Re: [freesci-develop] Making FreeSCI code C++ friendly

[Jordi Vilalta]

Hi,

2007/3/25, Christoph Reichenbach:
> 1. Even without cfsml it will be big.  For processing, I'd suggest splitting
>    it up (anyone could do that, of course, wouldn't have to be you) to be able
>    to split up the work of merging it in, if we are to accept it.

I'm doing it now.

> 2. I noticed a few places here and there where you might have been able to
>    avoid typecasts altering local variable types.  These are small things,
>    and with sufficiently small patches, they should be easy to revisit.

I'll try to separate the obvious cases where no typecast can be
avoided and the cases where someone can spend more time trying to
clean them.

The attached patch (I hope this time it doesn't get lost) is about the
first obvious case where I think the casts can't be avoided, and it's
where allocation of dynamic memory takes place.

I'll be posting the rest of the patches as soon as I finish them.

> 3. I don't like the renaming of variables such as ``destroy'' and ``class''
>    and ``new'' much.  Now I do realise that those changes are neccessary to
>    convert the core of FreeSCI to be compilable with a present-day C++
>    compiler, but, as I mentioned before, I'm doubtful as to whether that is a
>    helpful change in the first place.

I'll do a separate patch about these.

Regards,

Jordi Vilalta