[Bug c++/30274] New: bool bit-field: wrong increment and decremenet

[Bug c++/30274] New: bool bit-field: wrong increment and decremenet

[s__nakayama at infoseek dot jp]

testcase

#include <stdio.h>

#ifndef __cplusplus
typedef _Bool bool;
#endif

struct S
{
  bool x: 4;
};

int main()
{
  struct S x = {1};
  x.x++;                // x.x: value becomes 2.
  printf("%d\n", x.x);
//  x.x--;                // accept invalid
  return 0;
}

GCC 4.2 result
$g++ a.c -o a; ./a
2

GCC 4.1 result
$g++ a.c -o a; ./a
1


GCC version:4.2.0 20061212
The same bug exists also in C-frontend since 4.0.


-- 
           Summary: bool bit-field: wrong increment and decremenet
           Product: gcc
           Version: 4.2.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c++
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: s__nakayama at infoseek dot jp


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=30274

[Bug c++/30274] [4.2/4.3 Regression] bool bit-field: wrong increment and decremenet

[pinskia at gcc dot gnu dot org]

------- Comment #1 from pinskia at gcc dot gnu dot org  2006-12-21 18:13 -------
The problem is in:
        if (same_type_p (TREE_TYPE (arg), boolean_type_node))
Which was in 4.2 and before:
        if (TREE_TYPE (arg) == boolean_type_node)

Confirmed.  I might be able to fix this.


-- 

pinskia at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |pinskia at gcc dot gnu dot
                   |                            |org
         AssignedTo|unassigned at gcc dot gnu   |pinskia at gcc dot gnu dot
                   |dot org                     |org
             Status|UNCONFIRMED                 |ASSIGNED
     Ever Confirmed|0                           |1
           Keywords|                            |wrong-code
   Last reconfirmed|0000-00-00 00:00:00         |2006-12-21 18:13:05
               date|                            |
            Summary|bool bit-field: wrong       |[4.2/4.3 Regression] bool
                   |increment and decremenet    |bit-field: wrong increment
                   |                            |and decremenet
   Target Milestone|---                         |4.2.0


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=30274

[Bug c++/30274] [4.2/4.3 Regression] bool bit-field: wrong increment and decremenet

[bangerth at dealii dot org]

------- Comment #2 from bangerth at dealii dot org  2006-12-22 00:52 -------
When is the justification that we expect a value of 2? Bool in C++ is
a one-bit type and when you do x.x++ I would imagine that you overflow
the range of that type. The fact that you declare it as a bit-field of
length 4 is immaterial, in my opinion: you also don't get a 128-bit integer
simply by declaring a variable as
struct X {
  int x : 128;
};

W.


-- 

bangerth at dealii dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |bangerth at dealii dot org
             Status|ASSIGNED                    |WAITING


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=30274

[Bug c++/30274] [4.2/4.3 Regression] bool bit-field: wrong increment and decremenet

[s__nakayama at infoseek dot jp]

------- Comment #3 from s__nakayama at infoseek dot jp  2006-12-22 08:39 -------
(In reply to comment #2)
> When is the justification that we expect a value of 2? Bool in C++ is
> a one-bit type and when you do x.x++ I would imagine that you overflow
> the range of that type. The fact that you declare it as a bit-field of
> length 4 is immaterial, in my opinion: you also don't get a 128-bit integer
> simply by declaring a variable as
> struct X {
>   int x : 128;
> };

It is a result of g++4.2 not a expectation that a value becomes 2.
Bit-size is immaterial.
The problem is that std 5.3.2 isn't applied for bit-field.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=30274

[Bug c++/30274] [4.2/4.3 Regression] bool bit-field: wrong increment and decremenet

[pinskia at gcc dot gnu dot org]

------- Comment #4 from pinskia at gcc dot gnu dot org  2006-12-22 08:45 -------
This is a very simple patch to fix this issue also:
-        if (same_type_p (TREE_TYPE (arg), boolean_type_node))
+        if (TREE_CODE (TREE_TYPE (arg)) == BOOLEAN_TYPE)


-- 

pinskia at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|WAITING                     |ASSIGNED
   Last reconfirmed|2006-12-21 18:13:05         |2006-12-22 08:45:50
               date|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=30274

[Bug c++/30274] [4.2/4.3 Regression] bool bit-field: wrong increment and decremenet

[pinskia at gcc dot gnu dot org]

------- Comment #5 from pinskia at gcc dot gnu dot org  2006-12-22 09:24 -------
Actually I was wrong in saying this is an easy fix.
That part of the fix is still needed but there is another issue in that we are
getting the wrong type for the increment:
 <component_ref 0xb7d2a118
    type <integer_type 0xb7d5e9b4 public unsigned QI
        size <integer_cst 0xb7c911e0 constant invariant 8>
        unit size <integer_cst 0xb7c911f8 constant invariant 1>
        align 8 symtab 0 alias set -1 precision 4 min <integer_cst 0xb7d61120
0> max <integer_cst 0xb7d61138 15>>

    arg 0 <var_decl 0xb7c9e210 x
        type <record_type 0xb7d5e8fc S type_1 type_5 QI size <integer_cst
0xb7c911e0 8> unit size <integer_cst 0xb7c911f8 1>
            align 8 symtab 0 alias set -1 fields <field_decl 0xb7d5e958 x>
            X() X(constX&) this=(X&) n_parents=0 use_template=0
interface-unknown
            pointer_to_this <pointer_type 0xb7d5eac8> chain <type_decl
0xb7d27958 S>>
        used tree_1 tree_2 tree_3 decl_5 QI file t.cc line 9 size <integer_cst
0xb7c911e0 8> unit size <integer_cst 0xb7c911f8 1>
        align 8 context <function_decl 0xb7d5f0e0 main> initial <constructor
0xb7d61330>>
    arg 1 <field_decl 0xb7d5e958 x type <integer_type 0xb7d5e9b4>
        used unsigned external nonlocal bit-field nonaddressable decl_3 decl_4
QI file t.cc line 4
        size <integer_cst 0xb7d235d0 constant invariant 4> unit size
<integer_cst 0xb7c911f8 1>
        align 1 offset_align 128
        offset <integer_cst 0xb7c91180 constant invariant 0>
        bit offset <integer_cst 0xb7c91930 constant invariant 0>
        bit_field_type <boolean_type 0xb7ca54ac bool public unsigned QI size
<integer_cst 0xb7c911e0 8> unit size <integer_cst 0xb7c911f8 1>
            align 8 symtab 0 alias set -1 precision 1 min <integer_cst
0xb7c915d0 0> max <integer_cst 0xb7c91600 1>> context <record_type 0xb7d5e8fc
S>
        chain <type_decl 0xb7d279c0 S type <record_type 0xb7d5e8fc S>
            nonlocal decl_4 VOID file t.cc line 3
            align 1 context <record_type 0xb7d5e8fc S>
           >>>


-- 

pinskia at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         AssignedTo|pinskia at gcc dot gnu dot  |unassigned at gcc dot gnu
                   |org                         |dot org
             Status|ASSIGNED                    |NEW


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=30274

[Bug c++/30274] [4.2/4.3 Regression] bool bit-field: wrong increment and decremenet

[mmitchel at gcc dot gnu dot org]

------- Comment #6 from mmitchel at gcc dot gnu dot org  2007-02-19 20:57 -------
This is not a bug.  The C++ standard says:

[expr.post.incr]

the value of the object is modified by adding 1 to it, unless the object is of
type bool, in which case it is set to true


-- 

mmitchel at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |INVALID


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=30274

[Bug c++/30274] [4.2/4.3 Regression] bool bit-field: wrong increment and decremenet

[s__nakayama at infoseek dot jp]

------- Comment #7 from s__nakayama at infoseek dot jp  2007-02-21 02:25 -------
(In reply to comment #6)
> This is not a bug.  The C++ standard says:
> 
> [expr.post.incr]
> 
> the value of the object is modified by adding 1 to it, unless the object is of
> type bool, in which case it is set to true

I do not understand why this is not a bug. 
Will someone explain why this is not a bug?


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=30274

[Bug c++/30274] [4.2/4.3 Regression] bool bit-field: wrong increment and decremenet

[bangerth at dealii dot org]

------- Comment #8 from bangerth at dealii dot org  2007-02-26 19:10 -------
Mark,
I think you misunderstood: the value *should* be set to 'true' (==1), but is
set to '2'. Or do you mean that '2' is another valid representation of 'true'
and that therefore this is ok?

W.


-- 

bangerth at dealii dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |mmitchel at gcc dot gnu dot
                   |                            |org
             Status|RESOLVED                    |REOPENED
         Resolution|INVALID                     |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=30274

[Bug c++/30274] [4.2/4.3 Regression] bool bit-field: wrong increment and decremenet

[mmitchel at gcc dot gnu dot org]

------- Comment #9 from mmitchel at gcc dot gnu dot org  2007-02-26 20:19 -------
I'm sorry; I did apparently misunderstand.  My reading of the report was that x
did not become 2, and that this made the submitter unhappy.

I agree that x should not become 2.  Although "2" is a valid *representation*
of "true", converting "x" to "int" must yield "1", not "2".  So, either we need
conversion code, or we need to avoid "x" being "2" in the first place.  The
latter is clearly more efficient.

Either the C++ front end or the middle end should transform any "++" operation
on "bool" to "= true".  I'm not sure if there are languages where the
equivalent of "b = true; ++b" should yield "false".  In any case, I suspect
this is most easily done in the front end.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=30274

[Bug c++/30274] [4.2/4.3 Regression] bool bit-field: wrong increment and decremenet

[mmitchel at gcc dot gnu dot org]

-- 

mmitchel at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Priority|P3                          |P1


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=30274

[Bug c++/30274] [4.2/4.3 Regression] bool bit-field: wrong increment and decremenet

[mmitchel at gcc dot gnu dot org]

-- 

mmitchel at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         AssignedTo|unassigned at gcc dot gnu   |mark at codesourcery dot com
                   |dot org                     |
             Status|REOPENED                    |ASSIGNED


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=30274

[Bug c++/30274] [4.2/4.3 Regression] bool bit-field: wrong increment and decremenet

[mmitchel at gcc dot gnu dot org]

------- Comment #10 from mmitchel at gcc dot gnu dot org  2007-03-11 03:08 -------
Subject: Bug 30274

Author: mmitchel
Date: Sun Mar 11 03:07:59 2007
New Revision: 122813

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=122813
Log:
        PR c++/30274
        * cp-tree.h (unlowered_expr_type): New function.
        * typeck.c (is_bitfield_expr_with_lowered_type): Handle
        COMPOUND_EXPR, MODIFY_EXPR, and SAVE_EXPR.
        (unlowered_expr_type): New function.
        (build_unary_op): Disallow predecrements of bool bitfields.
        * call.c (build_conditional_expr): Use unlowered_expr_type.
        * pt.c (type_unification_real): Likewise.
        PR c++/30274
        * g++.dg/expr/bitfield3.C: New test.
        * g++.dg/expr/bitfield4.C: Likewise.
        * g++.dg/expr/bitfield5.C: Likewise.
        * g++.dg/expr/bitfield6.C: Likewise.

Added:
    trunk/gcc/testsuite/g++.dg/expr/bitfield3.C
    trunk/gcc/testsuite/g++.dg/expr/bitfield4.C
    trunk/gcc/testsuite/g++.dg/expr/bitfield5.C
    trunk/gcc/testsuite/g++.dg/expr/bitfield6.C
Modified:
    trunk/gcc/cp/ChangeLog
    trunk/gcc/cp/call.c
    trunk/gcc/cp/cp-tree.h
    trunk/gcc/cp/pt.c
    trunk/gcc/cp/typeck.c
    trunk/gcc/testsuite/ChangeLog


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=30274

[Bug c++/30274] [4.2/4.3 Regression] bool bit-field: wrong increment and decremenet

[mmitchel at gcc dot gnu dot org]

------- Comment #11 from mmitchel at gcc dot gnu dot org  2007-03-11 03:09 -------
Subject: Bug 30274

Author: mmitchel
Date: Sun Mar 11 03:09:32 2007
New Revision: 122814

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=122814
Log:
        PR c++/30274
        * cp-tree.h (unlowered_expr_type): New function.
        * typeck.c (is_bitfield_expr_with_lowered_type): Handle
        COMPOUND_EXPR, MODIFY_EXPR, and SAVE_EXPR.
        (unlowered_expr_type): New function.
        (build_unary_op): Disallow predecrements of bool bitfields.
        * call.c (build_conditional_expr): Use unlowered_expr_type.
        * pt.c (type_unification_real): Likewise.
        PR c++/30274
        * g++.dg/expr/bitfield3.C: New test.
        * g++.dg/expr/bitfield4.C: Likewise.
        * g++.dg/expr/bitfield5.C: Likewise.
        * g++.dg/expr/bitfield6.C: Likewise.

Added:
    branches/gcc-4_2-branch/gcc/testsuite/g++.dg/expr/bitfield3.C
    branches/gcc-4_2-branch/gcc/testsuite/g++.dg/expr/bitfield4.C
    branches/gcc-4_2-branch/gcc/testsuite/g++.dg/expr/bitfield5.C
    branches/gcc-4_2-branch/gcc/testsuite/g++.dg/expr/bitfield6.C
Modified:
    branches/gcc-4_2-branch/gcc/cp/ChangeLog
    branches/gcc-4_2-branch/gcc/cp/call.c
    branches/gcc-4_2-branch/gcc/cp/cp-tree.h
    branches/gcc-4_2-branch/gcc/cp/pt.c
    branches/gcc-4_2-branch/gcc/cp/typeck.c
    branches/gcc-4_2-branch/gcc/testsuite/ChangeLog


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=30274

[Bug c++/30274] [4.2/4.3 Regression] bool bit-field: wrong increment and decremenet

[mmitchel at gcc dot gnu dot org]

------- Comment #12 from mmitchel at gcc dot gnu dot org  2007-03-11 03:10 -------
Fixed in 4.2.0, 4.3.0.


-- 

mmitchel at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |RESOLVED
         Resolution|                            |FIXED


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=30274