Closed torrance closed 1 year ago
Good catch! See commit 642a6f6 for the fix.
I will roll this out into v0.2.1
Just to be clear, this wasn't an problem with the changes not being flushed or written back to disk. Instead, there was a bug in how scalar indexes were handled when calling deleteat()
.
Changes should otherwise be written to disk when the Table
object in garbage collected at the very latest, but in general they seem to be written to disk during the call. In the future, I'd like to write a flush()
method to make this process explicit if it needs be.
Reported by Iniyan Natarajan: