#44: High CPU usage on empty or bad playlists
-------------------------+--------------------------------------------------
Reporter: toots | Owner: admin
Type: Bugs | Status: new
Priority: 1 | Milestone:
Component: Liquidsoap | Version: 0.3.3
Resolution: | Keywords:
-------------------------+--------------------------------------------------
Comment (by toots):
You're right that's an old bug, but I though it worth to be discussed
here.
You're also right, it's a defect by design, and not easy to solve.
My idea is that each polling has to carry an heuristic for delaying the
next attempt. So I would propose to follow an increasing delay between two
tries.
The idea is simply that there's little hope that something that failed
once will work immediatly, just like icecast restart..
I don't know what is a correct function, but my proposition is to add a
sequence of delays, d_0 = 0, d_1, ..., d_n for each requested URI and wait
for d_i before try i.
--
Ticket URL: <http://savonet.rastageeks.org/ticket/44#comment:2>
Savonet <http://savonet.rastageeks.org/>
Let's program our stream !
