Re: debug/2813: Type of variable is wrong in dwarf2 debugging info

Re: debug/2813: Type of variable is wrong in dwarf2 debugging info

[Mark Kettenis]

The following reply was made to PR debug/2813; it has been noted by GNATS.

From: Mark Kettenis <kettenis@science.uva.nl>
To: gcc-gnats@gcc.gnu.org, gcc-bugs@gcc.gnu.org
Cc:  
Subject: Re: debug/2813: Type of variable is wrong in dwarf2 debugging info
Date: Mon, 14 May 2001 23:19:33 +0200

    >Description:
    Compiling the attached test with -gdwarf-2 -dA, we find DW_AT_type
    of variable lave points at the DW_TAG_const_type die constructed
    for the argument cc, instead of pointing directly to the
    DW_TAG_base_type that represents type char.
 
    This is a regression from GCC 2.95.2.
 
 This patch is the culprit:
 
 2001-04-11  Mark Mitchell  <mark@codesourcery.com>
 
 	* dwarf2out.c (modified_type_die): Don't create new types here.
 	* tree.h (get_qualified_type): New function.
 	(build_qualified_type): Adjust comment.
 	* tree.c (get_qualified_type): New function.
 	(build_qualified_type): Use it.
 
 Basically, what's happening is that when creating a DIE for a
 qualified type, the code attaches the newly created DIE to the
 origional unqualified type:
 
   equate_type_number_to_die (type, mod_type_die);
 
 Thereafter the DIE for the qualified type will be used whenever an
 unqualified type is seen.
 
 I'm not sure how to solve this elegantly.  Adding
 
   if (qualified_type)
     type = qualified_type
 
 at the end of the
 
   if (code != ERROR_MARK)
     {
 
 block should do the trick, but perhaps a more experienced GCC hacker
 can come up with a better idea.
 
 Hope this helps,
 
 Mark
 

Re: debug/2813: Type of variable is wrong in dwarf2 debugging info

[Alexandre Oliva]

The following reply was made to PR debug/2813; it has been noted by GNATS.

From: Alexandre Oliva <aoliva@redhat.com>
To: Mark Mitchell <mark@codesourcery.com>
Cc: kettenis@wins.uva.nl, gcc-gnats@gcc.gnu.org, gcc-bugs@gcc.gnu.org,
        gcc-patches@gcc.gnu.org
Subject: Re: debug/2813: Type of variable is wrong in dwarf2 debugging info
Date: 17 May 2001 19:43:58 -0300

 On May 16, 2001, Mark Mitchell <mark@codesourcery.com> wrote:
 
 > Needs a comment.  Other than that, it's fine.
 
 Is /* We want to equate the qualified type to the die below.  */ enough?
 That's what I've just checked in.
 
 -- 
 Alexandre Oliva   Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
 Red Hat GCC Developer                  aoliva@{cygnus.com, redhat.com}
 CS PhD student at IC-Unicamp        oliva@{lsd.ic.unicamp.br, gnu.org}
 Free Software Evangelist    *Please* write to mailing lists, not to me

Re: debug/2813: Type of variable is wrong in dwarf2 debugging info

[Alexandre Oliva]

The following reply was made to PR debug/2813; it has been noted by GNATS.

From: Alexandre Oliva <aoliva@redhat.com>
To: Mark Kettenis <kettenis@wins.uva.nl>
Cc: gcc-gnats@gcc.gnu.org, gcc-bugs@gcc.gnu.org, mark@codesourcery.com
Subject: Re: debug/2813: Type of variable is wrong in dwarf2 debugging info
Date: 16 May 2001 03:49:42 -0300

 --=-=-=
 
 Mark Mitchell agreed this seemed like the right fix go, so I went
 ahead and prepared this patch (that I took the liberty to attribute to
 you; I hope you don't mind :-), that I'm bootstrapping as I write
 this.  It does indeed fix this particular bug report.
 
 Ok to install in branch and mainline if bootstrap succeeds?
 
 
 --=-=-=
 Content-Type: text/x-patch
 Content-Disposition: inline; filename=dwarf2out-qualtype.patch
 
 Index: gcc/ChangeLog
 from  Mark Kettenis  <kettenis@wins.uva.nl>
 
 	* dwarf2out.c (modified_type_die): Equate qualified type to die.
 
 Index: gcc/dwarf2out.c
 ===================================================================
 RCS file: /cvs/gcc/egcs/gcc/dwarf2out.c,v
 retrieving revision 1.266
 diff -u -p -r1.266 dwarf2out.c
 --- gcc/dwarf2out.c 2001/05/12 06:03:19 1.266
 +++ gcc/dwarf2out.c 2001/05/16 06:41:26
 @@ -6922,6 +6922,9 @@ modified_type_die (type, is_const_type, 
  	  if (mod_type_die == NULL)
  	    abort ();
  	}
 +
 +      if (qualified_type)
 +	type = qualified_type;
      }
  
    equate_type_number_to_die (type, mod_type_die);
 
 --=-=-=
 
 
 -- 
 Alexandre Oliva   Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
 Red Hat GCC Developer                  aoliva@{cygnus.com, redhat.com}
 CS PhD student at IC-Unicamp        oliva@{lsd.ic.unicamp.br, gnu.org}
 Free Software Evangelist    *Please* write to mailing lists, not to me
 
 --=-=-=--

debug/2813: Type of variable is wrong in dwarf2 debugging info

[aoliva]

>Number:         2813
>Category:       debug
>Synopsis:       Type of variable is wrong in dwarf2 debugging info
>Confidential:   no
>Severity:       serious
>Priority:       high
>Responsible:    unassigned
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Sun May 13 23:26:01 PDT 2001
>Closed-Date:
>Last-Modified:
>Originator:     aoliva@redhat.com
>Release:        GCC 3.0 20010507 and 3.1 20010513
>Organization:
>Environment:
alpha-unknown-linux-gnu and i686-pc-linux-gnu, native and x-mn10300-elf
>Description:
Compiling the attached test with -gdwarf-2 -dA, we find DW_AT_type of variable lave points at the DW_TAG_const_type die constructed for the argument cc, instead of pointing directly to the DW_TAG_base_type that represents type char.

This is a regression from GCC 2.95.2.
>How-To-Repeat:

>Fix:

>Release-Note:
>Audit-Trail:
>Unformatted:
----gnatsweb-attachment----
Content-Type: application/octet-stream; name="t.c"
Content-Transfer-Encoding: base64
Content-Disposition: attachment; filename="t.c"

aW50CnF1eDEgKGNvbnN0IGNoYXIgY2MpCnsKICBjaGFyIGxhdmUgPSAnQic7Cn0K