Re: bug with date_convert_universal()

On Mon, Mar 11, 2002 at 08:10:52PM +0100, Angel Fradejas wrote:
> Sorry for misleading you. The code in date_convert_universal is correct now.
> It was corrected in version 1.9 as you can see in CVS. It assumes 1-12
> input, as you correctly pointed out. The problem is the comment in date.h.

Umm, no.  The month is in range 0-11, and has always been until date.c
got broken in revision 1.9.  date.h says 0-11 in the comment,
date_parse_http() generates a month from 0 to 11, and 
date_convert_universal() expects a month from 0 to 11.  Where did you
see an inconsistency?

Currently it doesn't work for a 1-12 month range either, because leap
year calculation is wrong.

I recommend reversing the patch from revision 1.9, and then taking a look
at smsc_emi2 which apparently fills in a wrong value for the month field.
(Does anyone have specs for the EMI date format?)

-- 
Richard Braakman
Will write free software for money.
See http://www.xs4all.nl/~dark/resume.html