Re: Fwd: (file)Re: Date disappearing from database for certain photos in 5.4.2 and 6.2.1 (though not a problem in 4.7.1)

On 7/13/20 4:30 PM, Johannes Zarl-Zierl wrote:
> Thinking about this a little more, maybe another option presents itself here:
> We could give the user a choice to (a) ignore the image timestamp (b) set a 
> fuzzy time (c) adjust the timestamp for DST on the image (d) adjust the 
> timestamp for DST on the image and all other images (or just on the same 
> day?).
> 
> Note: that's all options I could come up with - maybe presenting all options 
> to the user is actually worse than a sensible subset.

Auto-adjusting images after that point is fraught with peril, doo.  We have no way of knowing when
(or if) the user ever adjusted the time on the camera.  About the only possible way I could see of
maintaining sequencing would be to adjust every frame that was taken within an hour of a frame taken
within that window, and to do it recursively.

In other words:

02:30 -> 03:30
03:01 -> 04:01
04:00 -> 05:00

but if the next such time were then 05:01, that would not need adjustment (the time would likely
still be wrong, but the sequencing wouldn't).  But any of us could easily find corner cases where
that might not work; if someone's doing time lapse photography, they might get into a situation
where the time could never not be adjusted (well, until the next transition back to standard time).
That seems like awfully complex logic.

> My favorite option as a user would probably be (d), followed by (b).
> 
>> Setting the date in the database this way isn't really "corrupt", either. 
>> It's simply a limitation of how the time is expressed in the camera. 
>> Arguably the timestamps on the files shot after 3 AM are the ones that are
>> wrong -- they should be DST-adjusted.
> 
> "Corruption" may be a strong word here, but in essence it *is* timestamp 
> corruption if a series of images end up being sorted incorrectly.

Ignoring the image timestamp and using the file modification time is corrupting the date in that
way, too -- but potentially to a much greater degree.