Steffen Märcker wrote > I can send you that code. Awesome. I'll email you. Although, I wonder how relevant this is given your answers down below. Are there many grammars available in Xtreams syntax to make this useful? Also, what is the license? Can I add it somewhere on GH under MIT (with attribution of course)?
Steffen Märcker wrote > However, I moved away from Xtreams own PEG flavor to Bryan Ford's > original PEG syntax (close but more common) Ah, interesting. I didn't realize that Xtreams used a custom PEG syntax. I wonder why?! Steffen Märcker wrote > From Bryan Ford's original PEG syntax: > 1. Xtreams Grammer and Actor that build an Xtreams parser > ... > All three are available for VW in the Cincom's public repository: > - Xtreasm-Parsing (8.2-3,stm): +PEG parser +some fixes to Xtreams parser I'd certainly like to port that at some point, but I'm currently fairly mystified about the best practice process. I just reached out to Pavel about the Ring2 approach on which he spoke at ESUG. Do you have a documented process or any pointers even? Steffen Märcker wrote > And 2, 3 for Pharo on GitHub, e.g. > - https://github.com/kursjan/petitparser2/tree/master/PetitParser2-PEG Great. I use PP2 a lot. So if IIUC, I can now feed a PEG-syntax grammar string and have a PP2 parser generated for it? Steffen Märcker wrote > Speaking of, I just noticed that translator form Xtreams.PEG to PEG is > missing to complete the picture here. ;-) Ah, yes that makes sense. Any idea how much effort would be involved? Steffen Märcker wrote > As far as I know, there is no BNF-like parser generator for Xtreams, > available. > ... > Is it an option for you to convert your BNF to PEG manually? I'm not a parsing expert, so that may have been what I'm already doing and I'm using the wrong terminology. I took the ABNF from rfc5322 [1] and adapted it slightly [2] to be consumed by Xtreams [3] [OT?] As a final aside, I've been wondering if there's any way to generate "hand-rolled" equivalent parsers from Xtreams, PP, etc. for use cases where none of the libraries are available. I have in mind Pharo's MailMessage. It doesn't seem like any full-featured parsing libraries will be integrated any time soon, so the lowest levels use painful, duplication-riddled hand-rolled parsers. It would be great to leverage all this great library tech to create and reason about those... [/OT] Thanks for the discussion! 1. https://tools.ietf.org/html/rfc5322 2. https://github.com/seandenigris/Xtreams-Pharo/blob/master/repository/Xtreams-Email.package/PEGParser.extension/class/grammarEmail.st 3. https://github.com/seandenigris/Xtreams-Pharo/blob/master/repository/Xtreams-Email.package/PEGParserEmailTest.class/instance/setUp.st ----- Cheers, Sean -- Sent from: http://forum.world.st/Pharo-Smalltalk-Users-f1310670.html
