UUID matching

[leaanthony]

Hey Alec! Awesome project and easily the best parser library I've used. Just wandering the best way to parse a UUID? The default lexer is perfect for my use case, apart from this one thing. Is it possible to extend the default lexer or would I need to recreate the default lexer and add my specific rule? Cheers!
EDIT: I've tried consuming all tokens in a greedy fashion and this worked great until an 'e' was present, then I got a "exponent has no digits" error, which appears to be an error thrown by the Go scanner.

[alecthomas]

Glad you enjoy using it :)

There's no reasonable way to extend it, but you can construct one via an existing text/scanner. There may be a way to configure a text/scanner to do this? If not you'll need to roll your own, which is not too hard.

[alecthomas]

Check out the _examples folders for some custom lexers.

[leaanthony]

Thanks for the quick reply Alec. Is the default lexer available for extending? I'll be honest, I've no idea where to start to recreate it :-)

[alecthomas]

You want this function. It depends on whether the text/scanner can be modified to do what you want. It has some configurability.

[leaanthony]

Thanks. I'm happy to just greedily consume the uuid and tried using the @~ syntax but I'm getting unexpected token ~.
Thing to parse: uuid(76eceacb-0462-4d94-ac83-423a6793775e)
Current attempt: 'uuid' '(' @~')' ')'

[alecthomas]

If ~ is a literal you will need to quote it the same way you've done with uuid, ( and ), eg. @'~'

[alecthomas]

Aah. You're probably on a tagged release. Use ! instead for now.

[alecthomas]

Yeah sorry about that, v2 is still in alpha so I'm still changing a few things. This construct was originally !<expr> but this turned out to be ambiguous so I had to change it to ~. The README reflects this change but I haven't tagged a new alphs release yet. Details in the original commit (5c4f519).

You could also just update to @master and ~ should work.

[leaanthony]

Thanks Alec. I'll try that 👍

[leaanthony]

Looks like it doesn't work.
Input: uuid(76eceacb-0462-4d94-ac83-423a6793775e)
Rule: parser:"'uuid' '(' @!')' ')' "
Error: exponent has no digits

Looks like that's getting raised by the Go Scanner, trying to parse a number: https://github.com/golang/go/blob/master/src/go/scanner/scanner.go#L513

Also tried adding the String and Char type to see if it would stop it trying to parse the data as a number, to no avail. Any suggestions other than recreating the whole lexer? Cheers

[alecthomas]

I think you will almost certainly have to roll your own lexer, but it's not too onerous. Here's a simple example from the tests.