TZ-related changes in Draft 5 of the next POSIX standard

Garrett Wollman wollman at
Tue Jan 9 19:08:19 UTC 2001

<<On Tue, 9 Jan 2001 12:26:23 -0500 (EST), Jonathan Lennox <lennox at> said:

> Is there any hope of their adding timegm() as well?

I doubt it.  There are already problems with the existing definitions,
and adding such an interface would probably make it even worse.  One
of the conflicting intents of the specification is that one could
implement timegm() as a simple formula, an example of which is given
in the ``Base Definitions'' under ``Seconds Since the Epoch''.  That
definition has been the source of much conflict over the history of
the POSIX effort, and much of the traffic on the austin-group
mailing-list last month arose from the same issue.

I am told there are not supposed to be any new interfaces added at
this point in the process, even if consensus could be reached.


More information about the tz mailing list