Re: [PATCH 2/7] PR gdb/15224 fix for Unify interactivity tests to use input_from_terminal_p

[Pedro Alves <palves@redhat.com>]

On 05/09/2013 12:08 PM, mbilal wrote:
> On Wednesday, May 08, 2013 9:22 PM Pedro Alves wrote:
> 
>>I'd guess that input_from_terminal_p already does some of the
>>checks this is also making, so that this could be simplified.
> 
> I don't think that because input_from_terminal_p
> checks the condition instream==stdin that
> is equal in both sanario , but input_from_terminal_p
> also return 1 when instream==NULL.
> So I think that there is no any condition that is overlapping with
> input_from_terminal_p.

  /* If INSTREAM is unset, and we are not in a user command, we
     must be in Insight.  That's like having a terminal, for our
     purposes.  */
  if (instream == NULL && !in_user_command)
    return 1;

This very much looks like a case where history should work, but
won't unless we simplify to check input_from_terminal_p only.

The point of unifying interactivity tests is exactly so that
all these little details are centralized in a single place.

Doesn't

-    if (instream == stdin
-      && ISATTY (stdin) && *linebuffer)
+    if (*linebuffer && input_from_terminal_p ())

work?  If not, why not?  It'd be great to try that on
GNU/Linux, mingw32 gdb running under cmd.exe, and mingw32 gdb under
Cygwin bash, the latter with and without "set interactive-mode on".
I suspect this change would fix history for the forced
"set interactive-mode on" case too.

> Index: event-top.c
> ===================================================================
> RCS file: /cvs/src/src/gdb/event-top.c,v
> retrieving revision 1.94
> diff -u -p -r1.94 event-top.c
> --- event-top.c    21 Mar 2013 17:37:29 -0000    1.94
> +++ event-top.c    9 May 2013 10:23:21 -0000
> @@ -604,7 +604,8 @@ command_line_handler (char *rl)
> 
>    /* Add line to history if appropriate.  */
>    if (instream == stdin
> -      && ISATTY (stdin) && *linebuffer)
> +      && ISATTY (stdin) && *linebuffer
> +      && input_from_terminal_p ())
>      add_history (linebuffer);

Please grep for add_history.  There's at least one more call in
the tree.

-- 
Pedro Alves