Re: PEG Parser

[Michael Lucy]

On Wed, Jan 26, 2011 at 8:38 PM, Noah Lavine <address@hidden> wrote:
> Hello,
>
>> I'm the guy that originally wrote this for GSOC, so I figured I'd jump
>> in.  I'd be happy to help with getting the PEG module merge-ready.
>
> Great!
>
>> keyword-flatten is described in api-peg.texi.  It's basically a
>> special case of context-flatten which collapses S-expressions
>> according to the symbol they start with.  From the documentation:
>> @deffn {Scheme Procedure} keyword-flatten terms lst
>> A less general form of @code{context-flatten}. Takes a list of
>> terminal atoms @code{terms} and flattens @var{lst} until all elements
>> are either atoms, or lists which have an atom from @code{terms} as
>> their first element.
>> @lisp
>> (keyword-flatten '(a b) '(c a b (a c) (b c) (c (b a) (c a)))) @result{}
>> (c a b (a c) (b c) c (b a) c a)
>> @end lisp
>
> Okay. I was confused about the behavior of keyword-flatten when the
> car of its argument is not in its list of keywords. I expected that to
> change, but it didn't. Looking at it now, is it true that the car of
> the argument never changes? So keyword-flatten flattens the sublists,
> but has to leave the overall one alone?

Yeah, the overall list is always there; only the sublists will ever be
collapsed down.  But the car can change.

It flattens until you have a list where each element either:
1. Is an atom.
or 2. Is a list whose first element is in the list of keywords.

So the car of the argument will change if it's a list that doesn't
start with the right keyword.  E.g.:
(keyword-flatten '(a) '((c (a b)) (a b) (b a))) ->
(c (a b) (a b) b a)

>
>> peg-string-compile is a function that will compile PEGs expressed as
>> strings into lambda expressions.  It does this by first parsing the
>> string using the PEG-parsing-PEG, then turning the output into an
>> S-expression representation of a PEG, compressing it, and passing the
>> compressed S-expression to peg-sexp-compile.  It's used if you e.g.
>> call peg-match with a string instead of an S-expression as the first
>> argument.
>
> I tried running peg-string-compile and then passing the result to
> eval, but I got an error because there was no symbol 'Begin. (capital
> B intended) Do you know what's going on with that? It might be
> something weird involving memoized symbols. Also, just to make sure I
> get it - this function produces Scheme code, right? (As opposed to
> Tree-IL or any of Guile's other languages.)

That's odd.  Could you paste in the lambda expression it generates?

(Yeah, it produces Scheme code.)

>
> Noah
>