On Monday, 16 September 2013 at 15:52:26 UTC, Bruno Medeiros
wrote:
> On 13/09/2013 08:46, eles wrote:
>> On Saturday, 7 September 2013 at 19:05:03 UTC, Walter Bright wrote:
> It's not clear to me what any of these measures would help with.

Glad that you answered.

It will help with:

1) people already contributing to the D chains generally use
github, so they are familiar with the workflow, the interface,
the code-review etc. This will increase the probability to
contribute with a PR, even if for small glitches (or
documentation) in the beginning (BTW me, for one, I find the
interface on code.google to be awful at best, when compared to
github).

2) people that are in the D community know about Eclipse/DDT.
However, those passing by and wondering about what the D language
means, *and* are Eclipse users, they tend to not really know
about it. Having it officially endorsed, visible on the download
page and so on, will help them in using it (and, with a bit of
chance, contributing to it).

As a sidenote: we use a lot Eclipse/CDT when working in C/C++,
but with an external Makefile. You have best of two worlds: an
IDE, and an autotools/make integration. Let's not even speak that
the Makefile itself is just a shell around scons/SConstruct.

On 17/09/2013 15:48, eles wrote:
> On Monday, 16 September 2013 at 15:52:26 UTC, Bruno Medeiros
> wrote:
>> On 13/09/2013 08:46, eles wrote:
>>> On Saturday, 7 September 2013 at 19:05:03 UTC, Walter Bright wrote:
>> It's not clear to me what any of these measures would help with.
>
> Glad that you answered.
>
> It will help with:
>
> 1) people already contributing to the D chains generally use
> github, so they are familiar with the workflow, the interface,
> the code-review etc. This will increase the probability to
> contribute with a PR, even if for small glitches (or
> documentation) in the beginning (BTW me, for one, I find the
> interface on code.google to be awful at best, when compared to
> github).
>

This is more of an issue of project hosting than whether it's officially endorsed/listed or not. I could move the hosting of DDT from Google Code to Github and potentially reap some of those benefits, regardless of DLang endorsement or not. (In hindsight I do agree it might have been better to have it hosted on Github. Google Code seems to have stagnated a bit while Github is getting more popular and getting better - but things were different when I switched away from DSource.org)


-- 
Bruno Medeiros - Software Engineer

On Tuesday, 17 September 2013 at 15:34:15 UTC, Bruno Medeiros
wrote:
> On 17/09/2013 15:48, eles wrote:
>> On Monday, 16 September 2013 at 15:52:26 UTC, Bruno Medeiros
>> wrote:
>>> On 13/09/2013 08:46, eles wrote:
>>>> On Saturday, 7 September 2013 at 19:05:03 UTC, Walter Bright
> (In hindsight I do agree it might have been better to have it hosted on Github.

I encourage you to do that, in the beginning. After the VisualD
move will mature a bit and, hopefully, DDT will integrate with
the debugger, it will become official.

Speaking about that, did you have the time to have a look at
Descent's debugging module? Recently, Iain really improved gdc's
gdb debugging support, and supporting at least gdb (it is usable
on Windows, too) would be a break-through.

On 18/09/2013 07:58, eles wrote:
> Speaking about that, did you have the time to have a look at
> Descent's debugging module? Recently, Iain really improved gdc's
> gdb debugging support, and supporting at least gdb (it is usable
> on Windows, too) would be a break-through.

Not yet.

-- 
Bruno Medeiros - Software Engineer

On Wednesday, 18 September 2013 at 13:21:45 UTC, Bruno Medeiros
wrote:
> On 18/09/2013 07:58, eles wrote:
> Not yet.

How could I help?

On Sun, 15 Sep 2013 18:56:32 +0200
"deadalnix" <deadalnix@gmail.com> wrote:

> On Sunday, 15 September 2013 at 16:41:43 UTC, Paulo Pinto wrote:
> > I would say the selling points would mainly be:
> >
> > - system level programming features outside what unsafe blocks
> > allow for in C#
> > - meta-programming abilities
> > - more friendly ways to manage memory manually
> >
> 
> - slices !
> 
> > All capabilities related to low level coding.
> >

Yea, pretty much those things, although metaprogramming is much more than being a "low level tool", it's very much about productivity.

I might also add "no need for mono on non-MS platforms", just because
my understanding is that mono's not as mature/complete and doesn't get
as much attention or end-user installations as .NET. But I could
be wrong though: I've haven't actually used mono except once in a failed
attempt to run a specific VB.NET-based app on Linux a couple years ago.

On Mon, 16 Sep 2013 15:22:29 +0100
Bruno Medeiros <brunodomedeiros+dng@gmail.com> wrote:
> 
> But Eclipse bashing by people who use say, VisualStudio, that I don't understand. Last time I tried both toolchains, VS seemed as heavy and "bloated" as Eclipse (CDT) was. Yet CDT seemed quite ahead in terms of features, especially semantics-wise (open definition, code complete, etc.). Admittedly this was 3-4 years ago, and I only toyed lightly with C/C++ code, I didn't do any serious development. But I doubt the situation changed such that VS got much better than CDT, if anything, the opposite is more likely.
> 

Unless things have changed since I last tried, Eclipse is very Java-centric (or JVM-centric) at a fairly fundamental level. Support for other languages (from what I've tried) tends to get crammed into a Java-oriented model, which can make things confusing and leave a bad impression.

On 18/09/2013 20:53, Nick Sabalausky wrote:
> Unless things have changed since I last tried, Eclipse is very
> Java-centric (or JVM-centric) at a fairly fundamental level. Support
> for other languages (from what I've tried) tends to get crammed into a
> Java-oriented model, which can make things confusing and leave a bad
> impression.

That's true, there is a lot of Java/JDT likeness in other Eclipse IDEs. I don't see why that would be a problem though, unless there is really very little customization in the IDE, and things are too much Java-centric. But some Java/JDT concepts apply well to other IDEs/languages too, I believe. For example the concept of projects with source folders, etc., applies well to D.

Do you remember any specific example you thought confusion or bad? And which language/IDE was it applying to?

-- 
Bruno Medeiros - Software Engineer

On Tuesday, 17 September 2013 at 13:03:46 UTC, Manu wrote:
> On 17 September 2013 21:48, Bruno Medeiros <brunodomedeiros+dng@gmail.com>wrote:
>
>> On 17/09/2013 12:37, Bruno Medeiros wrote:
>>
>>> On 16/09/2013 22:39, Brad Roberts wrote:
>>>
>>>> On 9/16/13 8:52 AM, Bruno Medeiros wrote:
>>>>
>>>>> Correct me if I'm wrong, but I think Manu's point with regards with
>>>>> IDE "official endorsement" was
>>>>> more to try to have the D language organization devs (Walter, Andrei,
>>>>> etc.) *use* VisualD or another
>>>>> IDE and understand the issues around it (especially with regards to
>>>>> compiler/debugger integration).
>>>>>
>>>>
>>>> If that's the definition of official endorsement, then sorry, not likely
>>>> to ever happen.  Demanding that the core devs develop with specific
>>>> tools is ridiculous in concept.  Would you switch because someone told
>>>> you to?  Me either.  I've been using vi(m) for about 20 years now.  My
>>>> fingers know what to do without conscious control.. I don't have the
>>>> free time nor the desire to retrain myself like that.
>>>>
>>>>
>>> I agree, I don't think it's a realistic to expect that. I was just
>>> pointing out Manu's idea, not agreeing with it.
>>>
>>>
>> Clarification: I think it's unrealistic to expected the core devs to use
>> the IDE for all of their D development, yes. But it would be good to have
>> them *try* it, to see how it works, to understand how others users would
>> develop in D, what quality issues there could be with it, etc. In this
>> regard I agree with Manu's comments.
>
>
> I'll happily wear that my assertion was heavy handed, mostly due to
> long-term frustration, and to some extent, this is just the way I talk
> (which never comes across in text to people who don't know me).
> Regardless of how I phrased it however, I'm encouraged to see the message
> was generally well received and actions have been taken. I'm keen to see
> if/how it affects the ecosystem in the future.
> I hope it does increase the overall attention/awareness.

It's certains if the D language organization devs use "official" IDEs the quality will be improve greatly.

Good reports can only come from real usages, IDEs are critical for productivity the GUI polish is something important but hard to achieved.

On my project DQuick, VisualD just can't find members of classes aren't directly defined in the current module. I also try the beta of next version for code coverage with unittest without any success.

On 18/09/2013 14:41, eles wrote:
> On Wednesday, 18 September 2013 at 13:21:45 UTC, Bruno Medeiros
> wrote:
>> On 18/09/2013 07:58, eles wrote:
>> Not yet.
>
> How could I help?

If you're willing to spend enough time, you can try implementing debugger support yourself:
http://www.eclipse.org/articles/Article-Debugger/how-to.html

But otherwise just doing something half-way is not gonna help much (if that means I have to spend a lot of time reviewing and modyfing the contribution, close to as much as if I were implementing it myself).

-- 
Bruno Medeiros - Software Engineer