Eric,

Eric S Fraga <e.fraga@ucl.ac.uk> writes:

> Guido Van Hoecke <guivho@gmail.com> writes:
>
> [...]
>
>> My ics file had a.o. my elder sister's birthday, and unfortunately her's
>> as well as mine is (way) before the start of the epoch, so mktime
>> returns a negative timestamp at line 63 and strftime at line 143 doesn't
>> grok it.
>
> Ah, I see.  The script should at least not break just because some of us
> (me included) are older than Unix... ;-)
>
> The fix should be straightforward.  Could you try adding the lines
>
>    if (timestamp < 0) timestamp = 0;
>
> after the call to mktime in the datetimestamp function?  This should at
> least make the script not crap out although obviously the date will be
> wrong (set to start of epoch).  If this works, I will update the script
> on Worg and put in a warning message of some sort, maybe even in the
> entry itself.

With this change the script assigns the wrong date, but no longer stops.

> Suggestions on how to handle this case would be welcome, of course.

I created a patch so that the date and time string is built directly
from the iCal data for times before the epoch. This function produces
valid date/time strings albeit without weekday info.

Patch file is attached.