libstdc++ test case ext/headers.cc failed on arm-none-eabi

[Jonathan Wakely]

On 7 August 2013 14:34, Bin.Cheng wrote:
>
> Thanks all of you for helping.
> Since I know little about c++, here comes more questions:
>
> 1) for the warning message:
> g.h:2:14: warning: non-static data member initializers only available
> with -std=c++11 or -std=gnu++11 [enabled by default]
>  #define INIT 0
>               ^
> What does the "[enabled by default]" mean? If "-std=gnu++11" is
> enabled by default, why gcc gives this warning message then?

It means the warning is enabled by default, not -std=gnu++11 is
enabled by default. That's nothing to do with C++, it's how GCC
displays warnings.

> 2) The case works for x86 regression test, could this be a target (or
> target testing setup) specific problem? What kind of imformation
> should I provide to identify the issue?

It seems to depend on the 'single' thread model, i.e. gthr-single.h

> 3) According to the explanation, this possibly reveals two issues, one
> for the warning message itself, and the other for the inaccurate
> warning message, right?

There's an existing PR for the innacurate diagnostic.