The reason is that apr_time_exp_* functions explode and implode the same epoch as unix (1/1/70 midnight). As with all Unix time functions, months are counted from 0 (Jan).
This assures that time constants are portable from platform to platform. Bill At 07:05 AM 2/11/2003, Erwin De Wolff wrote: >Using platform windows XP, APR library version 0.9.2. > >I wonder if it is correct to get the result: > >apr_time_exp_tz(&test, 0, 0); >the month 0, (zero? not 01 ?!?) >year 70 (I understand that) >day 1 >etc. >It seems that it is the number of microseconds since 01-00-1970 00:00:00 GMT >(month 00). Is that on purpose done or a bug? >On the other hand 01-01-1970 00:00:00 GMT and implode it gives 2678400000000 >which is 31 one days in microseconds. So it is consistent at least. > >If it is meant like this I can easily correct it, but I don't see a reason >for this. > >Greetings > >Erwin de Wolff
