[2/3] RFC: *stopped and CLI commands

[2/3] RFC: *stopped and CLI commands

[Tom Tromey]

I noticed while working on the PR that if you type "run" to an MI-mode
gdb, then the *stopped notification will not have any 'reason' field, or
other fields emitted by the print_it method.

The fix is simple: call bpstat_print at the right place in MI.
There is already a little special case in this code for printing the
frame, so I think a bit more here doesn't hurt.

The test case is in the next patch.

Tom

2011-11-16  Tom Tromey  <tromey@redhat.com>

	* mi/mi-interp.c (mi_on_normal_stop): Call bpstat_print.

From 9554d74f7a6b4b0db7f416b970ae94f55e4e8aa0 Mon Sep 17 00:00:00 2001
From: Tom Tromey <tromey@redhat.com>
Date: Tue, 15 Nov 2011 14:35:23 -0700
Subject: [PATCH 2/3] make *stopped show reason for CLI commands as well

---
 gdb/ChangeLog      |    4 ++++
 gdb/mi/mi-interp.c |    1 +
 2 files changed, 5 insertions(+), 0 deletions(-)

diff --git a/gdb/mi/mi-interp.c b/gdb/mi/mi-interp.c
index 805e3f8..2aec1c7 100644
--- a/gdb/mi/mi-interp.c
+++ b/gdb/mi/mi-interp.c
@@ -429,6 +429,7 @@ mi_on_normal_stop (struct bpstats *bs, int print_frame)
 	  struct ui_out *saved_uiout = current_uiout;
 
 	  current_uiout = mi_uiout;
+	  bpstat_print (bs);
 	  print_stack_frame (get_selected_frame (NULL), 0, SRC_AND_LOC);
 	  current_uiout = saved_uiout;
 	}
-- 
1.7.6.4

Re: [2/3] RFC: *stopped and CLI commands

[Pedro Alves]

On Wednesday 16 November 2011 20:00:30, Tom Tromey wrote:
> diff --git a/gdb/mi/mi-interp.c b/gdb/mi/mi-interp.c
> index 805e3f8..2aec1c7 100644
> --- a/gdb/mi/mi-interp.c
> +++ b/gdb/mi/mi-interp.c
> @@ -429,6 +429,7 @@ mi_on_normal_stop (struct bpstats *bs, int print_frame)
>           struct ui_out *saved_uiout = current_uiout;
>  
>           current_uiout = mi_uiout;
> +         bpstat_print (bs);
>           print_stack_frame (get_selected_frame (NULL), 0, SRC_AND_LOC);
>           current_uiout = saved_uiout;
>         }

Hmm, I should have read this one first.  This means that with
patch 3 we'll print "Stopped due to ..." twice?

-- 
Pedro Alves

Re: [2/3] RFC: *stopped and CLI commands

[Tom Tromey]

>>>>> "Pedro" == Pedro Alves <pedro@codesourcery.com> writes:

Pedro> On Wednesday 16 November 2011 20:00:30, Tom Tromey wrote:
>> diff --git a/gdb/mi/mi-interp.c b/gdb/mi/mi-interp.c
>> index 805e3f8..2aec1c7 100644
>> --- a/gdb/mi/mi-interp.c
>> +++ b/gdb/mi/mi-interp.c
>> @@ -429,6 +429,7 @@ mi_on_normal_stop (struct bpstats *bs, int print_frame)
>> struct ui_out *saved_uiout = current_uiout;
>> 
>> current_uiout = mi_uiout;
>> +         bpstat_print (bs);
>> print_stack_frame (get_selected_frame (NULL), 0, SRC_AND_LOC);
>> current_uiout = saved_uiout;
>> }

Pedro> Hmm, I should have read this one first.  This means that with
Pedro> patch 3 we'll print "Stopped due to ..." twice?

Yes, I hadn't noticed that.  Oops.

However, we still need this call in order to get the reason= output in
the MI notification when the command was a CLI command.  That is, if you
omit this call, you get:

    (gdb) 
    run
    &"run\n"
    ~"Starting program: /home/tromey/Space/SecondArcher/build/gdb/testsuite/gdb.mi/solib-main \n"
    =thread-group-started,id="i1",pid="32683"
    =thread-created,id="1",group-id="i1"
    ^running
    *running,thread-id="all"
    (gdb) 
    =library-loaded,id="/lib64/ld-linux-x86-64.so.2",target-name="/lib64/ld-linux-x86-64.so.2",host-name="/lib64/ld-linux-x86-64.so.2",symbols-loaded="0",thread-group="i1"
    ~"Stopped due to shared library event\n"
    *stopped,frame={addr="0x0000003a15c0ea20",func="__GI__dl_debug_state",args=[],file="dl-debug.c",fullname="/usr/src/debug/glibc-2.14-34-ge2a3090/elf/dl-debug.c",line="77"},thread-id="1",stopped-threads="all",core="2"



Note how the *stopped doesn't have a reason.

The fix is to leave this call to bpstat_print, but to change the
printing of "Stopped due to shared library event" to use ui_out_text
rather than printf_filtered.  This eliminates the double printing.

Tom

Re: [2/3] RFC: *stopped and CLI commands

[Pedro Alves]

On Tuesday 22 November 2011 20:05:17, Tom Tromey wrote:
> >>>>> "Pedro" == Pedro Alves <pedro@codesourcery.com> writes:

> Pedro> Hmm, I should have read this one first.  This means that with
> Pedro> patch 3 we'll print "Stopped due to ..." twice?
> 
> Yes, I hadn't noticed that.  Oops.
> 
> However, we still need this call in order to get the reason= output in
> the MI notification when the command was a CLI command.  That is, if you
> omit this call, you get:

...

> Note how the *stopped doesn't have a reason.
> 
> The fix is to leave this call to bpstat_print, but to change the
> printing of "Stopped due to shared library event" to use ui_out_text
> rather than printf_filtered.  This eliminates the double printing.

Yeah, makes sense.

-- 
Pedro Alves