I rediscovered this thread while pondering an (existential?!) problem: Why do we keep having to write, and rewrite, and rewrite, ad infinitum, parsers for well known domains like rfc specs? The parser world in many ways feels like a modern post-"Tower of Babel". I was really excited about the reverse parsing stuff done in Squeak by Ted Kaehler and Alessandro Warth [1], which never seemed to get picked up.
Steffen Märcker wrote > I wrote a PetitParserGenerator that takes the DSL and builds a > PetitParser. I don't know how I could've missed this gem! I hope Steffen is still subscribed. I googled "PetitParserGenerator", but only found these ML posts :/ IIRC Xtreams can take a BNF and generate a parser. I was thinking about implementing a BNF parser in PetitParser, but would love to avoid that, wouldn't mind a two step BNF -> Xtreams -> PP process. Although there was this SO reply [2] where Lukas said that one can't necessarily blindly feed a BNF to a PEG. It also got me thinking about sharing parsers between PP and PP2, since in many cases it seems that only the internal class names would differ... [1] http://www.vpri.org/pdf/m2008001_parseback.pdf [2] https://stackoverflow.com/a/9443024/424245 ----- Cheers, Sean -- Sent from: http://forum.world.st/Pharo-Smalltalk-Users-f1310670.html
