Conversation
Q.U.I.N.N.
icedquinn@blob.cat
maybe tomorrow see about implementing the pattern match paper 
although it looks like in practice most people just do a very direct translation from match arms to if/else trees. some of the papers were like "we can get around backtracking all you have to do is
"
and its like yes cool i love finite state automata in my codebase i like 800 lines of _state = 80086; goto fhqwgads;
although it looks like in practice most people just do a very direct translation from match arms to if/else trees. some of the papers were like "we can get around backtracking all you have to do is
"and its like yes cool i love finite state automata in my codebase i like 800 lines of _state = 80086; goto fhqwgads;
2
0
1
elizabeth worm🔅
lizzie@brain.worm.pink
@icedquinn tbh i just assumed they all used finite state automata
1
0
0
Q.U.I.N.N.
icedquinn@blob.cat
@lizzie the simple backtrack strategy makes the most reasonable code (but it might test stuff repeatedly.) i tend to use that in parsing because PEGs were the only things that made sense for so long (and i don't use the packrat memoize table, i just backtrack, which is fine as long as your grammar is reasonable.)
DFA is like. okay, you avoided potentially testing twice. you just have this O(NY^2) mini compiler that generates a 126 state monstrocity instead
DFA is like. okay, you avoided potentially testing twice. you just have this O(NY^2) mini compiler that generates a 126 state monstrocity instead
0
0
1