Porting to Spirit2

[Braden McDaniel]

Okay, so it's really more of a rewrite than a port... but I've finally
started recasting uri_grammar in terms of Spirit2. More than that, I'm
following the modern URI grammar in RFC 3986.

I expect to have something worth committing in the next few days.

--
Braden McDaniel <bra...@endoframe.com>

--
You received this message because you are subscribed to the Google Groups "uri-grammar" group.
To post to this group, send email to uri-g...@googlegroups.com.
To unsubscribe from this group, send email to uri-grammar...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/uri-grammar?hl=en.

[Braden McDaniel]

On Thu, 2010-04-22 at 17:42 -0400, Braden McDaniel wrote:
> Okay, so it's really more of a rewrite than a port... but I've finally
> started recasting uri_grammar in terms of Spirit2. More than that, I'm
> following the modern URI grammar in RFC 3986.
>
> I expect to have something worth committing in the next few days.

This is taking a bit longer than I'd hoped. After encountering some
Spirit bugs that are supposedly fixed in Boost 1.43, I've decided to
postpone further work here until that release is available. That
shouldn't be too long, as I understand it.

[Braden McDaniel]

On Thu, 2010-04-22 at 17:42 -0400, Braden McDaniel wrote:

> Okay, so it's really more of a rewrite than a port... but I've finally
> started recasting uri_grammar in terms of Spirit2. More than that, I'm
> following the modern URI grammar in RFC 3986.
>
> I expect to have something worth committing in the next few days.

I have committed this, along with a lot more tests (especially for the
IPv6 stuff).

* I have made use of qi::raw to get iterator pairs to the various
URI components. This pretty much eliminated the need for
user-supplied semantic actions.
* There may be room for further componentization into
qi::grammars, particularly where IPv6 address parsing is
concerned.
* I have changed the parse-uri test program to use
Boost.Program_options rather than GNU argp. (The reason it
didn't already is that when I first wrote this years ago,
Boost.Program_options was too broken to use.)