On Tuesday, 27 August 2013 at 13:35:36 UTC, Dmitry Olshansky wrote:
> 27-Aug-2013 17:33, Dmitry Olshansky пишет:
>> 27-Aug-2013 11:45, monarch_dodra пишет:
>>> Koenig lookup in C++ ( also known as
>>> http://en.wikipedia.org/wiki/Argument-dependent_name_lookup ), is
>>> a mechanism that alows functions (in particular templates) to
>>> call a function that operates on an object of a specific
>>> namespace, without having to qualify the namespace  ...
>>
>> ADL has a "bug of surprise" written all over it.
>
> *Bag.. though hardly a typo )
>
>> It doesn't work all
>> that well in C++ to being with. Even after all these years:
>> http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2006/n2103.pdf

Well, I don't expect we'd actually change the way things are done *now*, and for all the "bugs" ADL had, it also allowed for some very powerful and *extendable* generic programming.

D has a way of doing things that C++ can, but usually, with a twist that makes it safer and/or more powerful, or with some very well thought mechanics.

In this case, it really feels as if the whole a host of functionality is just both supported, *and* with no workarounds :/

If we don't have ADL I kind of like the notion of "forced import infection" though...

On Tuesday, 27 August 2013 at 13:51:51 UTC, monarch_dodra wrote:
> D has a way of doing things that C++ can, but usually, with a twist that makes it safer and/or more powerful, or with some very well thought mechanics.
>
> In this case, it really feels as if the whole a host of functionality is just both supported, *and* with no workarounds :/
>
> If we don't have ADL I kind of like the notion of "forced import infection" though...

What do you mean by that ? Being able to force feed an import in a template instantiation ? That'd be great, because of UFCS.

On Tuesday, 27 August 2013 at 15:10:45 UTC, deadalnix wrote:
> On Tuesday, 27 August 2013 at 13:51:51 UTC, monarch_dodra wrote:
>> D has a way of doing things that C++ can, but usually, with a twist that makes it safer and/or more powerful, or with some very well thought mechanics.
>>
>> In this case, it really feels as if the whole a host of functionality is just both supported, *and* with no workarounds :/
>>
>> If we don't have ADL I kind of like the notion of "forced import infection" though...
>
> What do you mean by that ? Being able to force feed an import in a template instantiation ? That'd be great, because of UFCS.

I'm unsure ^^

you brought it up: "We should be able to inject import into templates."

Did you have anything specific in mind?

On Tuesday, 27 August 2013 at 16:59:14 UTC, monarch_dodra wrote:
> On Tuesday, 27 August 2013 at 15:10:45 UTC, deadalnix wrote:
>> On Tuesday, 27 August 2013 at 13:51:51 UTC, monarch_dodra wrote:
>>> D has a way of doing things that C++ can, but usually, with a twist that makes it safer and/or more powerful, or with some very well thought mechanics.
>>>
>>> In this case, it really feels as if the whole a host of functionality is just both supported, *and* with no workarounds :/
>>>
>>> If we don't have ADL I kind of like the notion of "forced import infection" though...
>>
>> What do you mean by that ? Being able to force feed an import in a template instantiation ? That'd be great, because of UFCS.
>
> I'm unsure ^^
>
> you brought it up: "We should be able to inject import into templates."
>
> Did you have anything specific in mind?

We can already do this (sortof) by passing some imports in a string/mixin template to the template. Obviously you have to write the template with this in mind though, which isn't ideal. Also you'd have to check for duplicate imports.

On 8/27/2013 6:35 AM, Dmitry Olshansky wrote:
> 27-Aug-2013 17:33, Dmitry Olshansky пишет:
>> ADL has a "bug of surprise" written all over it.
>
> *Bag.. though hardly a typo )

I think you've coined a new term!