[RFC,8.3,0/3] Some style fixes

> Date: Tue, 12 Mar 2019 17:08:33 +0000 (UTC)
> From: "Hannes Domani via gdb-patches" <gdb-patches@sourceware.org>
> 
> In copy_source_line() it checks if (column < first_col), and because of the ++column directly
> before, it basically starts with 1 instead of 0.
> 
> Attached is a patch that fixes most of 2) and 3), but I ignored the handling of escaped
> characters, because I just don't have them in any of my sources.

I can confirm that this patch fixes the problems with horizontal
scrolling, with or without TAB characters in the sources.  I think we
should push it.

I also found another regression in the TUI configuration: if you step
with, say, "next" through the program, and execution winds up on a
line that is longer than the source window-width, then the current
line gets redrawn in reverse video, and as result it overwrites the
window-frame and a portion of the next source line.  This happens
because we only update one source line, and we do that in a way that
doesn't take the window width and horizontal scrolling offset into
account.

Here's the patch that fixes this regression:

On 03/13/2019 03:44 PM, Eli Zaretskii wrote:
> I also found another regression in the TUI configuration: if you step
> with, say, "next" through the program, and execution winds up on a
> line that is longer than the source window-width, then the current
> line gets redrawn in reverse video, and as result it overwrites the
> window-frame and a portion of the next source line.  This happens
> because we only update one source line, and we do that in a way that
> doesn't take the window width and horizontal scrolling offset into
> account.
> 
> Here's the patch that fixes this regression:
> 
> --- gdb/tui/tui-winsource.c~5	2019-03-12 16:51:50.000000000 +0200
> +++ gdb/tui/tui-winsource.c	2019-03-13 08:21:02.303443600 +0200
> @@ -382,12 +387,30 @@ tui_set_is_exec_point_at (struct tui_lin
>          {
>            changed++;
>            content[i]->which_element.source.is_exec_point = new_state;
> -          tui_show_source_line (win_info, i + 1);
>          }
>        i++;
>      }
>    if (changed)
> -    tui_refresh_win (&win_info->generic);
> +    {
> +      struct gdbarch *gdbarch = win_info->detail.source_info.gdbarch;
> +      struct symtab *s = NULL;
> +
> +      if (win_info->generic.type == SRC_WIN)
> +	{
> +	  struct symtab_and_line cursal
> +	    = get_current_source_symtab_and_line ();
> +
> +	  if (cursal.symtab == NULL)
> +	    s = find_pc_line_symtab (get_frame_pc (get_selected_frame (NULL)));
> +	  else
> +	    s = cursal.symtab;
> +	}
> +      tui_update_source_window_as_is (win_info, gdbarch, s,
> +				      win_info->generic.content[0]
> +					->which_element.source.line_or_addr,
> +				      FALSE);
> +      tui_refresh_win (&win_info->generic);
> +    }

The substance of the patch LGTM.

But all this code above looks very much like the new tui_refill_source_window
function, added today by 6f11e6824e15.  I think you can replace all
the above with:

  -  tui_refresh_win (&win_info->generic);
  +  tui_refill_source_window (win_info);

Thanks,
Pedro Alves

> Cc: gdb-patches@sourceware.org
> From: Pedro Alves <palves@redhat.com>
> Date: Thu, 14 Mar 2019 20:25:25 +0000
> 
> The substance of the patch LGTM.
> 
> But all this code above looks very much like the new tui_refill_source_window
> function, added today by 6f11e6824e15.  I think you can replace all
> the above with:
> 
>   -  tui_refresh_win (&win_info->generic);
>   +  tui_refill_source_window (win_info);

You are right; I pushed the above to both branches.