|
Boost : |
Subject: Re: [boost] [Review] Phoenix review results
From: dan marsden (danmarsden_at_[hidden])
Date: 2008-10-11 04:54:13
Joel de Guzman wrote:
>Eric Niebler wrote:
>>
>> Joel de Guzman wrote:
>>> Now it's time to grab a beer and celebrate :-)
>>
>> And to get (back) to work on Phoenix v3! ;-) Congrats.
>
>Part of Phoenix' success is due to the easy extension mechanism.
>It would be super awesome if you can find some ways to simplify
>V3 extension to something as easy as V2's. I intend to apply
>whatever technique you can come up with to Spirit2. Yeah, I'm
>not happy with its extension mechanism too. Nevertheless, I
>am starting a Spirit 2.1 using Phoenix V3 as the model.
I think the problem with the v2 extension mechanism is that we
exposed how the expression templates were implemented. That left
us with no choice but to make a breaking change in the extension
mechanism between v2 and v3.
IMO ideally any extension mechanism for v3 would not expose directly that
proto is the underlying ET engine. I'm not sure how practical this
will be to do though, as we may need a mass of indirection to avoid
exposing proto directly?
Cheers
Dan
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk