Kevin Falcone on Sun, 19 Nov 2000 23:19:15 -0500 (EST) |
>>>>> "Md" == MaD dUCK <madduck@madduck.net> writes: Md> but would i then not be filling up the FIFO to the point of no Md> return??? how is a FIFO stored? in memory? /usr/src/linux/fs/fifo.c for the gory details Md> it would be nice to have a FIFO that could act as a "packet Md> buffer" of length 1, where i define a packet to be something Md> like the output from one invocation. sort of like a super market Md> shelf with capacity one: customer service puts product in, then Md> only as soon as customer removes the product, a new one is Md> placed in. >From purely empirical evidence, it appears to be a buffer holding one invocation. The limit is how fast your daemon can keep refilling it. Running sigrand, and then cat'ing .signature means that it will block if sigrand hasn't refilled the FIFO. Decreasing the time between loops (or blocking on writing to the pipe) would result in virtually asynch filling. Also, sigrand used to run for days, trying every second to write to the FIFO, and it didn't fill up ram, so somehow I doubt that it is caching all of the prints. -kevin -- "I sense much NT in you. NT leads to Blue Screen, Blue Screen leads to downtime, downtime leads to suffering. NT is the path to the Dark Side." ______________________________________________________________________ Philadelphia Linux Users Group - http://www.phillylinux.org Announcements-http://lists.phillylinux.org/mail/listinfo/plug-announce General Discussion - http://lists.phillylinux.org/mail/listinfo/plug
|
|