Thread overview
[D-runtime] D runtime initialization from C fails on OSX
May 23, 2012
Jacob Carlborg
May 23, 2012
Walter Bright
May 24, 2012
Jacob Carlborg
May 25, 2012
Jacob Carlborg
May 23, 2012
I'm trying to fix this issue:

http://d.puremagic.com/issues/show_bug.cgi?id=7995

I created a pull request:

https://github.com/D-Programming-Language/druntime/pull/205

But that apparently broke some test, although everything in the auto tester passed as far as I could see.

I don't understand the test case that Walter gives in the comments:

https://github.com/D-Programming-Language/druntime/pull/205#issuecomment-5675090

The test case behaves the same for me with the 2.059 release as with my pull request applied. I'm writing here now because I was hoping that Walter or someone else could give some more information. As usual I'm basically only getting a "Fails" if the pull request breaks something. No information about what the behavior is and what the expected behavior should be.

-- 
/Jacob Carlborg

_______________________________________________
D-runtime mailing list
D-runtime@puremagic.com
http://lists.puremagic.com/mailman/listinfo/d-runtime

May 23, 2012
I replied on the github page too:

The exit status of the program must be non-zero. It was zero. Sorry for not making that clear earlier. I tested it on OSX.

On 5/23/2012 1:14 PM, Jacob Carlborg wrote:
> I'm trying to fix this issue:
>
> http://d.puremagic.com/issues/show_bug.cgi?id=7995
>
> I created a pull request:
>
> https://github.com/D-Programming-Language/druntime/pull/205
>
> But that apparently broke some test, although everything in the auto tester passed as far as I could see.
>
> I don't understand the test case that Walter gives in the comments:
>
> https://github.com/D-Programming-Language/druntime/pull/205#issuecomment-5675090
>
> The test case behaves the same for me with the 2.059 release as with my pull request applied. I'm writing here now because I was hoping that Walter or someone else could give some more information. As usual I'm basically only getting a "Fails" if the pull request breaks something. No information about what the behavior is and what the expected behavior should be.
>
_______________________________________________
D-runtime mailing list
D-runtime@puremagic.com
http://lists.puremagic.com/mailman/listinfo/d-runtime

May 24, 2012

On May 24, 2012, at 12:20 AM, Walter Bright <walter@digitalmars.com> wrote:

> I replied on the github page too:
>
> The exit status of the program must be non-zero. It was zero. Sorry for not making that clear earlier. I tested it on OSX.

Thanks, that was just what I needed :) . I would be great if you could give this information at the same time you revert a pull request. It makes life easier both for the the one who did the pull request and for you.

May 25, 2012

On May 24, 2012, at 12:20 AM, Walter Bright <walter@digitalmars.com> wrote:

> I replied on the github page too:
>
> The exit status of the program must be non-zero. It was zero. Sorry for not making that clear earlier. I tested it on OSX.

I've made some modifications to the previous code and your test behaves as expected now. All unit tests in Phobos pass. Although I want to double check, is it ok that the unit tests are not run when just calling "rt_init", i.e. starting the runtime from a C application? The unit tests are still run as normal when starting from a D application.

BTW is this test case part of some test suite that is publicly available? If not, I think it should be added.

https://github.com/jacob-carlborg/druntime/commit/ad5fca3000c6e25e9c7f35fddf93776322462d66

--
/Jacob Carlborg