Hello jmerrill,

jmc> There is an absolute standard for this -- the timezone that's in
jmc> an internet-style (date and) time string says that earlier times
jmc> are a negative offset from GMT.  (For example, the U.S. has a negative 
offset from GMT.)

jmc> I think you did it backwards.  It must match the internet time
jmc> standard -- when there are only two choices, our "some standard"
jmc> should be the one that has already been chosen by the world.

Thanks for the input, there is a big point in what you're saying.
However, another though came to my mind: one of the primary purposes
for the method suggested is to be able to get GMT time from local one
easily. Then my current idea makes sense as gmtime = localtime +
gmtoffset - simple & easy. However, still open to changing, and I
believe Nicolas should have a final word here.. ;)

Mike

jmc> J. Merrill
jmc> Senior Software Engineer
jmc> 3M Health Information Systems, Inc.



jmc> Michael Pliskin <[EMAIL PROTECTED]> 
jmc> Sent by: [EMAIL PROTECTED] 
jmc> 07/14/2008 12:02 PM 
jmc> Please respond to
jmc> Michael Pliskin <[EMAIL PROTECTED]>; Please respond to
jmc> Neko intermediate language mailing list <[email protected]>

jmc> To
jmc> Neko intermediate language mailing list <[email protected]> 
jmc> cc
jmc> Subject
jmc> Re: [Neko] API for local/GMT time conversion





jmc> Hello Nicolas,

MP>> Hmm then we're reducing the solution to just get_tz_offset function -
MP>> minutes seem non-standard for me, maybe seconds is better. Ok, will
MP>> make it and send a patch soon.

jmc> ok, here is a simple patch. Note that I am subtracting local from GMT,
jmc> so GMT+2 will have negative offset. Not sure if it is the best but we
jmc> just need to stick to some convention.




-- 
Best regards,
 Michael                            mailto:[EMAIL PROTECTED]


-- 
Neko : One VM to run them all
(http://nekovm.org)

Reply via email to