Re: [Duplicity-talk] Losing incremental backups

[Greg Hartman]

My first patch didn't completely solve the problem.
Here is a second patch that ensures that

stringtotime(timetostring(t)) == t

The new code compensates for a number of cases,
including the fact that 1 AM will happen twice on days
like 11/4 in the US when DST kicks out. It also
compensates when the timezone of the system parsing
date strings doesn't match time timezone of the system
writing time.

In some cases timetostring(stringtotime(s)) may not be
equal to s. This happens if the original time was in a
different timezone or if it was written by an earlier
version of duplicity that wrongly assumed that the
alternate timezone applied. However this mismatch
doesn't create a problem since duplicity uses the
number of seconds from the epoch to compare times
rather than just comparing time strings.

I tried to put some comments in the patch to explain
in more detail.

Greg


      
____________________________________________________________________________________
Get easy, one-click access to your favorites. 
Make Yahoo! your homepage.
http://www.yahoo.com/r/hs

patch
Description: 4046725619-patch