View Single Post
Old 03-24-2020, 09:24 PM   #3995
JimmXinu
Plugin Developer
JimmXinu ought to be getting tired of karma fortunes by now.JimmXinu ought to be getting tired of karma fortunes by now.JimmXinu ought to be getting tired of karma fortunes by now.JimmXinu ought to be getting tired of karma fortunes by now.JimmXinu ought to be getting tired of karma fortunes by now.JimmXinu ought to be getting tired of karma fortunes by now.JimmXinu ought to be getting tired of karma fortunes by now.JimmXinu ought to be getting tired of karma fortunes by now.JimmXinu ought to be getting tired of karma fortunes by now.JimmXinu ought to be getting tired of karma fortunes by now.JimmXinu ought to be getting tired of karma fortunes by now.
 
JimmXinu's Avatar
 
Posts: 7,024
Karma: 4604635
Join Date: Dec 2011
Location: Midwest USA
Device: Kobo Clara Colour running KOReader
New Test Version Posted

2020-03-24
- Make sure all timestamps are using Calibre's local_tz.

@Atherton25,

I've tried to follow your steps exactly, but I cannot get an error from doing a #fffcheck:<5daysago search with anything I try.

Are you talking about FFF's Last Checked Column feature? That is not the same thing as just setting a date column to the updated date metadata, which is what you appear to describe above.

Because I can get an astimezone() cannot be applied to a naive datetime error while trying to update a non-FFF story while Last Checked Column is enabled, but it only appears in the debug log where it is otherwise ignored.

It's very easy to argue that FFF shouldn't be setting anything for books that aren't from FFF, even if the user tries to. But there are cases when a book may have a had a valid story URL in past, but the site isn't supported anymore. And then I'd be left with only leaving out books with no URL at all, which is even more chaotic...

On one occasion while working on this, I was able to get Calibre to give me an astimezone() cannot be applied to a naive datetime error while editing metadata after trying various updates.

And now I cannot get that error trying to replicate it either way above. (sigh).

So what I've done is looked at the few places in FFF that wasn't already explicitly assigning the same timezone as Calibre (a few datetime.now() calls) and forced them to Calibre's TZ also.

That's my best guess, and it seems logical. So hopefully it will help.
JimmXinu is offline   Reply With Quote