The binary KORE pipeline still has one step where textual KORE is required - parsing a program from surface syntax into a KORE term requires the use of --gen-glr-bison-parser and the corresponding parser_SORT executable.
Currently, this isn't a big bottleneck (as the motivating example from the C semantics comes from the concatenation of several individual terms into a single translation unit). However, if a semantics were to generate a particularly huge individual program term, parsing the textual KORE into binary after the initial parsing step would be a problem.
The binary KORE pipeline still has one step where textual KORE is required - parsing a program from surface syntax into a KORE term requires the use of
--gen-glr-bison-parser
and the correspondingparser_SORT
executable.Currently, this isn't a big bottleneck (as the motivating example from the C semantics comes from the concatenation of several individual terms into a single translation unit). However, if a semantics were to generate a particularly huge individual program term, parsing the textual KORE into binary after the initial parsing step would be a problem.