Using the standard PEG syntax sure would be nice.

[rlpowell]

I am, to be clear, not actually expecting anything to happen here, just thought I'd express a wish.

As I find myself writing a Pest parser to parse the original PEG grammar syntax (i.e. the one in https://bford.info/pub/lang/peg.pdf ) so that I can port what is probably the largest PEG grammar ever written (it's in two parts, https://mw.lojban.org/papri/BPFK_Section:_PEG_Morphology_Algorithm and https://www.digitalkingdom.org/~rlpowell/hobbies/lojban/grammar/lojban.peg.txt ) into Pest's format, I find myself wishing that I didn't have to do that, that Pest used a syntax that mostly aligned with the original PEG syntax.

This is, to be clear, not remotely unique to Pest! Almost all the Rust PEG parsers have their own idiosyncratic syntax. But since there's a clear mapping directly between Pest's syntax and the original, it just seems wasteful to have to convert between the two.

Thank you for coming to my TED talk. :)

[tomtau]

Not sure if this is actionable in this repo directly. Unless I'm mistaken, the differences are:

• rules are assigned with <- instead of =
• rule definitions don't have braces
• comments start with ; instead of //
• the choice operator is / instead of |
• there's no sequence operator (it's implicit / from white spaces)

I was thinking whether it could be made to implement a potential feature in pest3: for example, someone was proposing a feature to be able to mark some tokens as parameters and be able to instantiate them at runtime. That could cover all cases (rules don't need braces in pest3), except for the last one (unless one could replace the sequence operator as an empty string or something)