PeerBlock's tracelog functionality should store non-logged (i.e. too-highly
verbose) messages in e.g. a ringbuffer, so that they can be logged to
peerblock.log in the case where an error happens.
We have lots of messages, but don't want to log them by default. If an error
condition happens though, all these extra messages can be of much use in
helping to pinpoint the exact code path leading up to that error. So we should
store the last N messages (overwriting the oldest one once our buffer fills
up), and then if we hit an exception or other detectable error condition we can
log all the additional messages leading up to that problem.
This should help with troubleshooting problems, and minimize the requirement to
reproduce the problem with a higher tracelog verbosity level.
Original issue reported on code.google.com by peerbloc...@gmail.com on 23 Jul 2010 at 4:17
Original issue reported on code.google.com by
peerbloc...@gmail.comon 23 Jul 2010 at 4:17