[v2,1/3] gdb: Rename gdb_flush to ui_file_flush.

‐‐‐‐‐‐‐ Original Message ‐‐‐‐‐‐‐
On Wednesday, 27 November 2019 00:00, Iain Buclaw <ibuclaw@gdcproject.org> wrote:

> ‐‐‐‐‐‐‐ Original Message ‐‐‐‐‐‐‐
> On Tuesday, 26 November 2019 21:24, Pedro Alves palves@redhat.com wrote:
>
> > On 11/26/19 12:49 PM, Iain Buclaw wrote:
> >
> > > The significance of this is that printf_unfiltered writes messages to wrap_buffer, whereas puts_unfiltered pushes them immediately to stdout, resulting in "post-" messages being printed out of order.
> >
> > It sounds quite surprising that two _unfiltered functions could behave differently
> > like that. That sounds like a bug that should be fixed, instead of worked around
> > by having to recall to use printf vs puts.
> > Thanks,
> > Pedro Alves
>
> I think the best way to avoid the discrepancy is to treat both fputs_filtered and fputs_unfiltered equally by forwarding both calls to fputs_maybe_filtered.
>
> To avoid recursion, flush_wrap_buffer and fputs_maybe_filtered have had calls to fputs_unfiltered replaced with stream->puts().
>
> While attempting to grok my head around fputs_maybe_filtered, I also noticed that buffer_clearer is being removed by the compiler as dead code.
>

Except that doesn't work as some parts of gdb require that direct writing/flushing stdout is kept in place, so what I ultimately ended up doing is defining two new functions for this, then repurposing fputs_unfiltered and gdb_flush to interact with fputs_maybe_unfiltered.

I've split the patch into two logical parts, each dealing with one function at a time.  Running the testsuitelocally, I can see no new regressions as a result of applying this.

This first patch renames gdb_flush to ui_file_flush.  Redefining gdb_flush in utils.c, with a new behaviour to flush the wrap_buffer if necessary before flushing the stream.

--
Iain

---
gdb/ChangeLog:

2019-11-28  Iain Buclaw  <ibuclaw@gdcproject.org>

        * gdb/event-loop.c (gdb_wait_for_event): Update.
        * gdb/printcmd.c (printf_command): Update.
        * gdb/remote-fileio.c (remote_fileio_func_write): Update.
        * gdb/remote-sim.c (gdb_os_flush_stdout): Update.
        (gdb_os_flush_stderr): Update.
        * gdb/remote.c (remote_console_output): Update.
        * gdb/ui-file.c (gdb_flush): Rename to...
        (ui_file_flush): ...this.
        (stderr_file::write): Update.
        (stderr_file::puts): Update.
        * gdb/ui-file.h (gdb_flush): Rename to...
        (ui_file_flush): ...this.
        * gdb/utils.c (gdb_flush): Add function.
        * gdb/utils.h (gdb_flush): Add declaration.

---

* Iain Buclaw <ibuclaw@gdcproject.org> [2019-11-28 23:53:48 +0000]:

> ‐‐‐‐‐‐‐ Original Message ‐‐‐‐‐‐‐
> On Wednesday, 27 November 2019 00:00, Iain Buclaw <ibuclaw@gdcproject.org> wrote:
> 
> > ‐‐‐‐‐‐‐ Original Message ‐‐‐‐‐‐‐
> > On Tuesday, 26 November 2019 21:24, Pedro Alves palves@redhat.com wrote:
> >
> > > On 11/26/19 12:49 PM, Iain Buclaw wrote:
> > >
> > > > The significance of this is that printf_unfiltered writes messages to wrap_buffer, whereas puts_unfiltered pushes them immediately to stdout, resulting in "post-" messages being printed out of order.
> > >
> > > It sounds quite surprising that two _unfiltered functions could behave differently
> > > like that. That sounds like a bug that should be fixed, instead of worked around
> > > by having to recall to use printf vs puts.
> > > Thanks,
> > > Pedro Alves
> >
> > I think the best way to avoid the discrepancy is to treat both fputs_filtered and fputs_unfiltered equally by forwarding both calls to fputs_maybe_filtered.
> >
> > To avoid recursion, flush_wrap_buffer and fputs_maybe_filtered have had calls to fputs_unfiltered replaced with stream->puts().
> >
> > While attempting to grok my head around fputs_maybe_filtered, I also noticed that buffer_clearer is being removed by the compiler as dead code.
> >
> 
> Except that doesn't work as some parts of gdb require that direct
> writing/flushing stdout is kept in place, so what I ultimately ended
> up doing is defining two new functions for this, then repurposing
> fputs_unfiltered and gdb_flush to interact with
> fputs_maybe_unfiltered.
> 
> I've split the patch into two logical parts, each dealing with one function at a time.  Running the testsuitelocally, I can see no new regressions as a result of applying this.
> 
> This first patch renames gdb_flush to ui_file_flush.  Redefining gdb_flush in utils.c, with a new behaviour to flush the wrap_buffer if necessary before flushing the stream.
> 
> --
> Iain
> 
> ---
> gdb/ChangeLog:
> 
> 2019-11-28  Iain Buclaw  <ibuclaw@gdcproject.org>
> 
>         * gdb/event-loop.c (gdb_wait_for_event): Update.
>         * gdb/printcmd.c (printf_command): Update.
>         * gdb/remote-fileio.c (remote_fileio_func_write): Update.
>         * gdb/remote-sim.c (gdb_os_flush_stdout): Update.
>         (gdb_os_flush_stderr): Update.
>         * gdb/remote.c (remote_console_output): Update.
>         * gdb/ui-file.c (gdb_flush): Rename to...
>         (ui_file_flush): ...this.
>         (stderr_file::write): Update.
>         (stderr_file::puts): Update.
>         * gdb/ui-file.h (gdb_flush): Rename to...
>         (ui_file_flush): ...this.
>         * gdb/utils.c (gdb_flush): Add function.
>         * gdb/utils.h (gdb_flush): Add declaration.

Iain,  thanks for looking at this issue.

When posting patches for review you should include the commit message
for review too.  This should consist of a concise but informative
title, a description of the problem, maybe how you discovered it, or
what problems it was causing, and then an overview of the solution.

> 
> ---
> 

> diff --git a/gdb/event-loop.c b/gdb/event-loop.c
> index affa00b4aa..0d12ac9d41 100644
> --- a/gdb/event-loop.c
> +++ b/gdb/event-loop.c
> @@ -747,8 +747,8 @@ gdb_wait_for_event (int block)
>    int num_found = 0;
>  
>    /* Make sure all output is done before getting another event.  */
> -  gdb_flush (gdb_stdout);
> -  gdb_flush (gdb_stderr);
> +  ui_file_flush (gdb_stdout);
> +  ui_file_flush (gdb_stderr);
>  
>    if (gdb_notifier.num_fds == 0)
>      return -1;
> diff --git a/gdb/printcmd.c b/gdb/printcmd.c
> index fe0efd371a..e0295940d8 100644
> --- a/gdb/printcmd.c
> +++ b/gdb/printcmd.c
> @@ -2718,7 +2718,7 @@ printf_command (const char *arg, int from_tty)
>    ui_printf (arg, gdb_stdout);
>    reset_terminal_style (gdb_stdout);
>    wrap_here ("");
> -  gdb_flush (gdb_stdout);
> +  ui_file_flush (gdb_stdout);
>  }
>  
>  /* Implement the "eval" command.  */
> diff --git a/gdb/remote-fileio.c b/gdb/remote-fileio.c
> index bc7c71ffdd..e2ee311bf3 100644
> --- a/gdb/remote-fileio.c
> +++ b/gdb/remote-fileio.c
> @@ -641,7 +641,7 @@ remote_fileio_func_write (remote_target *remote, char *buf)
>        case FIO_FD_CONSOLE_OUT:
>  	ui_file_write (target_fd == 1 ? gdb_stdtarg : gdb_stdtargerr,
>  		       (char *) buffer, length);
> -	gdb_flush (target_fd == 1 ? gdb_stdtarg : gdb_stdtargerr);
> +	ui_file_flush (target_fd == 1 ? gdb_stdtarg : gdb_stdtargerr);
>  	ret = length;
>  	break;
>        default:
> diff --git a/gdb/remote-sim.c b/gdb/remote-sim.c
> index 67b4690945..f9c2f605c3 100644
> --- a/gdb/remote-sim.c
> +++ b/gdb/remote-sim.c
> @@ -363,7 +363,7 @@ gdb_os_write_stdout (host_callback *p, const char *buf, int len)
>  static void
>  gdb_os_flush_stdout (host_callback *p)
>  {
> -  gdb_flush (gdb_stdtarg);
> +  ui_file_flush (gdb_stdtarg);
>  }
>  
>  /* GDB version of os_write_stderr callback.  */
> @@ -388,7 +388,7 @@ gdb_os_write_stderr (host_callback *p, const char *buf, int len)
>  static void
>  gdb_os_flush_stderr (host_callback *p)
>  {
> -  gdb_flush (gdb_stdtargerr);
> +  ui_file_flush (gdb_stdtargerr);
>  }
>  
>  /* GDB version of printf_filtered callback.  */
> diff --git a/gdb/remote.c b/gdb/remote.c
> index 3fc9a2608e..054802f744 100644
> --- a/gdb/remote.c
> +++ b/gdb/remote.c
> @@ -6784,7 +6784,7 @@ remote_console_output (const char *msg)
>        tb[1] = 0;
>        fputs_unfiltered (tb, gdb_stdtarg);
>      }
> -  gdb_flush (gdb_stdtarg);
> +  ui_file_flush (gdb_stdtarg);
>  }
>  
>  struct stop_reply : public notif_event
> diff --git a/gdb/ui-file.c b/gdb/ui-file.c
> index 71b74bba19..d7d113856e 100644
> --- a/gdb/ui-file.c
> +++ b/gdb/ui-file.c
> @@ -91,7 +91,7 @@ null_file::write_async_safe (const char *buf, long sizeof_buf)
>  
>  
>  void
> -gdb_flush (struct ui_file *file)
> +ui_file_flush (struct ui_file *file)
>  {
>    file->flush ();
>  }
> @@ -315,7 +315,7 @@ stdio_file::can_emit_style_escape ()
>  void
>  stderr_file::write (const char *buf, long length_buf)
>  {
> -  gdb_flush (gdb_stdout);
> +  ui_file_flush (gdb_stdout);
>    stdio_file::write (buf, length_buf);
>  }
>  
> @@ -325,7 +325,7 @@ stderr_file::write (const char *buf, long length_buf)
>  void
>  stderr_file::puts (const char *linebuffer)
>  {
> -  gdb_flush (gdb_stdout);
> +  ui_file_flush (gdb_stdout);
>    stdio_file::puts (linebuffer);
>  }
>  
> diff --git a/gdb/ui-file.h b/gdb/ui-file.h
> index 3f6f38a68f..711a888a2e 100644
> --- a/gdb/ui-file.h
> +++ b/gdb/ui-file.h
> @@ -100,7 +100,7 @@ public:
>  /* A preallocated null_file stream.  */
>  extern null_file null_stream;
>  
> -extern void gdb_flush (ui_file *);
> +extern void ui_file_flush (ui_file *);
>  
>  extern int ui_file_isatty (struct ui_file *);
>  
> diff --git a/gdb/utils.c b/gdb/utils.c
> index f7fae35729..5d6f680bce 100644
> --- a/gdb/utils.c
> +++ b/gdb/utils.c
> @@ -1544,6 +1544,13 @@ flush_wrap_buffer (struct ui_file *stream)
>      }
>  }
>  
> +void
> +gdb_flush (struct ui_file *stream)
> +{
> +  flush_wrap_buffer (stream);
> +  ui_file_flush (stream);
> +}

All new functions need to have a header comment.  The GDB style would
be to add this comment here:

  /* See utils.h.  */

And then add a full description of the function in the header file.

> +
>  /* Indicate that if the next sequence of characters overflows the line,
>     a newline should be inserted here rather than when it hits the end.
>     If INDENT is non-null, it is a string to be printed to indent the
> diff --git a/gdb/utils.h b/gdb/utils.h
> index 79c8a6fc8d..185f156a0a 100644
> --- a/gdb/utils.h
> +++ b/gdb/utils.h
> @@ -323,6 +323,8 @@ extern struct ui_file **current_ui_gdb_stdin_ptr (void);
>  extern struct ui_file **current_ui_gdb_stderr_ptr (void);
>  extern struct ui_file **current_ui_gdb_stdlog_ptr (void);
>  
> +extern void gdb_flush (struct ui_file *);

So you should add a comment here, ideally this would make it clear
when I would call this instead of ui_file_flush.

Ideally, given you've gone to the effort of understanding all this, it
would be nice to have a commented added on ui_file_flush too as that
was missing comments before.

Thanks,
Andrew

> +
>  /* The current top level's ui_file streams.  */
>  
>  /* Normal results */