On 4/13/2014 10:38 AM, Dicebot wrote:
> I think most important thing about built-in unittests is how this feature
> encourages to add tests when contributing to projects that don't currently have
> any real coverage. It becomes so easy that there is not reason to not do it and
> this help open-source a lot.

Exactly. You can do unittests with other languages, too, but having it be so convenient, and no arguments about which unittest tool, is a game changer. The built-in -cov is also a game changer.

Am 13.04.2014 20:40, schrieb Walter Bright:
> On 4/13/2014 10:38 AM, Dicebot wrote:
>> I think most important thing about built-in unittests is how this feature
>> encourages to add tests when contributing to projects that don't
>> currently have
>> any real coverage. It becomes so easy that there is not reason to not
>> do it and
>> this help open-source a lot.
>
> Exactly. You can do unittests with other languages, too, but having it
> be so convenient, and no arguments about which unittest tool, is a game
> changer. The built-in -cov is also a game changer.

Until one gets lousy management in the picture. :(

Sorry, just got a lousy week with developers deleting unit tests to "improve" their builds.

--
Paulo

On 13/04/14 20:40, Walter Bright wrote:

> Exactly. You can do unittests with other languages, too, but having it
> be so convenient, and no arguments about which unittest tool, is a game
> changer. The built-in -cov is also a game changer.

The built-in unittest functionality is too limited. One needs some framework built on top of it. Then you get back to the arguments about which framework to use.

-- 
/Jacob Carlborg

Am Sun, 13 Apr 2014 09:27:17 -0700
schrieb Walter Bright <newshound2@digitalmars.com>:

> On 4/13/2014 4:04 AM, Michel Fortin wrote:
> > Interesting. As far as I know, the D GC is also a wrapper around malloc, and it will not return memory using free when an object is deallocated. That rant could also apply to D.
> 
> You're right in that a malloc debugger would be ineffective when the D GC is used. But the D GC offers its own memory debugging features. They can be turned on with version identifiers in the GC source.

...and disabled by default. The point is, programmers who trade security for performance will now and then lose that gamble. They WILL circumvent whatever the language offers if they can gain speed from it. In D there is minimallyInitializedArray() for them and void-initializers or pointer arithmetic.

-- 
Marco

On Sunday, 13 April 2014 at 18:45:21 UTC, Paulo Pinto wrote:
> Am 13.04.2014 20:40, schrieb Walter Bright:
>> On 4/13/2014 10:38 AM, Dicebot wrote:
>>> I think most important thing about built-in unittests is how this feature
>>> encourages to add tests when contributing to projects that don't
>>> currently have
>>> any real coverage. It becomes so easy that there is not reason to not
>>> do it and
>>> this help open-source a lot.
>>
>> Exactly. You can do unittests with other languages, too, but having it
>> be so convenient, and no arguments about which unittest tool, is a game
>> changer. The built-in -cov is also a game changer.
>
> Until one gets lousy management in the picture. :(
>
> Sorry, just got a lousy week with developers deleting unit tests to "improve" their builds.
>
> --
> Paulo

Wow, that would almost be a sacking offence where I work.