Hi all,


I'm back from a few long trips during which I got behind reading this group. I decided to mark everything as read and restart anew. If there are any topics that you believe have "extinguished" but need my attention, I'd be indebted if you mentioned them to me via private email or by replying to this.


Thanks,

Andrei

On 12-11-2012 20:42, Andrei Alexandrescu wrote:
> Hi all,
>
>
> I'm back from a few long trips during which I got behind reading this
> group. I decided to mark everything as read and restart anew. If there
> are any topics that you believe have "extinguished" but need my
> attention, I'd be indebted if you mentioned them to me via private email
> or by replying to this.
>
>
> Thanks,
>
> Andrei

The "Something needs to happen with shared, and soon." thread may be of interest to you, as well as any parts of "deprecate deprecated?" you may have missed.

Welcome back!

-- 
Alex Rønne Petersen
alex@lycus.org
http://lycus.org

Le 12/11/2012 20:42, Andrei Alexandrescu a écrit :
> Hi all,
>
>
> I'm back from a few long trips during which I got behind reading this
> group. I decided to mark everything as read and restart anew. If there
> are any topics that you believe have "extinguished" but need my
> attention, I'd be indebted if you mentioned them to me via private email
> or by replying to this.
>
>
> Thanks,
>
> Andrei

Topic on UDA obviously. Topic on range transience probably, as it is almost concluded.

On Monday, 12 November 2012 at 19:42:14 UTC, Andrei Alexandrescu wrote:
> Hi all,
>
>
> I'm back from a few long trips during which I got behind reading this group. I decided to mark everything as read and restart anew. If there are any topics that you believe have "extinguished" but need my attention, I'd be indebted if you mentioned them to me via private email or by replying to this.
>
>
> Thanks,
>
> Andrei

Welcome back!

Long discussion on const ref,
http://forum.dlang.org/thread/yhnbcocwxnbutylfeoxi@forum.dlang.org

here's the summary thread
http://forum.dlang.org/thread/zteryxwxyngvyqvukqkm@forum.dlang.org

--rt

On 11/12/12 11:45 AM, deadalnix wrote:
> Le 12/11/2012 20:42, Andrei Alexandrescu a écrit :
>> Hi all,
>>
>>
>> I'm back from a few long trips during which I got behind reading this
>> group. I decided to mark everything as read and restart anew. If there
>> are any topics that you believe have "extinguished" but need my
>> attention, I'd be indebted if you mentioned them to me via private email
>> or by replying to this.
>>
>>
>> Thanks,
>>
>> Andrei
>
> Topic on UDA obviously.

Thanks! I did follow that and in brief my opinion is:

* The feature looks good

* The choice of syntax is mostly motivated by the question: do we want to integrate the existing "@"-attributes with the user-defined ones? If so, the choice of the @ syntax is a consequence.

* I saw very little focus for attribute testing, i.e. "does symbol xyz have attribute abc"? I don't think we should relegate that to a __traits.

* The execution of that feature (merging in the trunk along with a bunch of unrelated work) was extremely poor. We need to acquire a sense of urgency about fixing our process, lest this mom-and-pop-shop style of getting work done will end up alienating our user base.

> Topic on range transience probably, as it is
> almost concluded.

I'm leaning toward doing nothing about this.


Andrei

On 2012-11-12 21:28, Andrei Alexandrescu wrote:

> * I saw very little focus for attribute testing, i.e. "does symbol xyz
> have attribute abc"? I don't think we should relegate that to a __traits.

The current implementation only has __traits(getAttributes), which will basically return a tuple of all the values attached to a symbol. On top of that we can build library functions for checking if an attribute exists for a symbol.

> * The execution of that feature (merging in the trunk along with a bunch
> of unrelated work) was extremely poor. We need to acquire a sense of
> urgency about fixing our process, lest this mom-and-pop-shop style of
> getting work done will end up alienating our user base.

Thank you for acknowledging this.

What's also being discussed in addition to your points are:

* Should built in types be allowed as attributes
* Should only types explicitly marked as an attribute (somehow) be allowed as attributes

-- 
/Jacob Carlborg

On Monday, November 12, 2012 21:24:06 Rob T wrote:
> Long discussion on const ref, http://forum.dlang.org/thread/yhnbcocwxnbutylfeoxi@forum.dlang.org
> 
> here's the summary thread http://forum.dlang.org/thread/zteryxwxyngvyqvukqkm@forum.dlang.org

Yeah. You have some very strong views and that issue which almost no one else actually understands, so a lot of people keep pushing for D's const ref to act like C++'s const&.

- Jonathan M Davis

Andrei Alexandrescu:

Welcome back.

> If there are any topics that you believe have "extinguished" but need my attention, I'd be indebted if you mentioned them to me via private email or by replying to this.

There are some threads I have opened, or where I have written some posts in:

In this thread
http://forum.dlang.org/thread/50A0EEA4.7010706@webdrake.net

Joseph Rushton Wakeling suggests code like this to compile:

struct Foo(_T) {
    alias _T T;
}
void bar(FooT)(FooT foo, FooT.T x) {
}
void main() {
    Foo!int foo;
    bar(foo, 1); // line 8
}



Here I have suggested to add a lenses-like function to Phobos:
http://forum.dlang.org/thread/dqvaohhgthbsygoctwmh@forum.dlang.org


Here I have suggested a little refinement for D purity:
http://forum.dlang.org/thread/hqtgzpnfbpbvdiriogin@forum.dlang.org


Here I have underlined a little design problem with opCast(bool):
http://forum.dlang.org/thread/eicpsjarvxvahknluqwu@forum.dlang.org


In the middle of this long thread I have exchanged several messages with Walter:
http://forum.dlang.org/thread/k7afq6$2832$1@digitalmars.com?page=10

where I have suggested to disallow:

void main() {
    auto data = [0, 1, 2, 3];
    foreach (x; data)
        x++;
}

And replace it with this:

void main() {
    auto data = [0, 1, 2, 3];
    foreach (@copy x; data)
        writeln(++x);
}

You will find many more details in the thread, especially in this post of mine where I have tried to explain the situation as much clearly as possible:
http://forum.dlang.org/thread/k7afq6$2832$1@digitalmars.com?page=12#post-hqdihcnzzgnlfvznhqbc:40forum.dlang.org

Bye,
bearophile

On Mon, Nov 12, 2012 at 2:28 PM, Andrei Alexandrescu <SeeWebsiteForEmail@erdani.org> wrote:
>
> * The execution of that feature (merging in the trunk along with a bunch of unrelated work) was extremely poor. We need to acquire a sense of urgency about fixing our process, lest this mom-and-pop-shop style of getting work done will end up alienating our user base.
>
> Andrei

+1

On 11/12/12 12:50 PM, Jacob Carlborg wrote:
> What's also being discussed in addition to your points are:
>
> * Should built in types be allowed as attributes
> * Should only types explicitly marked as an attribute (somehow) be
> allowed as attributes

Saw that too. I don't have an opinion about such, though the comparison with exception types is rather interesting.

Andrei