masks (path, brush): use NAN as sentinel instead of magic number
the should partially fix #10805
#2 Updated by Ulrich Pegelow over 4 years ago
- System changed from Debian to all
- % Done changed from 0 to 10
- Status changed from New to Confirmed
I don't get a crash but something is certainly broken here. In my case darktable even fails to read the mask shapes from the XMP file -> no masks in the impage after import. If I use the "load sidecar file" option 'darktable -d sql' reveals that no "insert into masks ..." statement is prepared.
#3 Updated by Ulrich Pegelow over 4 years ago
First things first, the non-reading of the masks from the XMP file seems to be caused by one of the shapes having an empty name: mask_name is set to <rdf:li/>.
Once I fixed it manually the original bug report is confirmed and everything wreaks havoc. darktable runs endlessly, eating up all available system memory and finally crashes.
#6 Updated by Jean-Luc Coulon over 4 years ago
Well I was trying/testing everything.
The masks were in the circuit.
I remarked quite many strange things with the masks. For some things in the mask UI, I was replied the developper is no more active. So I supposed it was the same for all the things related to masks... And I left that...
I'm rebuilding dt. I will play more with the masks :)
#7 Updated by Ulrich Pegelow over 4 years ago
- % Done changed from 50 to 100
- Status changed from In Progress to Fixed
Applied in changeset darktable|8298829cfbc41f8adbc6b07c0abee34a023a37bc.