Closed danielrh closed 8 years ago
@danielrh You're amazing, Daniel! The fix looks correct, and, yes, it's entirely reasonable to start with a fresh metablock when a new metablock header starts. I didn't do a good job at all testing for cases with multiple metablocks, and that debt's now asking to get paid. Thanks a lot for finding, reporting, and fixing this, Daniel!
Before only a few fields were cleared when a Metablock Header was encountered
However the btype_l may refer to an array of different dimension than the one for this metablock. This resets everything to the ground state.
Also added a test to demonstrate the behavior (test used to fail but now passes)