Deletable/Overwritable Cocouncil End Frame
Derived from this post. User reported that attempts to add a frame after the cross-examination resulted in adding the frame to the co-council conversation. After repeating, the entire conversation disappeared. Attempts to add a frame to the co-council conversation resulted in pushing frames up and out of the entire cross-examination. Several blank frames were created and then deleted. The user saved, and found the trial contained several frames with a null ID, rendering the trial inoperable. Loading of a prior backup showed that frame 62, the cocouncil_end_conv and end frame, did not exist. Restoring frame 62 appears to have made the backup usable again.
It is assumed that the root cause of this sequence was loss of frame #62. Copies of trial data from after saving (broyito_bug_original) and from loading the backup (broyito_bug_backup_loaded) are included.
Comments (7)
-
-
reporter - attached mock_trial_data_bugged.txt
Included a fairly minimal copy of valid trial data afflicted from the glitch as mock_trial_data_bugged.txt
-
reporter Still there, but now glitchier. I ran the same reproduction steps as before (open the CE statement, then create a new statement to the right) and got the following:
I've included a 9-frame trial that demonstrated the bug in the report.
-
-
assigned issue to
- marked as critical
-
assigned issue to
-
Commit eefc818 apparently fixes the issue here... Though I'm not entirely confident about it in complex situations. Hopefully it does not create other error cases.
-
reporter I've played around with this in the editor and haven't found any new problems. Given the complexity of the code and how long it took to find the last bug, other errors are still very possible.
I assume that we'll keep this open until the rewrite, just in case?
-
- changed status to resolved
Bug no longer reproducible with latest fixes.
Upcorming rewrite of the row map mechanism should however take this kind of issues into account and be better designed for it.
- Log in to comment
Enth, since you were able to reproduce the issue, can you check if the fix in commit 7af368a fixed it ?