Recent breakage -- disassemble

Recent breakage -- disassemble

[Michael Snyder]

Some time between February 18 and March 24, someone checked in a change
which broke pagination in the disassemble command.  Although "list" will
still print out a screen's worth of lines and then pause, "disassemble"
will not.

Sorry I can't pin it down closer than that -- my cvs client is giving me grief.

Michael

Re: Recent breakage -- disassemble

[Elena Zannoni]

Michael Snyder writes:
 > Some time between February 18 and March 24, someone checked in a change
 > which broke pagination in the disassemble command.  Although "list" will
 > still print out a screen's worth of lines and then pause, "disassemble"
 > will not.
 > 
 > Sorry I can't pin it down closer than that -- my cvs client is giving me grief.
 > 
 > Michael

Hmmm, this could be due to the new disassembler. The old one was using
printf_filtered while the new one uses ui_out functions. But the
change happened in November 2002. Do a cvs log of printcmd.c. If you
back out those changes, does the problem go away?

elena

Re: Recent breakage -- disassemble

[Elena Zannoni]

Elena Zannoni writes:
 > Michael Snyder writes:
 >  > Some time between February 18 and March 24, someone checked in a change
 >  > which broke pagination in the disassemble command.  Although "list" will
 >  > still print out a screen's worth of lines and then pause, "disassemble"
 >  > will not.
 >  > 
 >  > Sorry I can't pin it down closer than that -- my cvs client is giving me grief.
 >  > 
 >  > Michael
 > 
 > Hmmm, this could be due to the new disassembler. The old one was using
 > printf_filtered while the new one uses ui_out functions. But the
 > change happened in November 2002. Do a cvs log of printcmd.c. If you
 > back out those changes, does the problem go away?
 > 
 > elena

Ok, the problem is not that filtered prints were not used, but that
there is code in fputs_maybe_filtered to bypass the filtered print, if
this condition is true:

  /* Don't do any filtering if it is disabled.  */
  if ((stream != gdb_stdout) || !pagination_enabled
      || (lines_per_page == UINT_MAX && chars_per_line == UINT_MAX))
    {
      fputs_unfiltered (linebuffer, stream);
      return;
    }

In this case, stream is different from gdb_stdout.

Can you try this patch?

elena

Index: utils.c
===================================================================
RCS file: /cvs/uberbaum/gdb/utils.c,v
retrieving revision 1.99
diff -u -p -r1.99 utils.c
--- utils.c	2 Apr 2003 03:02:46 -0000	1.99
+++ utils.c	22 Apr 2003 14:28:00 -0000
@@ -1887,7 +1887,7 @@ fputs_maybe_filtered (const char *linebu
     return;
 
   /* Don't do any filtering if it is disabled.  */
-  if ((stream != gdb_stdout) || !pagination_enabled
+  if (!ui_file_isatty (stream) || !pagination_enabled
       || (lines_per_page == UINT_MAX && chars_per_line == UINT_MAX))
     {
       fputs_unfiltered (linebuffer, stream);

Re: Recent breakage -- disassemble

[Michael Snyder]

Elena Zannoni wrote:
> 
> Elena Zannoni writes:
>  > Michael Snyder writes:
>  >  > Some time between February 18 and March 24, someone checked in a change
>  >  > which broke pagination in the disassemble command.  Although "list" will
>  >  > still print out a screen's worth of lines and then pause, "disassemble"
>  >  > will not.
>  >  >
>  >  > Sorry I can't pin it down closer than that -- my cvs client is giving me grief.
>  >  >
>  >  > Michael
>  >
>  > Hmmm, this could be due to the new disassembler. The old one was using
>  > printf_filtered while the new one uses ui_out functions. But the
>  > change happened in November 2002. Do a cvs log of printcmd.c. If you
>  > back out those changes, does the problem go away?
>  >
>  > elena
> 
> Ok, the problem is not that filtered prints were not used, but that
> there is code in fputs_maybe_filtered to bypass the filtered print, if
> this condition is true:
> 
>   /* Don't do any filtering if it is disabled.  */
>   if ((stream != gdb_stdout) || !pagination_enabled
>       || (lines_per_page == UINT_MAX && chars_per_line == UINT_MAX))
>     {
>       fputs_unfiltered (linebuffer, stream);
>       return;
>     }
> 
> In this case, stream is different from gdb_stdout.
> 
> Can you try this patch?

Thanks, Elena, your patch makes it work again.
Michael



> 
> Index: utils.c
> ===================================================================
> RCS file: /cvs/uberbaum/gdb/utils.c,v
> retrieving revision 1.99
> diff -u -p -r1.99 utils.c
> --- utils.c     2 Apr 2003 03:02:46 -0000       1.99
> +++ utils.c     22 Apr 2003 14:28:00 -0000
> @@ -1887,7 +1887,7 @@ fputs_maybe_filtered (const char *linebu
>      return;
> 
>    /* Don't do any filtering if it is disabled.  */
> -  if ((stream != gdb_stdout) || !pagination_enabled
> +  if (!ui_file_isatty (stream) || !pagination_enabled
>        || (lines_per_page == UINT_MAX && chars_per_line == UINT_MAX))
>      {
>        fputs_unfiltered (linebuffer, stream);

Re: Recent breakage -- disassemble

[Andrew Cagney]

> Ok, the problem is not that filtered prints were not used, but that
> there is code in fputs_maybe_filtered to bypass the filtered print, if
> this condition is true:
> 
>   /* Don't do any filtering if it is disabled.  */
>   if ((stream != gdb_stdout) || !pagination_enabled
>       || (lines_per_page == UINT_MAX && chars_per_line == UINT_MAX))
>     {
>       fputs_unfiltered (linebuffer, stream);
>       return;
>     }
> 
> In this case, stream is different from gdb_stdout.
> 
> Can you try this patch?

I think this patch has the potential for causing gdb_stdlog and 
gdb_stderr to also paginate.  ui_file_isatty(gdb_stderr) also returns 1.

Andrew

> Index: utils.c
> ===================================================================
> RCS file: /cvs/uberbaum/gdb/utils.c,v
> retrieving revision 1.99
> diff -u -p -r1.99 utils.c
> --- utils.c	2 Apr 2003 03:02:46 -0000	1.99
> +++ utils.c	22 Apr 2003 14:28:00 -0000
> @@ -1887,7 +1887,7 @@ fputs_maybe_filtered (const char *linebu
>      return;
>  
>    /* Don't do any filtering if it is disabled.  */
> -  if ((stream != gdb_stdout) || !pagination_enabled
> +  if (!ui_file_isatty (stream) || !pagination_enabled
>        || (lines_per_page == UINT_MAX && chars_per_line == UINT_MAX))
>      {
>        fputs_unfiltered (linebuffer, stream);
> 
> 
> 
> 
>