This has particularly been a problem due to how lexd handles combining diacritics, but in any case where an att file has a multichar symbol that isn't <[^<>]*>, lt-comp will add it to the alphabet but lt-proc will ignore it. I think the simplest solution is to have lt-comp split multichar symbols into multiple transitions (possibly triggered by a command-line option rather than always).
mr-martian
commented
2 years ago
This has particularly been a problem due to how lexd handles combining diacritics, but in any case where an att file has a multichar symbol that isn't
<[^<>]*>,lt-compwill add it to the alphabet butlt-procwill ignore it. I think the simplest solution is to havelt-compsplit multichar symbols into multiple transitions (possibly triggered by a command-line option rather than always).