =).
Right, this would only become a problem if someone wanted closed-source machines, which is unlikely anyway as a binary-only plugin wouldn't even work across different versions of the library
Thanks, I'll let you know about any new insights.
cool, keep me posted
[Prev in Thread] | Current Thread | [Next in Thread] |