Darren J Moffat wrote: > Why isn't the implementation of this using the port_associate(3C) > facility to watch for the default timezone file being updated ?
localtime(3C) is very performance sensitive. We don't want to make any extra system calls or threads interactions from the function. > Using port_associate(3C) means it is possible to update per process > without requiring the admin to run a (privileged) command. I really > don't like the idea of the requirement of the privileged command. Is > this because it has patch binding and port_associate(3C) isn't available > in all releases this is intended to be patched to ? If so I would like > to see that tzreload is "Obsolete Committed" and the OpenSolaris/ONNV > release of this use port_associate(3C) instead of the method proposed. > > The case says ctime(3C). Isn't it really calling of > localtime_r(3C),tzset(3C) and mktime(3C) that causes the update to the > cached TZ ?, ie those functions in libc that call the static > getsystemTZ() function. All functions listed in the ctime(3C) except for gmtime* and asctime* will update the cached timezone information. -Nakano
