On 01/08/2014 10:30 AM, Keith Winston wrote:
well, fair enough. Generally, the issue involves instances when Python will come back, but it might take several minutes or much longer. And weird behaviour ensues: like timers I put on the program don't actually time the amount of time it's busy (by a very, very large margin). Also, often several minutes after such a thing (sometimes quite a bit later), things will suddenly start working quickly again. Also, on my laptop I can actually tell when it's struggling, b/c the fan turns on and/or speeds up, and in many of these cases it will go into all-out mode, even though I'm not doing anything. But your point about having no lever with which to move the world is a good one.
It smells like a logical bug, or lack of simplicity / clarity, leading to a behaviour with "exponential" cost (in cpu usage, thus time, more or less in proportion), in some given circumstances. Or maybe just an unavoidable exponential cost routine, but the quantity of iput shoulde remain small, while in some (buggy) cases it is big.
[Just wildly guessing, don't know your app & code. Can you reduce it to a minimal still manifesting similar bug?]
Denis _______________________________________________ Tutor maillist - Tutor@python.org To unsubscribe or change subscription options: https://mail.python.org/mailman/listinfo/tutor
