February 22, 2018
On Tuesday, 20 February 2018 at 09:20:15 UTC, aberba wrote:
> I've taken a close look at the Api and its looks good. Its been around for a while. Why is it not yet in phobos?
>
> https://code.dlang.org/packages/std_data_json

A good reason is that a dmd regression broke it:

https://issues.dlang.org/show_bug.cgi?id=18267

I haven't been able to update from dmd 2.077.1 at work because of this.

Atila
February 22, 2018
On Thursday, February 22, 2018 09:35:06 Atila Neves via Digitalmars-d wrote:
> On Tuesday, 20 February 2018 at 09:20:15 UTC, aberba wrote:
> > I've taken a close look at the Api and its looks good. Its been around for a while. Why is it not yet in phobos?
> >
> > https://code.dlang.org/packages/std_data_json
>
> A good reason is that a dmd regression broke it:
>
> https://issues.dlang.org/show_bug.cgi?id=18267
>
> I haven't been able to update from dmd 2.077.1 at work because of this.

Well, while that's definitely a problem, it really has nothing to do with why it's not in Phobos (no attempt to get it into Phobos has occurred since well before 2.077.1 was released), and if it _were_ in Phobos, then the regression would have been caught by the autotester, and the code that caused it would never have been merged - which is one advantage to having something be in Phobos rather than on code.dlang.org.

- Jonathan M Davis

February 22, 2018
On Thursday, 22 February 2018 at 09:45:25 UTC, Jonathan M Davis wrote:
> Well, while that's definitely a problem, it really has nothing to do with why it's not in Phobos (no attempt to get it into Phobos has occurred since well before 2.077.1 was released), and if it _were_ in Phobos, then the regression would have been caught by the autotester, and the code that caused it would never have been merged - which is one advantage to having something be in Phobos rather than on code.dlang.org.
>
> - Jonathan M Davis

I believe it's a common practice in other languages to do something like testing the compiler against top 20 (or more) most popular packages, while I doubt this package would make it to top 20 (do we even have such metrics from dub?), do you know if something like this is being done for D too?
February 22, 2018
On Thursday, February 22, 2018 10:32:34 JN via Digitalmars-d wrote:
> On Thursday, 22 February 2018 at 09:45:25 UTC, Jonathan M Davis
>
> wrote:
> > Well, while that's definitely a problem, it really has nothing to do with why it's not in Phobos (no attempt to get it into Phobos has occurred since well before 2.077.1 was released), and if it _were_ in Phobos, then the regression would have been caught by the autotester, and the code that caused it would never have been merged - which is one advantage to having something be in Phobos rather than on code.dlang.org.
> >
> > - Jonathan M Davis
>
> I believe it's a common practice in other languages to do something like testing the compiler against top 20 (or more) most popular packages, while I doubt this package would make it to top 20 (do we even have such metrics from dub?), do you know if something like this is being done for D too?

There is some testing that is done against a set of projects on code.dlang.org, though I get the impression that it's not very reliable (it frequently seems to fail due to reasons unrelated to the PRs being tested), and for better or worse, it doesn't have to pass for something to be merged. So, even if std_data_json is on the list of projects being tested, I'm not sure that it would have helped much. I don't know if std_data_json is on the list though.

CI stuff like that has been improving, but there's still work to be done on that front.

- Jonathan M Davis

February 22, 2018
On 22/02/2018 11:32 PM, JN wrote:
> On Thursday, 22 February 2018 at 09:45:25 UTC, Jonathan M Davis wrote:
>> Well, while that's definitely a problem, it really has nothing to do with why it's not in Phobos (no attempt to get it into Phobos has occurred since well before 2.077.1 was released), and if it _were_ in Phobos, then the regression would have been caught by the autotester, and the code that caused it would never have been merged - which is one advantage to having something be in Phobos rather than on code.dlang.org.
>>
>> - Jonathan M Davis
> 
> I believe it's a common practice in other languages to do something like testing the compiler against top 20 (or more) most popular packages, while I doubt this package would make it to top 20 (do we even have such metrics from dub?), do you know if something like this is being done for D too?

Yes http://code.dlang.org/api/packages/dump

	"stats": {
		"updatedAt": "2018-02-22T10:21:07.711Z",
		"downloads": {
			"total": 226,
			"monthly": 0,
			"weekly": 0,
			"daily": 0
		},
		"repo": {
			"stars": 7,
			"watchers": 2,
			"forks": 6,
			"issues": 1
		},
		"rating": 1.62909460067749
	},
February 22, 2018
On Thursday, 22 February 2018 at 10:40:11 UTC, Jonathan M Davis wrote:
> There is some testing that is done against a set of projects on code.dlang.org, though I get the impression that it's not very reliable (it frequently seems to fail due to reasons unrelated to the PRs being tested), and for better or worse, it doesn't have to pass for something to be merged.

It has to pass since a few weeks.
February 22, 2018
On Thursday, 22 February 2018 at 10:32:34 UTC, JN wrote:
> I believe it's a common practice in other languages to do something like testing the compiler against top 20 (or more) most popular packages, while I doubt this package would make it to top 20 (do we even have such metrics from dub?), do you know if something like this is being done for D too?

We do even better - an example from a recent PR:

https://ci.dlang.io/blue/organizations/jenkins/dlang-org%2Fphobos/detail/PR-6207/1

It's ~40 projects.

(though it's really a PITA to test all these projects because (1) Jenkins is a horrible piece of software and (2) some of these projects have randomness (and thus random failures) in their testsuites.)
February 22, 2018
On Thursday, 22 February 2018 at 07:56:12 UTC, Jonathan M Davis wrote:
> On Thursday, February 22, 2018 07:24:30 aberba via Digitalmars-d wrote:
>> [...]
>
> I did use Java in previous jobs, but I don't curently use it. In recent years, I've done far more with C++ and D, and my current job mostly involves D. And I would almost certainly turn down a job that involved much in the way of Java unless I were desperate. It's just too primitive to be a pleasant language to program in.
>
> [...]

Nice.
1 2
Next ›   Last »