SSK at aaaaaaaaaaa,bbbbbbbbb/3c07a530-mag
>
> but, if I instead requested
>
> SSK at aaaaaaaaaaa,bbbbbbbbb/mag[5 days ago]
>
> then it would first ignore the parts inside [] and request the same DBR
> as before, but when interpreting the DBR data, it would use the time of
> five days ago rather than the current time. That way it wouldn't matter
> what weird baseline or wacky update schedule the person used, because it
> would calculate the key for 5 days ago the same way it would have
> calculated the key.
>
> Wouldn't this make everybody happy?
>
>
> <>
> --
>
> Oskar Sandberg
> oskar at freenetproject.org
>

I think it should...  I just wonder what units should be allowed.  How about
just seconds, minutes, hours, days, and weeks.  These are easy to define as
containing so many seconds.  I usually hate the idea of figuring out what a
month is and what a year is since these do not map easily back to a fixed
number of days and therefore seconds.

I must admit I miss the 0.3 DBR format, but I think overall the hex version
is better because it should have resulted in easier to maintain and
understand code resulting in fewer bugs.  I also like being able to do stuff
like this in bash:

../fcpget -htl 25 SSK@<publickey>/$(sh -c 'printf %x $[((`date +%s` - 86400
* DAYSAGO) / 86400) * 86400]')-<SiteName>

Ed


_______________________________________________
Devl mailing list
Devl at freenetproject.org
http://lists.freenetproject.org/mailman/listinfo/devl

Reply via email to