Just check for C++ version.

Just check for C++ version.

This fixes build breakage introduced with r15736.

TBR=svenpanne@chromium.org

Committed: https://code.google.com/p/v8/source/detail?r=15737

[Benedikt Meurer, 2013-07-18 07:50:38]

Committed patchset #1 manually as r15737 (presubmit successful).

[Sven Panne, 2013-07-18 07:57:57]

lgtm

https://codereview.chromium.org/19671009/diff/1/src/checks.h
File src/checks.h (left):

https://codereview.chromium.org/19671009/diff/1/src/checks.h#oldcode236
src/checks.h:236: (defined(__has_feature) && __has_feature(cxx_static_assert))
|| \
If the need for the additional terms of the disjunction comes up again, we could
do it differently:

#if _cplusplus >= 201103L
#define HAS_STATIC_ASSERT 1
#endif

#if defined(__has_feature)
#if __has_feature(cxx_static_assert)
#define HAS_STATIC_ASSERT 1
#endif
#endif
...
(same for cxx_static_assert)
(use HAS_STATIC_ASSERT below)

[Benedikt Meurer, 2013-07-18 08:01:40]

On 2013/07/18 07:57:57, Sven Panne wrote:
> lgtm
> 
> https://codereview.chromium.org/19671009/diff/1/src/checks.h
> File src/checks.h (left):
> 
> https://codereview.chromium.org/19671009/diff/1/src/checks.h#oldcode236
> src/checks.h:236: (defined(__has_feature) && __has_feature(cxx_static_assert))
> || \
> If the need for the additional terms of the disjunction comes up again, we
could
> do it differently:
> 
> #if _cplusplus >= 201103L
> #define HAS_STATIC_ASSERT 1
> #endif
> 
> #if defined(__has_feature)
> #if __has_feature(cxx_static_assert)
> #define HAS_STATIC_ASSERT 1
> #endif
> #endif

That should be in #elif then. :-)

[Sven Panne, 2013-07-18 08:04:45]

On 2013/07/18 08:01:40, Benedikt Meurer wrote:
> That should be in #elif then. :-)

To continue the nit-picking: I think it should even be OK without #elif, because
redefining macros with the same text is OK IIRC. OTOH, the preprocessor alone is
more complex than any complete programming language should be, so who knows...
:-D

[Benedikt Meurer, 2013-07-18 08:05:31]

On 2013/07/18 08:04:45, Sven Panne wrote:
> On 2013/07/18 08:01:40, Benedikt Meurer wrote:
> > That should be in #elif then. :-)
> 
> To continue the nit-picking: I think it should even be OK without #elif,
because
> redefining macros with the same text is OK IIRC. OTOH, the preprocessor alone
is
> more complex than any complete programming language should be, so who knows...
> :-D

I'm pretty sure we'll find compilers with preprocessors that are going to bail
out here. :-)