Flush ^running

Flush ^running

[Vladimir Prus]

[-- Attachment #1: Type: text/plain, Size: 355 bytes --]


I've run into a case where gdb would not print ^running in response
to -exec-continue -- at least not during the time frontend is willing
to wait for the response. The problem only happened for my when gdb
is driven by a frontend, not on command line; I don't know what frontend
does differently. I've checked in the below patch to fix this.

- Volodya


[-- Attachment #2: flush.diff --]
[-- Type: text/x-diff, Size: 1022 bytes --]

Index: gdb/ChangeLog
===================================================================
RCS file: /cvs/src/src/gdb/ChangeLog,v
retrieving revision 1.9872
diff -u -p -r1.9872 ChangeLog
--- gdb/ChangeLog	3 Oct 2008 22:00:09 -0000	1.9872
+++ gdb/ChangeLog	4 Oct 2008 14:56:26 -0000
@@ -1,3 +1,7 @@
+2008-10-04  Vladimir Prus  <vladimir@codesourcery.com>
+
+	* mi/mi-interp.c (mi_on_resume): Flush raw_stdout.
+
 2008-10-03  Paul Pluzhnikov  <ppluzhnikov@google.com>
 
 	PR gdb/2384:
Index: gdb/mi/mi-interp.c
===================================================================
RCS file: /cvs/src/src/gdb/mi/mi-interp.c,v
retrieving revision 1.35
diff -u -p -r1.35 mi-interp.c
--- gdb/mi/mi-interp.c	5 Jul 2008 17:57:24 -0000	1.35
+++ gdb/mi/mi-interp.c	4 Oct 2008 14:56:26 -0000
@@ -362,6 +362,7 @@ mi_on_resume (ptid_t ptid)
       if (!target_is_async_p ())
 	fputs_unfiltered ("(gdb) \n", raw_stdout);
     }
+  gdb_flush (raw_stdout);
 }
 
 extern initialize_file_ftype _initialize_mi_interp; /* -Wmissing-prototypes */

Re: Flush ^running

[Daniel Jacobowitz]

On Sat, Oct 04, 2008 at 07:01:40PM +0400, Vladimir Prus wrote:
> 
> I've run into a case where gdb would not print ^running in response
> to -exec-continue -- at least not during the time frontend is willing
> to wait for the response. The problem only happened for my when gdb
> is driven by a frontend, not on command line; I don't know what frontend
> does differently. I've checked in the below patch to fix this.

Try "cat | gdb -i=mi 2>&1 | cat" to reproduce this sort of thing.
If isatty() returns true, most C runtime libraries will put stdout in
unbuffered mode.

-- 
Daniel Jacobowitz
CodeSourcery

RE: Flush ^running

[Marc Khouzam]

Hi,

I've also run into that problem.

When continuing or stepping over a sleep(), I would only see the
^running and *running
after the sleep() call had finished. Just like you , it only happened
when using
the frontend, although is happened with CLI commands from the frontend
(continue, next).

Your patch fixes this issue.

One less thing to worry about.
Thanks!

Marc




> -----Original Message-----
> From: gdb-patches-owner@sourceware.org 
> [mailto:gdb-patches-owner@sourceware.org] On Behalf Of Vladimir Prus
> Sent: Saturday, October 04, 2008 11:02 AM
> To: gdb-patches@sources.redhat.com
> Subject: Flush ^running
> 
> 
> I've run into a case where gdb would not print ^running in response
> to -exec-continue -- at least not during the time frontend is willing
> to wait for the response. The problem only happened for my when gdb
> is driven by a frontend, not on command line; I don't know 
> what frontend
> does differently. I've checked in the below patch to fix this.
> 
> - Volodya
> 
>