On 1/29/11, Akakima <akakima33@gmail.com> wrote:
>
> "Andrej Mitrovic" <andrej.mitrovich@gmail.com>
>
>>I hope this gets pasted right, otherwise I'll have to open a freakin' blog.
>>
>>:)
>>
>> Up until a few months ago, I had no experience with templates whatsoever.
>>
>
> That's very good stuff! Well written and very helpful.
>
> May i suggest you to put that somewhere in a [ maybe new ] section of digitalmars site reserved to other documentation like that.
>
> As a newbee, i can tell you that this would help.
>
> Thanks
>

Thanks.

Well, I'm not the maintainer of the D website. Maybe I could put it up on wiki4d (http://prowiki.org/wiki4d/wiki.cgi?FrontPage). Someone will have to direct the newbies there.

Save this somewhere or it will be lost here ;)

Done: http://prowiki.org/wiki4d/wiki.cgi?D__Tutorial/D2Templates

I've replaced my obvious English mistakes with slightly less obvious ones. Kinda wish prowiki had a nicer code display though (syntax highlighting would be nice).

Wait, why won't this compile?:

template foo(int value)
{
    int result;
    alias result foo;
}

void main()
{
    writeln( foo!(4) );
}

Oh yeah, it's that damn bug where you can't have multiple declarations, otherwise aliases don't work.

I still don't know if that bug will ever be fixed or if I should add that to the tutorial.

Well I've updated the tutorial anyway. I just left a note and used an templateImpl companion template.

Well now I'm feeling really stupid because apparently this is how it *should* work, at least according to the docs.

I'd love to see 'alias symbol this' work even with multiple declarations. It'd be nicee..

On Saturday 29 January 2011 16:01:06 Andrej Mitrovic wrote:
> Well now I'm feeling really stupid because apparently this is how it *should* work, at least according to the docs.
> 
> I'd love to see 'alias symbol this' work even with multiple declarations. It'd be nicee..

It's how it should work according to TDPL. It'll get fixed It's just a question of when.

- Jonathan M Davis

I would like to thank you for that such a great explanation of the Template-based
programming in that unary example. I think it's great you have uploaded that to
the wiki4d, and it's definite will help a lot of people that come from other
common languages background (like C, C# and Python) that don't support full
templating, to understand that design pattern.
And with that close tracking to the functional.d code, I thing but not sure that
we've found some bugs. I will do some testing later today, and if I will see what
I think to be wrong, I'll update.

and again,
 Thanks

On Sun, Jan 30, 2011 at 12:03, Tom <tom@gmail.com> wrote:
> I would like to thank you for that such a great explanation of the Template-based
> programming in that unary example. I think it's great you have uploaded that to
> the wiki4d, and it's definite will help a lot of people that come from other
> common languages background (like C, C# and Python) that don't support full
> templating, to understand that design pattern.
> And with that close tracking to the functional.d code, I thing but not sure that
> we've found some bugs. I will do some testing later today, and if I will see what
> I think to be wrong, I'll update.

If anyone is interested, I coded a n-args version of unaryFun/binaryFun called naryFun. If you use 'a', 'b', ... as args names, it can automatically determine the templated function arity.

So naryFun!"a + b * c - sin(d-a)" is a 4-args template function.

It's there:

http://svn.dsource.org/projects/dranges/trunk/dranges/docs/functional.html (look for naryFun at the bottom)

code is here: http://dsource.org/projects/dranges/browser/trunk/dranges/functional.d

It could be simpler now: when I did it 18 months ago, CTFE wasn't so powerful. The looping templates can now easily be done with a simple foreach. This has been on my todo list for quite some time...


Philippe