Re: Patch to gdb.texinfo

[Eli Zaretskii <eliz@delorie.com>]

>    > -For complicated cases, you can specify an arbitrary number of @r{-T}@var{section} @var{address}                                               
>    > +For complicated cases, you can specify an arbitrary number of @samp{@r{-T}@var{section} @var{address}}                                        

>    I don't see the rationale for this change.  Could you please explain?
>
> Look at the printable copy.  Single quotes around `-Tsection address'
> looks better.

You don't need a @samp to add the quotes.

What I wanted to point out was that a @samp whose effect is
effectively undone by @r and @var is not necessary.

I'll let Stan decide on this one.
From dima@Chg.RU Mon Mar 13 01:44:00 2000
From: Dmitry Sivachenko <dima@Chg.RU>
To: eliz@is.elta.co.il
Cc: gdb-patches@sourceware.cygnus.com
Subject: Re: Patch to gdb.texinfo
Date: Mon, 13 Mar 2000 01:44:00 -0000
Message-id: <200003130944.MAA64255@netserv1.chg.ru>
References: <200003130906.MAA63437@netserv1.chg.ru> <200003121611.TAA45044@netserv1.chg.ru> <200003130859.DAA25938@indy.delorie.com> <200003130933.EAA25971@indy.delorie.com>
X-SW-Source: 2000-03/msg00224.html
Content-length: 771

   >    > -For complicated cases, you can specify an arbitrary number of @r{-T}@var{section} @var{address}                                               
   >    > +For complicated cases, you can specify an arbitrary number of @samp{@r{-T}@var{section} @var{address}}                                        

   >    I don't see the rationale for this change.  Could you please explain?
   >
   > Look at the printable copy.  Single quotes around `-Tsection address'
   > looks better.

   You don't need a @samp to add the quotes.

   What I wanted to point out was that a @samp whose effect is
   effectively undone by @r and @var is not necessary.

   I'll let Stan decide on this one.

OK, may be it is better to replace @samp with explicit quotes (`.....').

--dima
From Peter.Schauer@regent.e-technik.tu-muenchen.de Mon Mar 13 01:48:00 2000
From: "Peter.Schauer" <Peter.Schauer@regent.e-technik.tu-muenchen.de>
To: gdb-patches@sourceware.cygnus.com
Subject: RFD: infrun.c:  No bpstat_stop_status call after proceed over break ?
Date: Mon, 13 Mar 2000 01:48:00 -0000
Message-id: <200003130948.KAA07547@reisser.regent.e-technik.tu-muenchen.de>
X-SW-Source: 2000-03/msg00226.html
Content-length: 4730

I am currently trying to fix a GDB bug with missing watchpoint triggers
after proceeding over a breakpoint on x86 targets.

Here is an example, using gdb.c++/annota2:

(gdb) b main
Breakpoint 1 at 0x8048b3a: file annota2.cc, line 21.
(gdb) r
Starting program: annota2

Breakpoint 1, main () at annota2.cc:21
21        a.x = 0;
(gdb) watch a.x
Watchpoint 2: a.x
(gdb) c
Continuing.
Watchpoint 2: a.x

Old value = -536882292
New value = 1
main () at annota2.cc:23
23        a.y = 2;
(gdb)


The breakpoint at main is at the instruction which should cause the
watchpoint trigger (a.x = 0).
When continuing over the breakpoint, breakpoints are removed and the target
is single stepped with trap_expected set to one.
After the step GDB does not reexamine the stop reason, missing the watchpoint
trigger at a.x = 0 and stops too late at the second watchpoint trigger.

Here is the relevant code from handle_inferior_event:

        /* Don't even think about breakpoints
           if just proceeded over a breakpoint.

           However, if we are trying to proceed over a breakpoint
           and end up in sigtramp, then through_sigtramp_breakpoint
           will be set and we should check whether we've hit the
           step breakpoint.  */
        if (stop_signal == TARGET_SIGNAL_TRAP && trap_expected
            && through_sigtramp_breakpoint == NULL)
          bpstat_clear (&stop_bpstat);
        else
          {
            /* See if there is a breakpoint at the current PC.  */
            stop_bpstat = bpstat_stop_status
              (&stop_pc,
		.
		.

I currently have no idea why we need to special case the proceed over
a breakpoint here, perhaps it is some historic remnant (although I do have
this nagging feeling that I might miss something obvious).

The patch below would get rid of the special case and fix the problem, it
causes no testsuite regressions. Note that bpstat_clear is called and
stop_print_frame is set to 1 unconditionally already a few lines above the
bpstat_stop_status call, so there is no need to do it again.


Any suggestions why we should need the old special case ?


2000-03-12  Peter Schauer  <pes@regent.e-technik.tu-muenchen.de>

	* infrun.c (handle_inferior_event):  Remove special case for stop
	after proceeding over a breakpoint, it caused missed watchpoints.

*** gdb/infrun.c.orig	Thu Feb 24 13:41:46 2000
--- gdb/infrun.c	Mon Mar 13 10:11:50 2000
***************
*** 2076,2110 ****
  	    return;
  	  }
  
! 	/* Don't even think about breakpoints
! 	   if just proceeded over a breakpoint.
! 
! 	   However, if we are trying to proceed over a breakpoint
! 	   and end up in sigtramp, then through_sigtramp_breakpoint
! 	   will be set and we should check whether we've hit the
! 	   step breakpoint.  */
! 	if (stop_signal == TARGET_SIGNAL_TRAP && trap_expected
! 	    && through_sigtramp_breakpoint == NULL)
! 	  bpstat_clear (&stop_bpstat);
! 	else
! 	  {
! 	    /* See if there is a breakpoint at the current PC.  */
! 	    stop_bpstat = bpstat_stop_status
! 	      (&stop_pc,
! 	    /* Pass TRUE if our reason for stopping is something other
! 	       than hitting a breakpoint.  We do this by checking that
! 	       1) stepping is going on and 2) we didn't hit a breakpoint
! 	       in a signal handler without an intervening stop in
! 	       sigtramp, which is detected by a new stack pointer value
! 	       below any usual function calling stack adjustments.  */
! 		(currently_stepping (ecs)
! 		 && !(step_range_end
! 		      && INNER_THAN (read_sp (), (step_sp - 16))))
! 	      );
! 	    /* Following in case break condition called a
! 	       function.  */
! 	    stop_print_frame = 1;
! 	  }
  
  	if (stop_signal == TARGET_SIGNAL_TRAP)
  	  ecs->random_signal
--- 2092,2114 ----
  	    return;
  	  }
  
! 	/* See if there is a breakpoint at the current PC.
! 	   Older versions of GDB did not call bpstat_stop_status after
! 	   proceeding over a breakpoint, causing missed watchpoints when
! 	   proceeding over a breakpoint on an instruction which triggers
! 	   a watchpoint.  */
! 	stop_bpstat = bpstat_stop_status
! 	  (&stop_pc,
! 	/* Pass TRUE if our reason for stopping is something other
! 	   than hitting a breakpoint.  We do this by checking that
! 	   1) stepping is going on and 2) we didn't hit a breakpoint
! 	   in a signal handler without an intervening stop in
! 	   sigtramp, which is detected by a new stack pointer value
! 	   below any usual function calling stack adjustments.  */
! 	    (currently_stepping (ecs)
! 	     && !(step_range_end
! 		  && INNER_THAN (read_sp (), (step_sp - 16))))
! 	  );
  
  	if (stop_signal == TARGET_SIGNAL_TRAP)
  	  ecs->random_signal

-- 
Peter Schauer			pes@regent.e-technik.tu-muenchen.de
From Peter.Schauer@regent.e-technik.tu-muenchen.de Mon Mar 13 01:48:00 2000
From: "Peter.Schauer" <Peter.Schauer@regent.e-technik.tu-muenchen.de>
To: gdb-patches@sourceware.cygnus.com
Subject: RFA: infrun.c, breakpoint.c: Kludge for Solaris x86 hardware watchpoint support
Date: Mon, 13 Mar 2000 01:48:00 -0000
Message-id: <200003130947.KAA07528@reisser.regent.e-technik.tu-muenchen.de>
X-SW-Source: 2000-03/msg00225.html
Content-length: 2985

Unfortunately I'd need the following kludge to work around a Solaris x86
kernel problem with hardware watchpoint support.
See the comment in the patches for a description of the problem.

Any objections ?

2000-03-12  Peter Schauer  <pes@regent.e-technik.tu-muenchen.de>

	breakpoint.c, breakpoint.h (remove_hw_watchpoints):  New function.
	infrun.c (resume):  Remove hardware watchpoints before stepping
	when CANNOT_STEP_HW_WATCHPOINTS is nonzero.

*** gdb/breakpoint.c.orig	Thu Feb 24 13:41:33 2000
--- gdb/breakpoint.c	Sat Mar  4 11:42:48 2000
***************
*** 1080,1085 ****
--- 1081,1107 ----
  }
  
  int
+ remove_hw_watchpoints ()
+ {
+   register struct breakpoint *b;
+   int val;
+ 
+   ALL_BREAKPOINTS (b)
+   {
+     if (b->inserted
+ 	&& (b->type == bp_hardware_watchpoint
+ 	    || b->type == bp_read_watchpoint
+ 	    || b->type == bp_access_watchpoint))
+       {
+ 	val = remove_breakpoint (b, mark_uninserted);
+ 	if (val != 0)
+ 	  return val;
+       }
+   }
+   return 0;
+ }
+ 
+ int
  reattach_breakpoints (pid)
       int pid;
  {
*** ./gdb/breakpoint.h.orig	Thu Feb  3 05:14:27 2000
--- ./gdb/breakpoint.h	Sat Mar  4 11:42:48 2000
***************
*** 562,567 ****
--- 562,568 ----
  extern int insert_breakpoints PARAMS ((void));
  
  extern int remove_breakpoints PARAMS ((void));
+ extern int remove_hw_watchpoints PARAMS ((void));
  
  /* This function can be used to physically insert eventpoints from the
     specified traced inferior process, without modifying the breakpoint
*** gdb/infrun.c.orig	Thu Feb 24 13:41:46 2000
--- gdb/infrun.c	Mon Mar 13 09:35:54 2000
***************
*** 296,301 ****
--- 296,308 ----
  #define HAVE_CONTINUABLE_WATCHPOINT 1
  #endif
  
+ #ifndef CANNOT_STEP_HW_WATCHPOINTS
+ #define CANNOT_STEP_HW_WATCHPOINTS 0
+ #else
+ #undef  CANNOT_STEP_HW_WATCHPOINTS
+ #define CANNOT_STEP_HW_WATCHPOINTS 1
+ #endif
+ 
  /* Tables of how to react to signals; the user sets them.  */
  
  static unsigned char *signal_stop;
***************
*** 796,801 ****
--- 803,820 ----
    if (step && breakpoints_inserted && breakpoint_here_p (read_pc ()))
      step = 0;
  #endif
+ 
+   /* Some targets (e.g. Solaris x86) have a kernel bug when stepping
+      over an instruction that causes a page fault without triggering
+      a hardware watchpoint. The kernel properly notices that it shouldn't
+      stop, because the hardware watchpoint is not triggered, but it forgets
+      the step request and continues the program normally.
+      Work around the problem by removing hardware watchpoints if a step is
+      requested, GDB will check for a hardware watchpoint trigger after the
+      step anyway.  */
+   if (CANNOT_STEP_HW_WATCHPOINTS && step && breakpoints_inserted)
+     remove_hw_watchpoints ();
+      
  
    /* Normally, by the time we reach `resume', the breakpoints are either
       removed or inserted, as appropriate.  The exception is if we're sitting

-- 
Peter Schauer			pes@regent.e-technik.tu-muenchen.de
From ac131313@cygnus.com Mon Mar 13 02:05:00 2000
From: Andrew Cagney <ac131313@cygnus.com>
To: GDB Patches <gdb-patches@sourceware.cygnus.com>
Subject: [PATCH] Fix C++ overloading, add support for seeing through references.
Date: Mon, 13 Mar 2000 02:05:00 -0000
Message-id: <38CCBD3A.35459C1C@cygnus.com>
X-SW-Source: 2000-03/msg00227.html
Content-length: 19325

FYI,

I've committed the attatched from Daniel Berlin (except for minor
edits).  It has a very interesting effect on the arm-elf -> arm-sim
testsuite results:

==> before

# of expected passes            5715
# of unexpected failures        36
# of unexpected successes       2
# of expected failures          193
# of unresolved testcases       79
# of untested testcases         18
# of unsupported tests          2

==> after

# of expected passes            5716
# of unexpected failures        36
# of unexpected successes       16
# of expected failures          178
# of unresolved testcases       79
# of untested testcases         18
# of unsupported tests          2

where the actual changes are ( ``<'' for old, ``>'' for new):

< XFAIL: gdb.c++/derivation.exp: print value of g_instance.afoo()
< XFAIL: gdb.c++/derivation.exp: print value of g_instance.bfoo()
< XFAIL: gdb.c++/derivation.exp: print value of g_instance.cfoo()
> XPASS: gdb.c++/derivation.exp: print value of g_instance.afoo()
> XPASS: gdb.c++/derivation.exp: print value of g_instance.bfoo()
> XPASS: gdb.c++/derivation.exp: print value of g_instance.cfoo()
< XFAIL: gdb.c++/overload.exp: print call overloaded func 1 arg
< XFAIL: gdb.c++/overload.exp: print call overloaded func 2 args
< XFAIL: gdb.c++/overload.exp: print call overloaded func 3 args
> XPASS: gdb.c++/overload.exp: print call overloaded func 1 arg
> XPASS: gdb.c++/overload.exp: print call overloaded func 2 args
> XPASS: gdb.c++/overload.exp: print call overloaded func 3 args
< XFAIL: gdb.c++/overload.exp: print call overloaded func void arg
> XPASS: gdb.c++/overload.exp: print call overloaded func void arg
< XFAIL: gdb.c++/overload.exp: print call overloaded func signed char
arg
< XFAIL: gdb.c++/overload.exp: print call overloaded func unsigned char
arg
> XPASS: gdb.c++/overload.exp: print call overloaded func signed char arg
> XPASS: gdb.c++/overload.exp: print call overloaded func unsigned char arg
< XFAIL: gdb.c++/overload.exp: print call overloaded func unsigned int
arg
> XPASS: gdb.c++/overload.exp: print call overloaded func unsigned int arg
< XFAIL: gdb.c++/templates.exp: print t5i.value
> PASS: gdb.c++/templates.exp: print t5i.value()
< XFAIL: gdb.c++/virtfunc.exp: print pAe->f()
< XFAIL: gdb.c++/virtfunc.exp: print pAa->f()
> XPASS: gdb.c++/virtfunc.exp: print pAe->f()
> XPASS: gdb.c++/virtfunc.exp: print pAa->f()
< XFAIL: gdb.c++/virtfunc.exp: print pDd->vg()
> XPASS: gdb.c++/virtfunc.exp: print pDd->vg()
< XFAIL: gdb.c++/virtfunc.exp: print pVB->vvb()
> XPASS: gdb.c++/virtfunc.exp: print pVB->vvb()

so fixed some long outstanding bugs.

	Andrew
Mon Mar 13 18:54:42 2000  Andrew Cagney  <cagney@b1.cygnus.com>

	From 2000-03-10 Daniel Berlin <dan@cgsoftware.com> Fix C++
 	overloading, add support for seeing through references:
	* valops.c (find_overload_match): Handle STABS overloading for
	C++.
	(find_overload_match): Look in right place for function arguments
	in the debug info.
	(find_overload_match): Rather than giving up when we have >1
	perfect match, just choose one, especially since the
	recommendation GDB gives ("disambiguate it by specifying function
	signature"), is basically impossible.
	(check_field_in): STREQ->strcmp_iw
	(search_struct_field): STREQ->strcmp_iw
	(find_method_list): STREQ->strcmp_iw
	* gdbtypes.c (rank_one_type): Add ability to see through
	references.
	(rank_one_type): strcmp->strcmp_iw, because the whitespace could
	be different.
	(rank_function): Rank function properly (was doing it wrong
	before, comparing the wrong parts of the arrays)
	(rank_one_type): Change #if 0 to #ifdef DEBUG_OLOAD.
	* gdbtypes.h: Add REFERENCE_CONVERSION_BADNESS for "badness"
	associated with converting a non-reference to a reference.

	* gdbtypes.c (rank_one_type): Add comment on how to eliminate the
 	#ifdef DEBUG_OLOAD.
	
Index: gdbtypes.c
===================================================================
RCS file: /cvs/src/src/gdb/gdbtypes.c,v
retrieving revision 1.3
diff -p -r1.3 gdbtypes.c
*** gdbtypes.c	2000/03/13 07:30:00	1.3
--- gdbtypes.c	2000/03/13 08:47:00
*************** lookup_struct_elt_type (type, name, noer
*** 994,1000 ****
      {
        char *t_field_name = TYPE_FIELD_NAME (type, i);
  
!       if (t_field_name && STREQ (t_field_name, name))
  	{
  	  return TYPE_FIELD_TYPE (type, i);
  	}
--- 994,1000 ----
      {
        char *t_field_name = TYPE_FIELD_NAME (type, i);
  
!       if (t_field_name && (strcmp_iw (t_field_name, name) == 0))
  	{
  	  return TYPE_FIELD_TYPE (type, i);
  	}
*************** rank_function (parms, nparms, args, narg
*** 2136,2143 ****
    LENGTH_MATCH (bv) = (nargs != nparms) ? LENGTH_MISMATCH_BADNESS : 0;
  
    /* Now rank all the parameters of the candidate function */
!   for (i = 1; i <= min_len; i++)
!     bv->rank[i] = rank_one_type (parms[i - 1], args[i - 1]);
  
    /* If more arguments than parameters, add dummy entries */
    for (i = min_len + 1; i <= nargs; i++)
--- 2136,2143 ----
    LENGTH_MATCH (bv) = (nargs != nparms) ? LENGTH_MISMATCH_BADNESS : 0;
  
    /* Now rank all the parameters of the candidate function */
!   for (i = 1; i < min_len; i++)
!     bv->rank[i] = rank_one_type (parms[i], args[i]);
  
    /* If more arguments than parameters, add dummy entries */
    for (i = min_len + 1; i <= nargs; i++)
*************** rank_one_type (parm, arg)
*** 2178,2185 ****
    if (parm == arg)
      return 0;
  
! #if 0
!   /* Debugging only */
    printf ("------ Arg is %s [%d], parm is %s [%d]\n",
        TYPE_NAME (arg), TYPE_CODE (arg), TYPE_NAME (parm), TYPE_CODE (parm));
  #endif
--- 2178,2198 ----
    if (parm == arg)
      return 0;
  
!   /* See through references, since we can almost make non-references
!      references. */
!   if (TYPE_CODE (arg) == TYPE_CODE_REF)
!     return (rank_one_type (TYPE_TARGET_TYPE (arg), parm)
! 	    + REFERENCE_CONVERSION_BADNESS);
!   if (TYPE_CODE (parm) == TYPE_CODE_REF)
!     return (rank_one_type (arg, TYPE_TARGET_TYPE (parm))
! 	    + REFERENCE_CONVERSION_BADNESS);
! 
! #ifdef DEBUG_OLOAD
!   /* Debugging only. */
!   /* FIXME/FYI: cagney/2000-03-13: No need to #ifdef this sort of
!      thing.  Instead add a command like ``set debug gdbtypes <int>''.
!      (A predicate to this is the addition of the ``set debug''). Also,
!      send the output to gdb_stderr and don't use printf. */
    printf ("------ Arg is %s [%d], parm is %s [%d]\n",
        TYPE_NAME (arg), TYPE_CODE (arg), TYPE_NAME (parm), TYPE_CODE (parm));
  #endif
*************** rank_one_type (parm, arg)
*** 2246,2261 ****
  		{
  		  if (TYPE_UNSIGNED (arg))
  		    {
! 		      if (!strcmp (TYPE_NAME (parm), TYPE_NAME (arg)))
  			return 0;	/* unsigned int -> unsigned int, or unsigned long -> unsigned long */
! 		      else if (!strcmp (TYPE_NAME (arg), "int") && !strcmp (TYPE_NAME (parm), "long"))
  			return INTEGER_PROMOTION_BADNESS;	/* unsigned int -> unsigned long */
  		      else
  			return INTEGER_COERCION_BADNESS;	/* unsigned long -> unsigned int */
  		    }
  		  else
  		    {
! 		      if (!strcmp (TYPE_NAME (arg), "long") && !strcmp (TYPE_NAME (parm), "int"))
  			return INTEGER_COERCION_BADNESS;	/* signed long -> unsigned int */
  		      else
  			return INTEGER_CONVERSION_BADNESS;	/* signed int/long -> unsigned int/long */
--- 2259,2274 ----
  		{
  		  if (TYPE_UNSIGNED (arg))
  		    {
! 		      if (!strcmp_iw (TYPE_NAME (parm), TYPE_NAME (arg)))
  			return 0;	/* unsigned int -> unsigned int, or unsigned long -> unsigned long */
! 		      else if (!strcmp_iw (TYPE_NAME (arg), "int") && !strcmp_iw (TYPE_NAME (parm), "long"))
  			return INTEGER_PROMOTION_BADNESS;	/* unsigned int -> unsigned long */
  		      else
  			return INTEGER_COERCION_BADNESS;	/* unsigned long -> unsigned int */
  		    }
  		  else
  		    {
! 		      if (!strcmp_iw (TYPE_NAME (arg), "long") && !strcmp_iw (TYPE_NAME (parm), "int"))
  			return INTEGER_COERCION_BADNESS;	/* signed long -> unsigned int */
  		      else
  			return INTEGER_CONVERSION_BADNESS;	/* signed int/long -> unsigned int/long */
*************** rank_one_type (parm, arg)
*** 2263,2271 ****
  		}
  	      else if (!TYPE_NOSIGN (arg) && !TYPE_UNSIGNED (arg))
  		{
! 		  if (!strcmp (TYPE_NAME (parm), TYPE_NAME (arg)))
  		    return 0;
! 		  else if (!strcmp (TYPE_NAME (arg), "int") && !strcmp (TYPE_NAME (parm), "long"))
  		    return INTEGER_PROMOTION_BADNESS;
  		  else
  		    return INTEGER_COERCION_BADNESS;
--- 2276,2284 ----
  		}
  	      else if (!TYPE_NOSIGN (arg) && !TYPE_UNSIGNED (arg))
  		{
! 		  if (!strcmp_iw (TYPE_NAME (parm), TYPE_NAME (arg)))
  		    return 0;
! 		  else if (!strcmp_iw (TYPE_NAME (arg), "int") && !strcmp_iw (TYPE_NAME (parm), "long"))
  		    return INTEGER_PROMOTION_BADNESS;
  		  else
  		    return INTEGER_COERCION_BADNESS;
Index: gdbtypes.h
===================================================================
RCS file: /cvs/src/src/gdb/gdbtypes.h,v
retrieving revision 1.3
diff -p -r1.3 gdbtypes.h
*** gdbtypes.h	2000/03/13 07:30:00	1.3
--- gdbtypes.h	2000/03/13 08:47:03
*************** count_virtual_fns PARAMS ((struct type *
*** 1129,1136 ****
  #define POINTER_CONVERSION_BADNESS     2
  /* Badness of conversion of pointer to void pointer */
  #define VOID_PTR_CONVERSION_BADNESS    2
! /* Badness of convering derived to base class */
  #define BASE_CONVERSION_BADNESS        2
  
  /* Non-standard conversions allowed by the debugger */
  /* Converting a pointer to an int is usually OK */
--- 1129,1138 ----
  #define POINTER_CONVERSION_BADNESS     2
  /* Badness of conversion of pointer to void pointer */
  #define VOID_PTR_CONVERSION_BADNESS    2
! /* Badness of converting derived to base class */
  #define BASE_CONVERSION_BADNESS        2
+ /* Badness of converting from non-reference to reference */
+ #define REFERENCE_CONVERSION_BADNESS   2
  
  /* Non-standard conversions allowed by the debugger */
  /* Converting a pointer to an int is usually OK */
Index: valops.c
===================================================================
RCS file: /cvs/src/src/gdb/valops.c,v
retrieving revision 1.5
diff -p -r1.5 valops.c
*** valops.c	2000/03/13 07:30:00	1.5
--- valops.c	2000/03/13 08:47:19
*************** search_struct_field (name, arg1, offset,
*** 2068,2074 ****
        {
  	char *t_field_name = TYPE_FIELD_NAME (type, i);
  
! 	if (t_field_name && STREQ (t_field_name, name))
  	  {
  	    value_ptr v;
  	    if (TYPE_FIELD_STATIC (type, i))
--- 2068,2074 ----
        {
  	char *t_field_name = TYPE_FIELD_NAME (type, i);
  
! 	if (t_field_name && (strcmp_iw (t_field_name, name) == 0))
  	  {
  	    value_ptr v;
  	    if (TYPE_FIELD_STATIC (type, i))
*************** search_struct_field (name, arg1, offset,
*** 2083,2089 ****
  	if (t_field_name
  	    && (t_field_name[0] == '\0'
  		|| (TYPE_CODE (type) == TYPE_CODE_UNION
! 		    && STREQ (t_field_name, "else"))))
  	  {
  	    struct type *field_type = TYPE_FIELD_TYPE (type, i);
  	    if (TYPE_CODE (field_type) == TYPE_CODE_UNION
--- 2083,2089 ----
  	if (t_field_name
  	    && (t_field_name[0] == '\0'
  		|| (TYPE_CODE (type) == TYPE_CODE_UNION
! 		    && (strcmp_iw (t_field_name, "else") == 0))))
  	  {
  	    struct type *field_type = TYPE_FIELD_TYPE (type, i);
  	    if (TYPE_CODE (field_type) == TYPE_CODE_UNION
*************** search_struct_field (name, arg1, offset,
*** 2128,2134 ****
           is not yet filled in.  */
        int found_baseclass = (looking_for_baseclass
  			     && TYPE_BASECLASS_NAME (type, i) != NULL
! 			     && STREQ (name, TYPE_BASECLASS_NAME (type, i)));
  
        if (BASETYPE_VIA_VIRTUAL (type, i))
  	{
--- 2128,2134 ----
           is not yet filled in.  */
        int found_baseclass = (looking_for_baseclass
  			     && TYPE_BASECLASS_NAME (type, i) != NULL
! 			     && (strcmp_iw (name, TYPE_BASECLASS_NAME (type, i)) == 0));
  
        if (BASETYPE_VIA_VIRTUAL (type, i))
  	{
*************** search_struct_method (name, arg1p, args,
*** 2314,2320 ****
  	  else if (cplus_demangle_opname (t_field_name, dem_opname, 0))
  	    t_field_name = dem_opname;
  	}
!       if (t_field_name && STREQ (t_field_name, name))
  	{
  	  int j = TYPE_FN_FIELDLIST_LENGTH (type, i) - 1;
  	  struct fn_field *f = TYPE_FN_FIELDLIST1 (type, i);
--- 2314,2320 ----
  	  else if (cplus_demangle_opname (t_field_name, dem_opname, 0))
  	    t_field_name = dem_opname;
  	}
!       if (t_field_name && (strcmp_iw (t_field_name, name) == 0))
  	{
  	  int j = TYPE_FN_FIELDLIST_LENGTH (type, i) - 1;
  	  struct fn_field *f = TYPE_FN_FIELDLIST1 (type, i);
*************** find_method_list (argp, method, offset, 
*** 2570,2576 ****
      {
        /* pai: FIXME What about operators and type conversions? */
        char *fn_field_name = TYPE_FN_FIELDLIST_NAME (type, i);
!       if (fn_field_name && STREQ (fn_field_name, method))
  	{
  	  *num_fns = TYPE_FN_FIELDLIST_LENGTH (type, i);
  	  *basetype = type;
--- 2570,2576 ----
      {
        /* pai: FIXME What about operators and type conversions? */
        char *fn_field_name = TYPE_FN_FIELDLIST_NAME (type, i);
!       if (fn_field_name && (strcmp_iw (fn_field_name, method) == 0))
  	{
  	  *num_fns = TYPE_FN_FIELDLIST_LENGTH (type, i);
  	  *basetype = type;
*************** find_overload_match (arg_types, nargs, n
*** 2740,2745 ****
--- 2740,2748 ----
    /* Get the list of overloaded methods or functions */
    if (method)
      {
+       int i;
+       int len;
+       struct type *domain;
        obj_type_name = TYPE_NAME (VALUE_TYPE (obj));
        /* Hack: evaluate_subexp_standard often passes in a pointer
           value rather than the object itself, so try again */
*************** find_overload_match (arg_types, nargs, n
*** 2756,2761 ****
--- 2759,2784 ----
  	       obj_type_name,
  	       (obj_type_name && *obj_type_name) ? "::" : "",
  	       name);
+       domain = TYPE_DOMAIN_TYPE (fns_ptr[0].type);
+       len = TYPE_NFN_FIELDS (domain);
+       /* NOTE: dan/2000-03-10: This stuff is for STABS, which won't
+          give us the info we need directly in the types. We have to
+          use the method stub conversion to get it. Be aware that this
+          is by no means perfect, and if you use STABS, please move to
+          DWARF-2, or something like it, because trying to improve
+          overloading using STABS is really a waste of time. */
+       for (i = 0; i < len; i++)
+ 	{
+ 	  int j;
+ 	  struct fn_field *f = TYPE_FN_FIELDLIST1 (domain, i);
+ 	  int len2 = TYPE_FN_FIELDLIST_LENGTH (domain, i);
+ 
+ 	  for (j = 0; j < len2; j++)
+ 	    {
+ 	      if (TYPE_FN_FIELD_STUB (f, j))
+ 		check_stub_method (domain, i, j);
+ 	    }
+ 	}
      }
    else
      {
*************** find_overload_match (arg_types, nargs, n
*** 2782,2795 ****
    /* Consider each candidate in turn */
    for (ix = 0; ix < num_fns; ix++)
      {
!       /* Number of parameters for current candidate */
!       nparms = method ? TYPE_NFIELDS (fns_ptr[ix].type)
! 	: TYPE_NFIELDS (SYMBOL_TYPE (oload_syms[ix]));
  
        /* Prepare array of parameter types */
        parm_types = (struct type **) xmalloc (nparms * (sizeof (struct type *)));
        for (jj = 0; jj < nparms; jj++)
! 	parm_types[jj] = method ? TYPE_FIELD_TYPE (fns_ptr[ix].type, jj)
  	  : TYPE_FIELD_TYPE (SYMBOL_TYPE (oload_syms[ix]), jj);
  
        /* Compare parameter types to supplied argument types */
--- 2805,2832 ----
    /* Consider each candidate in turn */
    for (ix = 0; ix < num_fns; ix++)
      {
!       if (method)
! 	{
! 	  /* For static member functions, we won't have a this pointer, but nothing
! 	     else seems to handle them right now, so we just pretend ourselves */
! 	  nparms=0;
  
+ 	  if (TYPE_FN_FIELD_ARGS(fns_ptr,ix))
+ 	    {
+ 	      while (TYPE_CODE(TYPE_FN_FIELD_ARGS(fns_ptr,ix)[nparms]) != TYPE_CODE_VOID)
+ 		nparms++;
+ 	    }
+ 	}
+       else
+ 	{
+ 	  /* If it's not a method, this is the proper place */
+ 	  nparms=TYPE_NFIELDS(SYMBOL_TYPE(oload_syms[ix]));
+ 	}
+ 
        /* Prepare array of parameter types */
        parm_types = (struct type **) xmalloc (nparms * (sizeof (struct type *)));
        for (jj = 0; jj < nparms; jj++)
! 	parm_types[jj] = method ? (TYPE_FN_FIELD_ARGS(fns_ptr,ix)[jj])
  	  : TYPE_FIELD_TYPE (SYMBOL_TYPE (oload_syms[ix]), jj);
  
        /* Compare parameter types to supplied argument types */
*************** find_overload_match (arg_types, nargs, n
*** 2826,2841 ****
  	  }
        free (parm_types);
  #ifdef DEBUG_OLOAD
        if (method)
  	printf ("Overloaded method instance %s, # of parms %d\n", fns_ptr[ix].physname, nparms);
        else
  	printf ("Overloaded function instance %s # of parms %d\n", SYMBOL_DEMANGLED_NAME (oload_syms[ix]), nparms);
!       for (jj = 0; jj <= nargs; jj++)
  	printf ("...Badness @ %d : %d\n", jj, bv->rank[jj]);
        printf ("Overload resolution champion is %d, ambiguous? %d\n", oload_champ, oload_ambiguous);
  #endif
      }				/* end loop over all candidates */
  
    if (oload_ambiguous)
      {
        if (method)
--- 2863,2885 ----
  	  }
        free (parm_types);
  #ifdef DEBUG_OLOAD
+       /* FIXME: cagney/2000-03-12: Send the output to gdb_stderr.  See
+          comments above about adding a ``set debug'' command. */
        if (method)
  	printf ("Overloaded method instance %s, # of parms %d\n", fns_ptr[ix].physname, nparms);
        else
  	printf ("Overloaded function instance %s # of parms %d\n", SYMBOL_DEMANGLED_NAME (oload_syms[ix]), nparms);
!       for (jj = 0; jj < nargs; jj++)
  	printf ("...Badness @ %d : %d\n", jj, bv->rank[jj]);
        printf ("Overload resolution champion is %d, ambiguous? %d\n", oload_champ, oload_ambiguous);
  #endif
      }				/* end loop over all candidates */
  
+   /* NOTE: dan/2000-03-10: Seems to be a better idea to just pick one
+      if they have the exact same goodness. This is because there is no
+      way to differentiate based on return type, which we need to in
+      cases like overloads of .begin() <It's both const and non-const> */
+ #if 0
    if (oload_ambiguous)
      {
        if (method)
*************** find_overload_match (arg_types, nargs, n
*** 2847,2852 ****
--- 2891,2897 ----
  	error ("Cannot resolve overloaded function %s to unique instance; disambiguate by specifying function signature",
  	       func_name);
      }
+ #endif
  
    /* Check how bad the best match is */
    for (ix = 1; ix <= nargs; ix++)
*************** check_field_in (type, name)
*** 2943,2949 ****
    for (i = TYPE_NFIELDS (type) - 1; i >= TYPE_N_BASECLASSES (type); i--)
      {
        char *t_field_name = TYPE_FIELD_NAME (type, i);
!       if (t_field_name && STREQ (t_field_name, name))
  	return 1;
      }
  
--- 2988,2994 ----
    for (i = TYPE_NFIELDS (type) - 1; i >= TYPE_N_BASECLASSES (type); i--)
      {
        char *t_field_name = TYPE_FIELD_NAME (type, i);
!       if (t_field_name && (strcmp_iw (t_field_name, name) == 0))
  	return 1;
      }
  
*************** check_field_in (type, name)
*** 2960,2966 ****
  
    for (i = TYPE_NFN_FIELDS (type) - 1; i >= 0; --i)
      {
!       if (STREQ (TYPE_FN_FIELDLIST_NAME (type, i), name))
  	return 1;
      }
  
--- 3005,3011 ----
  
    for (i = TYPE_NFN_FIELDS (type) - 1; i >= 0; --i)
      {
!       if (strcmp_iw (TYPE_FN_FIELDLIST_NAME (type, i), name) == 0)
  	return 1;
      }
  
From ac131313@cygnus.com Mon Mar 13 02:51:00 2000
From: Andrew Cagney <ac131313@cygnus.com>
To: GDB Patches <gdb-patches@sourceware.cygnus.com>
Subject: [RFC] Notes on QUIT and STREQ et.al.
Date: Mon, 13 Mar 2000 02:51:00 -0000
Message-id: <38CCC819.1071F28E@cygnus.com>
X-SW-Source: 2000-03/msg00228.html
Content-length: 1709

The attatched spells out the long term prospects of both STREQ et.al.
and QUIT.
Look OK to everyone?

	Andrew
Mon Mar 13 21:21:41 2000  Andrew Cagney  <cagney@b1.cygnus.com>

	* defs.h (STREQ): Note that the STR* macros do not need to be
 	used.
	(QUIT): Note that this can probably be replaced by a function.

Index: defs.h
===================================================================
RCS file: /cvs/src/src/gdb/defs.h,v
retrieving revision 1.11
diff -p -r1.11 defs.h
*** defs.h	2000/03/13 07:30:00	1.11
--- defs.h	2000/03/13 10:24:00
*************** extern int core_addr_greaterthan (CORE_A
*** 120,125 ****
--- 120,130 ----
  /* Gdb does *lots* of string compares.  Use macros to speed them up by
     avoiding function calls if the first characters are not the same. */
  
+ /* NOTE: cagney/2000-03-13: There is no reason for using these macros
+    in new code (which is just short of marking them as deprecated).
+    While old code can continue to refer to them, new code is better
+    off using the more familar strcmp(). */
+ 
  #define STRCMP(a,b) (*(a) == *(b) ? strcmp ((a), (b)) : (int)*(a) - (int)*(b))
  #define STREQ(a,b) (*(a) == *(b) ? !strcmp ((a), (b)) : 0)
  #define STREQN(a,b,c) (*(a) == *(b) ? !strncmp ((a), (b), (c)) : 0)
*************** extern int immediate_quit;
*** 152,157 ****
--- 157,167 ----
  extern int sevenbit_strings;
  
  extern void quit (void);
+ 
+ /* FIXME: cagney/2000-03-13: It has been suggested that the peformance
+    benefits of having a ``QUIT'' macro rather than a function are
+    marginal.  Before QUIT can be replaced by a function, a profile
+    confirming this, would be needed. */
  
  #ifdef QUIT
  /* do twice to force compiler warning */
From kettenis@wins.uva.nl Mon Mar 13 06:12:00 2000
From: Mark Kettenis <kettenis@wins.uva.nl>
To: ac131313@cygnus.com
Cc: gdb-patches@sourceware.cygnus.com
Subject: Re: [RFC] Notes on QUIT and STREQ et.al.
Date: Mon, 13 Mar 2000 06:12:00 -0000
Message-id: <200003131412.PAA16094@landau.wins.uva.nl>
References: <38CCC819.1071F28E@cygnus.com>
X-SW-Source: 2000-03/msg00229.html
Content-length: 573

   Date: Mon, 13 Mar 2000 21:51:05 +1100
   From: Andrew Cagney <ac131313@cygnus.com>

   The attatched spells out the long term prospects of both STREQ et.al.
   and QUIT.
   Look OK to everyone?

Well, STRCMP really doesn't make any sense.  A decent compiler in
combination with appropriate headers will take care of the
optimization.

I'm not sure if we want STREQ to go.  I think that `STREQ (a, b)' is
both easier to read and easier to type than `strcmp (a, b) == 0'.
There are quite a few GNU projects that use exactly the same macro in
exactly the same way.

Mark