Re: [PATCH v3 3/3] gdb/amdgpu: add precise-memory support

[Lancelot SIX <lsix@lancelotsix.com>]

Hi Simon,

It looks like your latest change is not c++11 compliant (see below).

I do have a WIP branch to bump the C++ version to C++17
(users/lsix/try-c++17).  Maybe it's time I write proper commit messages
and send this as RFC, see if there is a concensis to allow c++17.

> diff --git a/gdb/amd-dbgapi-target.c b/gdb/amd-dbgapi-target.c
> index 22c269b7992c..507824decf8d 100644
> --- a/gdb/amd-dbgapi-target.c
> +++ b/gdb/amd-dbgapi-target.c
> @@ -116,8 +117,9 @@ get_amd_dbgapi_target_inferior_created_observer_token ()
>  
>  struct amd_dbgapi_inferior_info
>  {
> -  explicit amd_dbgapi_inferior_info (inferior *inf)
> -    : inf (inf)
> +  explicit amd_dbgapi_inferior_info (inferior *inf,
> +				     bool precise_memory_requested = false)
> +    : inf {inf}, precise_memory {precise_memory_requested}
>    {}

If I build using -std=c++11 I get the following error:

../../gdb/amd-dbgapi-target.c: In constructor ‘amd_dbgapi_inferior_info::amd_dbgapi_inferior_info(inferior*, bool)’:
../../gdb/amd-dbgapi-target.c:122:18: error: no matching function for call to ‘amd_dbgapi_inferior_info::<unnamed struct>::._anon_207(<brace-enclosed initializer list>)’
  122 |     : inf {inf}, precise_memory {precise_memory_requested}
      |                  ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
../../gdb/amd-dbgapi-target.c:145:3: note: candidate: ‘amd_dbgapi_inferior_info::<unnamed struct>::<constructor>()’
  145 |   {
      |   ^

In c++17 mode, it works fine though…

Moving the initialization in the ctor's body would solve the issue

  explicit amd_dbgapi_inferior_info (inferior *inf,
                                     bool precise_memory_requested = false)
    : inf {inf}
  {
    precise_memory.requested = precise_memory_requested;
  }

The alternative would be to name
amd_dbgapi_inferior_info::precise_memory's type and give it a ctor.

Best,
Lancelot.

>  
>    /* Backlink to inferior.  */