Hot off the press! http://wiki.dlang.org/Vision/2016H1 -- Andrei

On 25/01/16 3:37 PM, Andrei Alexandrescu wrote:
> Hot off the press! http://wiki.dlang.org/Vision/2016H1 -- Andrei

There is a couple of things I want on there.

1. scope to be fixed and fully implemented
   (I'll bring some use cases to the table)
2. @assumenogc or something similar.
   That way IAllocator can be @nogc. Which to me is a requirement before it is out of experimental.

Number 1 is the most important for me. Otherwise there will be no review/PR for image library this year.

On 01/24/2016 10:07 PM, Rikki Cattermole wrote:
>
> 1. scope to be fixed and fully implemented
>     (I'll bring some use cases to the table)
> 2. @assumenogc or something similar.
>     That way IAllocator can be @nogc. Which to me is a requirement
> before it is out of experimental.

Both are under the larger headline "Memory Management." -- Andrei

On 25/01/16 4:13 PM, Andrei Alexandrescu wrote:
> On 01/24/2016 10:07 PM, Rikki Cattermole wrote:
>>
>> 1. scope to be fixed and fully implemented
>>     (I'll bring some use cases to the table)
>> 2. @assumenogc or something similar.
>>     That way IAllocator can be @nogc. Which to me is a requirement
>> before it is out of experimental.
>
> Both are under the larger headline "Memory Management." -- Andrei

Okay, I like specifics since then you can tick it off for next round ;)

On Monday, 25 January 2016 at 02:37:40 UTC, Andrei Alexandrescu wrote:
> Hot off the press! http://wiki.dlang.org/Vision/2016H1 -- Andrei

Something went wrong here:

> We fell short of our 2000 pull requests goal in H2 2015. We have had only 1 1378 pull requests.

In addition to the extraneous "1", the link is bad: "No results matched your search."

On Monday, 25 January 2016 at 02:37:40 UTC, Andrei Alexandrescu wrote:
> Hot off the press! http://wiki.dlang.org/Vision/2016H1 -- Andrei

For PRs, I suggest the goal to be number of PRs MERGED instead of created. That may provide the core team a subconsious incentive to look at long pending PRs and hit a good cycle.

For tooling, I suggest a look at GUI/IDEs, now that dlangui/dlangide seems a good candidate(native D, crossplatform). A good official supported GUI library will attract many people.

On Monday, 25 January 2016 at 02:37:40 UTC, Andrei Alexandrescu wrote:
> Hot off the press! http://wiki.dlang.org/Vision/2016H1 -- Andrei

Re: "Safety" and "Library additions", I anticipate that my `checkedint` module (a refinement and extension of Robert Schadek's `SafeInt` work) will be ready to start formal review within the next 1-2 months.

I uploaded it to DUB yesterday: http://code.dlang.org/packages/checkedint

TODO: Mostly documentation (coming this week, probably) and unit tests (it has tests, but they're too slow to stick directly in Phobos), but also miscellaneous small things.

On 25/01/16 4:21 PM, Puming wrote:
> On Monday, 25 January 2016 at 02:37:40 UTC, Andrei Alexandrescu wrote:
>> Hot off the press! http://wiki.dlang.org/Vision/2016H1 -- Andrei
>
> For PRs, I suggest the goal to be number of PRs MERGED instead of
> created. That may provide the core team a subconsious incentive to look
> at long pending PRs and hit a good cycle.
>
> For tooling, I suggest a look at GUI/IDEs, now that dlangui/dlangide
> seems a good candidate(native D, crossplatform). A good official
> supported GUI library will attract many people.

That won't be happening anytime soon.
Until we have image and windowing in Phobos (I'm working on both) there is no way a GUI toolkit is going in. And from what I know there will be a LOT of work to update it.

On Sun, 24 Jan 2016 21:37:40 -0500, Andrei Alexandrescu wrote:

> Hot off the press! http://wiki.dlang.org/Vision/2016H1 -- Andrei

I'm not fond of the militaristic terminology for participants. Novice, adept, master, maybe?

The section on safety is pretty short. I'd like to see in it:
* Guidelines for what should be made @trusted in Phobos (should we trust
Win/Posix API functions? should we only trust wrappers that take D arrays
rather than pointers? can we, for instance, create a @trusted wrapper
around curl?)
* Efficiency / safety tradeoff guidelines (should Phobos provide a
slightly faster implementations of things that aren't @safe? In those
cases, should it provide both @safe and fast alternatives?)
* @safe / @trusted IO as a goal

As is, there are a smattering of things in Phobos that aren't @safe but seem like they could or should be, with no explanation and no safe alternatives. I think almost all IO is @system. This makes it hard and sometimes confusing to try to write @safe code.

On Monday, 25 January 2016 at 03:49:56 UTC, Rikki Cattermole wrote:
> On 25/01/16 4:21 PM, Puming wrote:
>> On Monday, 25 January 2016 at 02:37:40 UTC, Andrei Alexandrescu wrote:
>>> Hot off the press! http://wiki.dlang.org/Vision/2016H1 -- Andrei
>>
>> For PRs, I suggest the goal to be number of PRs MERGED instead of
>> created. That may provide the core team a subconsious incentive to look
>> at long pending PRs and hit a good cycle.
>>
>> For tooling, I suggest a look at GUI/IDEs, now that dlangui/dlangide
>> seems a good candidate(native D, crossplatform). A good official
>> supported GUI library will attract many people.
>
> That won't be happening anytime soon.
> Until we have image and windowing in Phobos (I'm working on both) there is no way a GUI toolkit is going in. And from what I know there will be a LOT of work to update it.

Well I'm not saying that a GUI toolkit should go into Phobos.
I'd rather it stand alone, while taking some official support, say, link in D frontpage(like visualD).