|
Boost : |
From: Ani Taggu (ataggu_at_[hidden])
Date: 2002-04-23 17:08:11
I vote for acceptance -- with the understanding that the library will
continue to be refined with more user experience.
I recently did a porting (from Windows to Solaris) and the absence of a
portable date/time library was one of the many gotchas (threads was
another -- until I used the boost.threads version available then). The
original code used CTime/CTimeSpan Microsoft classes -- and I had to cook up
a primitive implementation based on time_t :(
I am no expert on date/time systems -- but right now, from available
documentation, the proposed library seem to support my day-to-day
requirements (and probably a fair majority of programmers). Even if GDTL was
less than perfect (which by my admission, I am not qualified to judge),
surely something is better than nothing :). The current discussion on GDTL
is interesting because it shows how difficult it is to design such a
library -- and its acceptance, while useful immediately -- will probably
trigger feedbacks with more and varied GDTL experience -- thus leading to
refinements later.
Regards,
-Ani Taggu
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk