Upcoming breaking file format changes

[Tobias Leupold]

E-Mail von Robert Krawitz via KPhotoAlbum vom Donnerstag, 24. April 2025, 17:41:
> On 4/24/25 06:40, Tobias Leupold wrote:
> > E-Mail von Robert Krawitz via KPhotoAlbum vom Donnerstag, 24. April 2025, 02:48:
> >> On 4/22/25 16:39, Johannes Zarl-Zierl wrote:
> >>> Hi!
> >>>
> >>> TL/DR: upcoming index.xml changes are not backwards compatible. Have a backup 
> >>> ready and see if you can find any bugs.
> >>
> >> This will certainly mean that kpa-util will need to be fixed...
> >>
> >> I've been through a lot and have had no time or energy to devote to hacking.  I lost my wife to
> >> cancer about 3 months ago; she was diagnosed about 3 years ago, and needless to say that's where all
> >> my cycles have gone...
> >
> > Hi Robert,
> > 
> > those are definitely not the news one wants to hear. My sincerest condolences!
> > This of course makes it fully understandable why we didn't hear much from you
> > the last time ... I wish you all the strength you need and all the best for
> > what may come!
> 
> Yes, the past 3 years have been extremely difficult.  Her needs were definitely my priority.
> Healing's going to be a slow process.
> 
> > Apart from that, kpa-util will defintely have to be fixed, as the way category
> > names are escaped to be used as XML attributes for the "compressed" format is
> > different now, plus all categories now have IDs (the attribute names are now
> > simply "tags_<category ID>"). Also, all area data is now stored within the
> > tags ID list, like so: "<ID>+a=<x y w h>,...".
> 
> That's going to be a non-trivial amount of work.  At least for the next month or so, I have no hope
> of getting to it (my wife's Celebration of Life will be next month, and we're also trying to
> establish an award in her honor, and a whole bunch more).

Fixing stuff in KPA is for sure not priority. Take your time!

> > It was quite simple to implement this, so I suppose it won't be too hard to
> > add this to the Perl code (for now, we could as well simply let the script
> > refuse to work on v11 databases).
>