Re: [boost] Re: review request: container_traits

From: Pavol Droba (droba_at_[hidden])
Date: 2003-10-04 17:22:43

Hi,

I understand your reasons now, although it is only partialy due to this 
long reasoning. 
Separated traits are most probably a best presentation layer for the
container_traits problem.

There was a great effort put into in the current implementation to
make it work on defficient compilers. Workarounds are easier to
implement using monolitic approach. This is probably also major reason
why I was supporting it. 

I would suggest to define the new interface, and implement it on
top of the current one for now. 

Regards,

Pavol

On Sat, Oct 04, 2003 at 03:47:03PM -0400, David Abrahams wrote:

[snip]
 
> Sorry, I find that argument unconvincing.  You can easily factor out
> common parts into separate metafunctions.  I'm going to stop arguing
> about this now; if I haven't convinced _you_ by now, I doubt anything
> will.  Let me just close by saying that "decoupled is better than
> coupled" is a well established principle, and its sad that we (me
> included) seem to have to keep relearning that lesson by making the
> same mistakes.  I'm just thankful Fredrik spotted this instance of the
> mistake, and I hope you and Thorsten will respond accordingly.