Brian Cuthie wrote:
Seems kind of unfair to bash Mark for asking here when you guys aren'tPersonally, I would prefer to just ignore things like this, because it is not sustainable for people to to just pick random lines of scripts, and ask "hey should I change this line", without even doing a little research into what the line is for.
even sure of what the answer is. This clearly isn't a -users question.
I think the big rule should be that if you haven't typed your request into google first, and looked there, you deserve to get ignored (and expect to be flamed) when you ask a question.
For example, if you had typed LD_ASSUME_KERNEL into google, you'd end up right at Ulrich Drepper's site, which would tell you all about how LD_ASSUME_KERNEL works, and how it basically switches thread implementations, and what NPTL is, and why some programs may not work well with it.
THEN, after doing some research, at the very least, when the question was posed, we'd all be able to have an intellegent discussion about this, and whether or not, or what would need to be done to ensure that asterisk was fully compatible with NPTL, and what impact that would have on performance.
As I recall, back from the pre-1.0 days, older asterisk versions would crash with NPTL, as it's much less tolerant of sloppy / erroneous programming than LinuxThreads; for example, with LinuxThreads, you could do pthread_mutex_t t = -1; [...] pthread_destroy(&t); and get away with it. With NPTL, you'd crash..
-SteveK
_______________________________________________ Asterisk-Dev mailing list [email protected] http://lists.digium.com/mailman/listinfo/asterisk-dev To UNSUBSCRIBE or update options visit: http://lists.digium.com/mailman/listinfo/asterisk-dev
