On 29 October 2015 at 23:15, Pedro Côrte-Real <pe...@pedrocr.net> wrote:

> There was a bug fixed about time display recently so maybe that is
> messing with your time.
>

I think the bug referred to is http://redmine.darktable.org/issues/10650.
I am the author of the patch.

What was happening prior to that patch was that the date/time from the
database was being formatted with the aid of a couple of functions one of
which (mktime) was being told daylight savings time (DST) was not in effect
on the supplied date/time regardless of the particular date/time in
question.  That function works in local time and would consult the timezone
information and, if DST was in effect for the date/time concerned it would
then adjust the time to take account of that.

The effect of this was that some pictures would display a time which was an
hour ahead of the date/time stamp from the DB (which is the date/time stamp
taken from the EXIF).  The solution was to tell mktime we don't know about
DST on the input date/time.  That seems to make mktime assume the input is
in correct local time (including DST where appropriate).

Unfortunately the muddle comes from the EXIF designers not designing for
image collections that span time zones and seasons.  EXIF, it seems, does
not record the timezone neither does it declare if the time is in local
time or UTC and, if local time, whether the clock on the camera includes
DST or is set to the local "standard time" even in summer.

At the moment, I think, darktable tries to treat dates/time as best it can
without knowing the timezone, i.e. it tries to be transparent but perhaps
that does not always work.

Personally I would prefer that darktable worked internally in UTC and did
conversions on import/export but this would mean import would need to know
the timezone of the EXIF date/time stamp.  This would almost certainly mean
there would have to be an option on import to select UTC or local time and
then maybe a list of time zones which defaults to the one from the
environment darktable is running in.  For the specific case the EXIF is in
local time and the image has a geotag DT could probably deduce the
timezone, though there may be some interesting boundary cases.

Regards,
Steve.
------------------------------------------------------------------------------
_______________________________________________
Darktable-users mailing list
Darktable-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/darktable-users

Reply via email to