Corruption of simple tree made from scratch in 5.2.1

[emyoulation]

While creating a simple Tree for experimenting with Godfather/Godchild Association, I managed to corrupt it. (Before actually getting to the real experimenting.)

I changed loaded trees to look up someone in a real tree. But trying to reload this tiny tree failed. And the recommended Repair found no problems. Now I cannot get far enough to export the data! The context menus don’t get built… so there’s no Export option.

A bug has been filed with the database folder as a ZIP attachment.

https://gramps-project.org/bugs/view.php?id=13276

System Information:

Gramps version: 5.2.1
Python version: 3.11.8
BSDDB version: not found
sqlite version: 3.42.0 (2.6.0)
LANG: en_US.UTF-8
OS: Linux
Distribution: 6.5.12-100.fc37.x86_64

GTK version : 3.24.41
gobject version: 3.46.0
cairo version : (1, 26, 0)

[Nick-Hall]

I can load the database. I don’t see any problems.

[emyoulation]

After loading 2 different trees, the problem is now resolved.

The next tree (which has run with no problem yesterday) reported that Gramps had a problem the last time it was loaded. That ‘Repair’ recovered whatever was wrong (preferences???) and opened that tree. Then loading the next tree reported no problem at all.

Then loading the small tree reported that there was a missing handle, It recovered from that and now there is no problem.

gramps.gen.errors.HandleError: Handle f971455fde919c6e895412e9fa not found

There is no Gramps52.log (or any .log file when searching the /home/districtsupport/.gramps folder.

[emyoulation]

Considering the behavior suggests a possibility.

I deleted a two people using the Toolbar - button before switching to another Tree. Could the ‘remember last active view’ feature have been the problem?

Maybe the saved preferences or Go history were pointing at the handle of the deleted person … or the historically previous (also deleted) person?

[ennoborg]

Switching trees triggers a similar problem here too, sometimes, in 5.1.