Fix build on Linux IA-64

The clone call used by death tests is not available on IA-64 architectures. see http://linux.die.net/man/2/clone

This patch merely disables death tests for IA-64.

[chandlerc, 2008-12-02 18:09:59]

Tiny nit, but otherwise this looks good.

http://codereview.appspot.com/8690/diff/1/2
File include/gtest/internal/gtest-port.h (right):

http://codereview.appspot.com/8690/diff/1/2#newcode332
Line 332: #if GTEST_HAS_STD_STRING && defined(GTEST_OS_LINUX) && GTEST_HAS_CLONE
Can we group the GTEST_HAS_* checks? I'd put the OS check first, but either way.
Just feels odd with them split.

[Hobbes, 2008-12-02 18:16:05]

reordered checks

[Hobbes, 2008-12-02 18:18:12]

and now fixing the newlines

[Zhanyong, 2008-12-02 18:27:07]

http://codereview.appspot.com/8690/diff/9/10
File include/gtest/internal/gtest-port.h (right):

http://codereview.appspot.com/8690/diff/9/10#newcode319
Line 319: // On IA-64 for example it isn't: see
http://linux.die.net/man/2/clone2
Please also say that non-Linux systems usually don't support it.

Nit: comments should be English sentences that end with a period.

http://codereview.appspot.com/8690/diff/9/10#newcode326
Line 326: #define GTEST_HAS_CLONE 1
clone() is a Linux thing.  We cannot define this to 1 if we are not on Linux. 
Please define it to 1 only if we are on Linux and __ia64__ is not defined.

Add an entry for this macro in the big block comment above (line 43).  Next to
other GTEST_HAS_* macros.  Now that the list is longer, we should sort it
alphabetically.

http://codereview.appspot.com/8690/diff/9/10#newcode332
Line 332: #if defined(GTEST_OS_LINUX) && GTEST_HAS_STD_STRING && GTEST_HAS_CLONE
Remove GTEST_OS_LINUX after you fix the definition of GTEST_HAS_CLONE.

[Hobbes, 2008-12-02 19:09:07]

fixing comments, reordering GTEST_HAS_ macros, only defining GTEST_HAS_CLONE on
Linux/non-IA64

[chandlerc, 2008-12-02 19:16:29]

LGTM

Much nicer, thanks!

http://codereview.appspot.com/8690/diff/14/15
File include/gtest/internal/gtest-port.h (right):

http://codereview.appspot.com/8690/diff/14/15#newcode43
Line 43: //   GTEST_HAS_CLONE          - Define it to 1/0 to indicate clone(2)
indicate _that_ clone(2) ...

[Zhanyong, 2008-12-02 20:01:45]

Looks good.

I'll do some testing and then commit it to trunk.  Cheers!

http://codereview.appspot.com/8690/diff/14/15
File include/gtest/internal/gtest-port.h (right):

http://codereview.appspot.com/8690/diff/14/15#newcode43
Line 43: //   GTEST_HAS_CLONE          - Define it to 1/0 to indicate clone(2)
On 2008/12/02 19:16:29, chandlerc wrote:
> indicate _that_ clone(2) ...

I can fix this when I commit the patch.

[Hobbes, 2008-12-02 20:09:44]

http://codereview.appspot.com/8690/diff/14/15
File include/gtest/internal/gtest-port.h (right):

http://codereview.appspot.com/8690/diff/14/15#newcode43
Line 43: //   GTEST_HAS_CLONE          - Define it to 1/0 to indicate clone(2)
On 2008/12/02 20:01:46, Zhanyong wrote:
> On 2008/12/02 19:16:29, chandlerc wrote:
> > indicate _that_ clone(2) ...
> 
> I can fix this when I commit the patch.

Thanks. You might wanna fix GTEST_HAS_TR1_TUPLE as well.