[v3] unify 3_4 and 3_5 libstdc++

[Paolo Carlini]

Paolo Carlini wrote:

> I'm now trying to isolate the problem, by varying the configure options.

Ok, passing --enable-libstdcxx-allocator=new happens this:

x86-64
------
XPASS: 26_numerics/c99_classification_macros_c.cc (test for excess errors)
XPASS: 27_io/fpos/14320-1.cc execution test

                === libstdc++ Summary ===

# of expected passes            2629
# of unexpected successes       2
# of expected failures          4

ia64
----
FAIL: 22_locale/num_put/put/char/wrapped_env.cc execution test
FAIL: 22_locale/num_put/put/char/wrapped_locale.cc execution test
FAIL: 22_locale/num_put/put/wchar_t/4.cc execution test
FAIL: 22_locale/num_put/put/wchar_t/wrapped_env.cc execution test
FAIL: 22_locale/num_put/put/wchar_t/wrapped_locale.cc execution test
XPASS: 26_numerics/c99_classification_macros_c.cc (test for excess errors)
XPASS: 27_io/fpos/14320-1.cc execution test

                === libstdc++ Summary ===

# of expected passes            2624
# of unexpected failures        5
# of unexpected successes       2
# of expected failures          4

=========

Definitely, *much* better. Therefore, the allocator seems implied.

Paolo.

P.S. Notice that num_put/put/*/4.cc (involved also in the wrapped_* tests)
uses locale + imbue + stringstream in a way completely different from 
that of
the other testcases, and maybe slightly defective. More about this later 
today.