3 ms·
This actually annoys me badly. One of the problems I see regularly is applications that fail to log enough of the stack to see what the entry point of the thing
by staofbur 9y ago
This actually annoys me badly. One of the problems I see regularly is applications that fail to log enough of the stack to see what the entry point of the thing that actually went wrong is because the syslog packet size is set to 512 bytes. The problem is clearly syslog then, not the 12KiB of stack your app throws when something goes pop!?!?
- andrewflnr 9y agoUm, yeah, the problem is someone setting an arbitrary limit because it was easier to implement. To argue otherwise is basically to claim that there is no possible justification for a deep stack, which, well, good luck proving that. Anyway, even if I don't like all that abstraction (I don't), I may not have a choice in platform (or logging system), so blaming useless syslog messages on the app developer is adding insult to injury. Not that blame is terribly useful when the best course of action is to just burn the whole thing down and start over. :)
- staofbur 9y agoThe arbitrary limit is a performance thing. Fits neatly in one UDP datagram and passes through everything unhampered by MTU etc. Agree that burning the whole thing down is the right thing when you get to this place :)
- Dylan16807 9y agoLogging the stack in what format? 500 bytes could fit as much as two hundred numeric entries, or as few as five fully detailed text entries.