For some reason, the unit tests on Mac OS X are failing for std.file.symlink. They pass just fine on Linux and FreeBSD, but the the Mac isn't so happy with them.

http://d.puremagic.com/test-results/

Does anyone with a Mac have time to look at it? It tries to create a symlink to /usr/include and verify that it worked, and it claims that the file exists after calling symlink, but it also claims that it's not a symlink (hence the failure). I have no access to a Mac, and so I can't do anything to fix it. We can disable to the failing unit tests if we have to, but we should really fix the issue rather than hiding the failing tests.

- Jonathan M Davis

I am not sure why they were added in the first place (couldn't find anything during a quick look at the commit logs and mailing list archives), but the problem were some OS X special cases in the code, namely using stat64() instead of lstat64().

I opened pull request 160 (https://github.com/D-Programming-Language/phobos/pull/160) to fix that ? if there was a reason I missed for not using lstat() on OS X or not running the other unit tests, please disregard it.

David


On 7/30/11 9:10 AM, Jonathan M Davis wrote:
> For some reason, the unit tests on Mac OS X are failing for std.file.symlink. They pass just fine on Linux and FreeBSD, but the the Mac isn't so happy with them.
>
> http://d.puremagic.com/test-results/
>
> Does anyone with a Mac have time to look at it? It tries to create a symlink to /usr/include and verify that it worked, and it claims that the file exists after calling symlink, but it also claims that it's not a symlink (hence the failure). I have no access to a Mac, and so I can't do anything to fix it. We can disable to the failing unit tests if we have to, but we should really fix the issue rather than hiding the failing tests.
>
> - Jonathan M Davis
> _______________________________________________
> phobos mailing list
> phobos at puremagic.com
> http://lists.puremagic.com/mailman/listinfo/phobos

Line 1821 in std.file, shouldn't it be "buffer.ptr" instead of just "buffer"?

On 30 jul 2011, at 10:22, David Nadlinger wrote:

> I am not sure why they were added in the first place (couldn't find anything during a quick look at the commit logs and mailing list archives), but the problem were some OS X special cases in the code, namely using stat64() instead of lstat64().
> 
> I opened pull request 160 (https://github.com/D-Programming-Language/phobos/pull/160) to fix that ? if there was a reason I missed for not using lstat() on OS X or not running the other unit tests, please disregard it.
> 
> David
> 
> 
> On 7/30/11 9:10 AM, Jonathan M Davis wrote:
>> For some reason, the unit tests on Mac OS X are failing for std.file.symlink. They pass just fine on Linux and FreeBSD, but the the Mac isn't so happy with them.
>> 
>> http://d.puremagic.com/test-results/
>> 
>> Does anyone with a Mac have time to look at it? It tries to create a symlink to /usr/include and verify that it worked, and it claims that the file exists after calling symlink, but it also claims that it's not a symlink (hence the failure). I have no access to a Mac, and so I can't do anything to fix it. We can disable to the failing unit tests if we have to, but we should really fix the issue rather than hiding the failing tests.
>> 
>> - Jonathan M Davis
>> _______________________________________________
>> phobos mailing list
>> phobos at puremagic.com
>> http://lists.puremagic.com/mailman/listinfo/phobos
> 
> _______________________________________________
> phobos mailing list
> phobos at puremagic.com
> http://lists.puremagic.com/mailman/listinfo/phobos

-- 
/Jacob Carlborg