Re: [RFC] Implement -list-thread-groups.

[Michael Snyder <msnyder@vmware.com>]

> diff --git a/gdb/thread.c b/gdb/thread.c
> index b1e318d..a32dff4 100644
> --- a/gdb/thread.c
> +++ b/gdb/thread.c
> @@ -636,9 +636,10 @@ set_stop_requested (ptid_t ptid, int stop)
>     use from MI.  
>     If REQUESTED_THREAD is not -1, it's the GDB id of the thread
>     that should be printed.  Otherwise, all threads are
> -   printed.  */
> +   printed.  
> +   If PID is not -1, only prints threads from the process PID.  */

For completeness and consistency, how about adding
"Otherwise, threads from all attached PIDs are printed."

>  void
> -print_thread_info (struct ui_out *uiout, int requested_thread)
> +print_thread_info (struct ui_out *uiout, int requested_thread, int pid)
>  {
>    struct thread_info *tp;
>    ptid_t current_ptid;
> @@ -646,6 +647,8 @@ print_thread_info (struct ui_out *uiout, int requested_thread)
>    char *extra_info;
>    int current_thread = -1;
>  
> +  gdb_assert (requested_thead == -1 || pid == -1);

I'm puzzled by this assert.
You don't think we'll ever want to specify both the pid and the thread?


Otherwise, these changes to thread.c seem fine.