For the speed in offering labels, I think there may be an easy way to cut this in half (we use beebasm's two pass scan of the code to pick up labels and symbols first, then check everything can be resolved on the second pass), so should be able to provide the labels after the first pass. Longer term may be able to consider incremental parsing of just changes but that's more of a fundamental change.
For the speed in offering labels, I think there may be an easy way to cut this in half (we use beebasm's two pass scan of the code to pick up labels and symbols first, then check everything can be resolved on the second pass), so should be able to provide the labels after the first pass. Longer term may be able to consider incremental parsing of just changes but that's more of a fundamental change.