clarified priority of "skipped" proofs, which might take long but do not produce relevant information (potential conflict of quick interactive feedback vs. performance in batch mode);

simplified / clarified execution priority: auto prints << 0, proofs < 0, eval = 0, print_state = 1;

paint unassigned/unchanged nodes as unprocessed -- relevant for editor_continuous_checking = false;

allow explicit indication of required node: full eval, no prints;

simplified flag for continuous checking: avoid GUI complexity and slow checking of all theories (including prints);

more work on (co)datatype docs

more (co)datatype documentation

merged

tuned proofs;

more uniform border;

proper PIDE markup for codegen arguments;

less aggressive flushing: cope with massive amounts of protocol messages, e.g. from threads_trace; propagate flush from ML to Scala via special protocol message; more formal type Message_Channel.message;

tuned -- more uniform ML vs. Scala;

recovered delay for Document.start_execution (see also 627fb639a2d9), which potentially improves throughput when many consecutive edits arrive;

more timing;

more timing; tuned -- Execution.is_running always holds due to immediate start;

removed spurious headings

more (co)datatype documentation

avoid DUP error in local context

sketched documentation for new (co)datatype package

tuned docs (the function package isn't so new anymore)

tuned comments;

tuned;

type theory is purely value-oriented;

pro-forma Execution.reset, despite lack of final join/commit;

tuned signature; removed notoriously outdated comments;

tuned;

de-assign execs that were not registered as running yet -- observe change of perspective more thoroughly;

merged

tuned intro