Re: [PATCH][gdb/testsuite] Fix gdb.cp/break-f-std-string.cc with older gcc

[Pedro Alves <pedro@palves.net>]

On 2022-05-12 12:45, Tom de Vries wrote:

> [gdb/testsuite] Fix gdb.cp/break-f-std-string.cc with older gcc
> 
> ---
>  gdb/testsuite/gdb.cp/break-f-std-string.cc | 14 ++++++++++++++
>  1 file changed, 14 insertions(+)
> 
> diff --git a/gdb/testsuite/gdb.cp/break-f-std-string.cc b/gdb/testsuite/gdb.cp/break-f-std-string.cc
> index 454ab4c42ea..cbbfeebac60 100644
> --- a/gdb/testsuite/gdb.cp/break-f-std-string.cc
> +++ b/gdb/testsuite/gdb.cp/break-f-std-string.cc
> @@ -17,6 +17,20 @@
>  
>  #include <string>
>  
> +#if _GLIBCXX_USE_CXX11_ABI == 1
> +#if defined (__GNUC__) && (__GNUC__ >= 5) && (__GNUC__ <= 8)
> +
> +// Work around missing std::string typedef before gcc commit
> +// "Define std::string and related typedefs outside __cxx11 namespace".

Since we try to follow the GDB conventions in the tests too, this should
use /**/ comments.

> +
> +namespace std {
> +using namespace __cxx11;

Why is this "using namespace __cxx11;" needed?  I guess something about the
"related typedefs" mentioned in the comment?  The test passes for me without
that, on GCC 6.5 (the only affected compiler I have handy atm), like:

namespace std {
typedef __cxx11::string string;
}

Otherwise LGTM.

Pedro Alves

> +typedef __cxx11::string string;
> +}
> +
> +#endif
> +#endif
> +
>  void
>  f (std::string s)
>  {
>