Re: [Boost-testing] Sandia-darwin-intel and... Sandia-darwin-intel!

From: Gennaro Prota (gennaro.prota_at_[hidden])
Date: 2008-09-02 15:32:43

K. Noel Belcourt wrote:
> 
> On Sep 2, 2008, at 11:01 AM, Gennaro Prota wrote:
> 
>> what's the difference between Sandia-darwin-intel-9.1 and
> 
> This is the intel-9.1 compiler on x86 hardware.
> 
>> Sandia-darwin-intel at
> 
> This is Apple's gcc-4.0.1 compiler (Boost build refers to this as the 
> darwin toolset, or, as it appears in the results page, the darwin-4.0.1 
> toolset).

Ah! :-) Thanks. Perhaps the top row names (or the toolset names) 
should be a tad standardized? I couldn't gather that it was gcc.

>> <http://beta.boost.org/development/tests/trunk/developer/dynamic_bitset.html>? 
>>
>>
>> Could the exact version of the Intel compiler (with build number) be
>> provided for the latter?
> 
> It's Apple's gcc compiler on x86 hardware, running under Tiger.
> 
>> BTW, almost all Intel versions are failing to recognize a constant
>> expression as such (sigh). To avoid messing up the SVN history, it
>> would be great if someone could help with a little local test. I'd
>> send the patch here in that case.
> 
> What can I do to help?

Could you please apply the attached patch to your trunk working copy 
and see if dyn_bitset_unit_tests3 compiles successfully? This is the 
classical "rephrase and move things around until the compiler is 
happy" fix, so I'd prefer not mess up with SVN, unless unavoidable.

Thanks a lot!

-- 
Genny

Index: dynamic_bitset.hpp
===================================================================
--- dynamic_bitset.hpp	(revision 48496)
+++ dynamic_bitset.hpp	(working copy)
@@ -1001,7 +1001,10 @@
 typename dynamic_bitset<Block, Allocator>::size_type
 dynamic_bitset<Block, Allocator>::count() const
 {
-    using namespace detail::dynamic_bitset_impl;
+    using detail::dynamic_bitset_impl::access_by_bytes;
+    using detail::dynamic_bitset_impl::access_by_blocks;
+    using detail::dynamic_bitset_impl::table_width;
+    using detail::dynamic_bitset_impl::value_to_type;
 
     // NOTE: I'm explicitly qualifying "bits_per_block" to workaround
     //       regressions of gcc 3.4.x