goyacc?The popular yacc tool has always suffered from serious design problems, even though it has been the go-to tool for creating lexers/parsers in C for more than 30 years. I just ran into it again in goyacc, which enabled gojq to exist. I absolutely love that project and their decisions and the performance looks solid so far. So I’m facing something of an internal dilemma. I have to implement the parser for BonzaiMark and I can use either one. The thing with goyacc is that you don’t get an AST out of it (at least I don’t think). With PEGN I could. In fact, PEGN is closer to regular expressions than a highly optimized parser. I could, perhaps, make the first PEGN parser in goyacc and then use it to generate a better one in all Go, that is the advantage of PEGN is that any language can be represented in it (like Antlr in Java).
I’ve not done a full grammar in yacc ever, even though I once owned the yacc/bison O’Reilly book. But I have to admit, I’m feeling really obsessed with learning it and doing it. That same feeling I’ve had about other things in the past, and then it faded after I mastered it. I really want to do it once to see the advantages and disadvantages versus PEGN, EBNF, ABNF and others.
There’s just not enough time!! So many things to learn and play with in life. How people ever get bored or don’t have any ideas on what to do or make is completely beyond me. I’m actually tense now because I want to clear my plate to play with goyacc of all things. Most people will just think I’m one fucking sick puppy. Oh well.