Angus Leeming <[EMAIL PROTECTED]> writes:

| Lars Gullik Bjønnes wrote:
>
>> Now I did it... and now I'll store this in some dark place until 1.5
>> started...
>> 
>> So just for your info.
>> 
>> The nice thing about this is that it prepares for further clean up and
>> the removal of the insetmailers. (this patch would really like
>> exceptions to be enabled as well...)
>> 
>> Have a look if you want to, else just ignore.
>
| Looks fine, although I note that it doesn't actually *do* anything yet. The
| change so far is to store 'argument' as a boost::any in FuncRequest,
| rather than as a string. However, it appears that 'argument' is still
| always encoded as a string so there's no gain at all yet.
>
| Have I got this right?

Yes, that was the idea.

| What will be interesting will be to see if you can actually start to encode
| things other than strings and use them in the core.

The plan is also to encode the requirements of the different lfuns
(number of parameters, type of parameters etc.) and actually check
those in the dispatch.

| Nonetheless, it's an interesting project.

On the back-burner, I might create a branch for it.

-- 
        Lgb

Reply via email to