Parsing issues

Since we have a new implementation now, based on completely different data structures, we have to take advantage and rethink some of the parsing steps we used in SDF to parse rules.

In the light of some recent bug reports, a few problems have emerged that we haven't really tackled in the old implementation:

1. Unique labels: #1839 Unrelated productions which had the same klabel, collided in the priority/associativity filter. We decided that we want to specify priorities external of blocks of syntax, by specifying tags or klabels: syntax priority '_*_ > '_+_. This ended up simplifying the KORE implementation but leaving out this small detail of having unique labels triggered some recurring bugs.

2. Overloaded productions: #1827 #1772