And why can't they reason the same way? They might have their own mutex like the interpreter mutex. Is it perhaps because they are embedded while Pike is not? But we'd like to make Pike embeddable, so if this specific issue is solved beforehand there'll be one less worry.
I can see no harm using localtime_r if it exists; if Xavier wants to fix that then it's ok with me.
/ Martin Stjernholm, Roxen IS
Previous text:
2003-06-17 19:47: Subject: Pike localtime(), threads safe and low level localtime() / localtime_r() ?
Nope. The only benefit would be protecting it from non pike threads (thus not holding the interpreter lock) using localtime at the same time. A rather questionable benefit though, since _those_ threads should really be using localtime_r anyway.
/ Marcus Comstedt (ACROSS) (Hail Ilpalazzo!)