Comments on: Python time.timezone / time.altzone edge case http://blogs.gnome.org/jamesh/2006/12/31/python-timetimezone-timealtzone-edge-case/ Random stuff Wed, 28 Oct 2009 02:53:32 +0800 http://wordpress.org/?v=2.8.6 hourly 1 By: mathew http://blogs.gnome.org/jamesh/2006/12/31/python-timetimezone-timealtzone-edge-case/comment-page-1/#comment-353 mathew Tue, 30 Nov 1999 00:00:00 +0000 http://blogs.gnome.org/jamesh/2006/12/31/python-timetimezone-timealtzone-edge-case/#comment-353 "Bazaar stores commit dates as a standard UNIX seconds since epoch value plus a time zone offset in seconds."<p/>Well, that's useful to know. Specifically, I now know never to go near bzr, as it's obviously designed by complete idiots.<p/>The only sensible way to record timestamps is to store them in UTC, and make converting to and from display format and display time zone be a client function. “Bazaar stores commit dates as a standard UNIX seconds since epoch value plus a time zone offset in seconds.”

Well, that’s useful to know. Specifically, I now know never to go near bzr, as it’s obviously designed by complete idiots.

The only sensible way to record timestamps is to store them in UTC, and make converting to and from display format and display time zone be a client function.

]]> By: James Henstridge http://blogs.gnome.org/jamesh/2006/12/31/python-timetimezone-timealtzone-edge-case/comment-page-1/#comment-354 James Henstridge Tue, 30 Nov 1999 00:00:00 +0000 http://blogs.gnome.org/jamesh/2006/12/31/python-timetimezone-timealtzone-edge-case/#comment-354 Matthew: I guess I was a bit ambiguous in the post above. Bazaar stores a "seconds since epoch field" plus a "time zone offset field". If you are using bzrlib, these are revision.timestamp and revision.timezone respectively.<p/>It was recording the UTC time correctly for my commits, but it was determining an incorrect time zone offset, and that problem seems to be caused by easily misused standard library features. Matthew: I guess I was a bit ambiguous in the post above. Bazaar stores a “seconds since epoch field” plus a “time zone offset field”. If you are using bzrlib, these are revision.timestamp and revision.timezone respectively.

It was recording the UTC time correctly for my commits, but it was determining an incorrect time zone offset, and that problem seems to be caused by easily misused standard library features.

]]>