Summary
- merged
- fixing more messy proofs
- more ambitious parallelism (in contrast to a8ee8e4884ec): pri = 1 ensures that internal proof tasks are executed before the already forked theory outline with pri = 0;
- eliminated pointless special case (see also a8ee8e4884ec, c4c4c2f01723);
- proper datatype for 8-bit characters
- corrected nonsense
- less ambitious parallelism: avoid problems with HOL-Proofs and threads=2 (congestion with many thousands futures and rather dense heap);
- clarified modules;