Re: IPv6 address parsing

On Tue, Dec 19, 2017 at 4:25 AM, Seth <[email protected]> wrote:
> On 17-12-17 16:42, Michael Powell wrote:
>
> i.e. repeat(1, 4)[hex] SHOULD reject "abcd1234" but it is not.
>
> Of course it SHOULD not. If you want that, program it. We've been over this
> many times before:

What are you talking about, "it SHOULD not"? I *DID* "program" it; I
want from 1 to 4 hexadecimal characters. *THAT'S IT*! No fewer, no
more. That's the whole point of repeat(1, 4)[...] isn't it?

It's possible I am misunderstanding the nature of HEX, per se, which,
after consulting again with the docs, probably I am.

http://www.boost.org/doc/libs/1_66_0/libs/spirit/doc/html/spirit/qi/reference/numeric/uint.html

That is to say, if *HEX* is excessively greedy, that's another matter,
entirely. Perhaps then the PEG hex that is provided is ill suited to
the task, and rather the uint_parser in the appropriate base-16 Radix
is the better solution to the problem at hand.

That's what I'm getting at here, the dynamics of repeating a specific
number of hex characters.

Same goes for bin and oct; I'm not sure how useful matching a general
form value of that pattern, or a specific xyz(num) is, when what I
want, at least in this instance, is a single "digit". But this does
not appear to be the nature of these particular parsers.

Sorry if that seems confusing; much of which seems to be mine. I'm
just trying to sort out the nuts and bolts of Spirit/Qi in order to
apply the best solution.

> On 08-12-17 16:53, Seth wrote:
>
> On 04-12-17 23:09, Michael Powell wrote:
>
> I should think that if there were any remaining input that would be a
> failure to match?
>
> You shouldn't. PEG grammars are left-to-right greedy. This is documented.
>
> I've said this before:
>
> On 30-11-17 00:40, Seth wrote:
>
> Bonus tip: consider adding `>> qi::eoi` at the end of the parse
> expression so you don't need to check whether all of the input has been
> consumed (it must).
>
> And also this warning about partially matching:
>
> On 27-11-17 13:51, Seth wrote:
>
> A more complete example:
>
> https://stackoverflow.com/questions/47243884/boost-spirit-qi-validating-input-parser/47244585#47244585
>
> Which makes the point that you might want to `lexeme [
> uint_parser<uint16_t, 10, 2, 5>() >> !(digit|'.') ]` to avoid
> partially matching longer numbers.
>
>

------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot