Re: Parsing with Perl 6

[Juerd]

Wow, very impressive. ++!

// <-[\n]>*:

I'd really want to take advantage of the new \N:

    // \N*

foreach my $function (@functions)

In Perl6, foreach will just be for. The C-style for loop will be called loop. And the for syntax will slightly change:

for @functions -> my $function {
    ...
}

rule quoted_string ($type)
{
    $type
    [
        <-[$type]>*:
        |
        [ <after \\ > $type ]
    ]*
    $type
}

Why not already have both types there?

rule quoted_string {
    $delimiter := ( <['"]> )
    [
        <!before $delimiter> .
        |
        <after \\> $delimiter
    ]*
    $delimiter
}

Hmm - In current regex stuff, I'd probably have written: /(['"]) (?: (?! \\ | \1) . | \\ . )* \1/x. Why not use something like that? (Or is it just a matter of taste?)

rule quoted_string {
    $delimiter := ( <['"]> )
    [
        <!before \\ | $delimiter> .
        |
        \\ .
    ]
    $delimiter
}

- Yes, I reinvent wheels.
- Spam: Visit eurotraQ.

[jryan]

Ah, you are completely right about \N, of course. I'm just a blockhead ;(

Concerning quotes: I think that the way I structured mine is easier to convert from logic. For instance, I defined a quoted string as:

- start quote
- string of (non-quote or backslashed quote) characters
- end quote

Which I then translated into (comments added this time):

rule quoted_string
{
    " # start quote

    [ # string
    
        <-["]>+: # non quote characters
        | # or 
        [ <after \\ > " ] # backslashed quote character

    ]* # end string

    " # end quote
}

It was then shown how to make it "dynamic" by simply adding an argument and subsituting that argument in for the quote delimeter. While yours is much more elegant and efficient, I thought that my example was easier to follow in the current context.