Time zones again

[Ray Osborn]

I guess you read the e-mail from Eric Boucher about building NAPI on a
Solaris.  Does anyone have any objection to making the current __MWERKS__
and WIN32 method of defining gmt_offset the default?  Unless there is a
clean way of finding out if tm_gmtoff is defined in a struct, this seems to
be the only ISO C-compliant way of doing things.  As I reported later, we do
have to call localtime again (or perhaps we should just call gmtime first),
because GNU C overwrites some of the struct fields.  It seems crazy to me
that you can assign a struct, only to have it overwritten when you assign a
second struct, but I've confirmed that there is a warning to this effect in
the GNU documentation.

Freddie, I think you defined most of the preprocessor checks in order to
cope with old versions of VMS.  Is it possible that all of these could be
simplified if we use the ISO C method?

Ray
-- 
Dr Ray Osborn                Tel: +1 (630) 252-9011
Materials Science Division   Fax: +1 (630) 252-7777
Argonne National Laboratory  E-mail: ROsborn at anl.gov
Argonne, IL 60439-4845