Re: [RFC 2/3] gdb: wait for inferior after attaching instead of adding a continuation

[Tom Tromey <tom@tromey.com>]

>>>>> "Tankut" == Tankut Baris Aktemur <tankut.baris.aktemur@intel.com> writes:

Tankut> In a less harmful case, the current implementation causes an unnatural
Tankut> positioning of the GDB prompt in a single target setting when
Tankut> executing the attach command in async mode using the '&' operator.

I'd prefer Pedro review this, since he's the infrun expert, but I was
curious about this point -- it seems to me that "attach &" should be
expected not to wait.

So maybe the issue is that somehow a continuation is being run for the
wrong inferior?  Like in this part of the final patch:

-      /* Do all continuations associated with the whole inferior (not
-	 a particular thread).  */
-      if (inferior_ptid != null_ptid)
-	current_inferior ()->do_all_continuations ();

... do inferior_ptid and current_inferior disagree here?
Or else why does this code get invoked?

Tom