Re: [cpp-netlib-devel] Moving the Parser Implementation to a .cpp

[Dean M. B. <mik...@gm...>]

On Fri, Dec 3, 2010 at 11:17 PM, Nelson, Erik - 2
<eri...@ba...> wrote:
>> Dean Michael Berris wrote on Thursday, December 02, 2010 11:35 PM
>> I think the easiest way to minimize the compile time cost would be to
>> move the Spirit.Qi based parser out to a .cpp and just build and
>> external lib. I'll try to make that happen locally and see how far
>> that goes, and I'll update everyone on what that's going to look like
>> for 0.9 soon as I'm done.
>>
>> Comments, feedback, violent reactions, etc. would be most appreciated.
>>
>
>
> Sounds like a good start.
>

Thanks Erik. :)

> I'm not sure if you've been following or went to any of the MSM stuff
> this year at BoostCon, but that's a pretty sweet implementation if you
> need a state machine in the course of making cpp-netlib.
>

Yes, I did -- and I was there! :D

Unfortunately that's yet another hammer I need to learn to use. The
implementation right now works with a hand-rolled incremental parser
-- it's going to be interesting to see how I (or someone else) can
port that to use Boost.MSM.

The EUML stuff is cool and I'd really like to give that a spin. But, I
don't have the time to much with that part of the implementation yet.
Maybe later on I'll get more familiar with it and start moving the
implementations to use Boost.MSM later on.

-- 
Dean Michael Berris
deanberris.com