>>>>> "KS" == K Stol <[EMAIL PROTECTED]> writes:
KS> Is it true that there will be operations for generating an event? KS> So, unlike interrupt signals, these events are generated 'inside' KS> the system. So, will there be an op like "postevent" or something? i expect it will be supported. all full event systems have what i call a plain or user level event. there may even be priority levels. this is very easy to do. dan has said that at critical times all high priority events (async i/o, timers, etc) will be flushed from the pending queue. when a special op is called (inserted by compilers), lower level events can be processed and flushed from the queue. those would include user events. i would expect the priorities would be simply to just flush higher ones and then lower ones. how many levels is undecided but only a few should be needed. enqueuing a user event should be simple. you need to pass in some event id or private data (an object pmc? code ref?) for the callback. this op code allocates an event queue thingy, stuffs it with the provided user data (and priority level) and enqueues it . it will eventually get triggered/callback'ed when that queue gets flushed by the special op code. uri -- Uri Guttman ------ [EMAIL PROTECTED] -------- http://www.stemsystems.com --Perl Consulting, Stem Development, Systems Architecture, Design and Coding- Search or Offer Perl Jobs ---------------------------- http://jobs.perl.org
