[Bug ada/21242] New: ACATS wrong array copy code c41203a c41203b c52102b c52102d c62002a

[Bug ada/21242] New: ACATS wrong array copy code c41203a c41203b c52102b c52102d c62002a

[laurent at guerby dot net]

These started failing immediately after 4.1.0 was able to bootstrap Ada
on x86_64-linux, that is to say between

LAST_UPDATED: Sun Apr  3 16:59:59 UTC 2005
LAST_UPDATED: Fri Apr 15 18:15:00 UTC 2005

Part of array copying generated code must be wrong in some way to get those test
failing.

,.,. C41203A ACATS 2.5 05-04-27 01:33:50
---- C41203A CHECK THAT THE NAME PART OF A SLICE MAY BE OF CERTAIN FORMS
                AND THAT THE APPROPRIATE SLICE IS ACCESSED (FOR STATIC
                INDICES).
   * C41203A WRONG TARGET FOR (IN) OUT PARAMETER - N1.
   * C41203A WRONG TARGET FOR (IN) OUT PARAMETER - N2.
   * C41203A WRONG TARGET FOR (IN) OUT PARAMETER - F2.
   * C41203A WRONG TARGET FOR (IN) OUT PARAMETER - N3.
   * C41203A WRONG TARGET FOR (IN) OUT PARAMETER - N4.
   * C41203A WRONG TARGET FOR (IN) OUT PARAMETER - C41203A.N1.
**** C41203A FAILED ****************************.

,.,. C41203B ACATS 2.5 05-04-27 01:33:53
---- C41203B CHECK THAT THE NAME PART OF A SLICE MAY BE OF CERTAIN FORMS
                AND THAT THE APPROPRIATE SLICE IS ACCESSED (FOR DYNAMIC 
                INDICES).
   * C41203B WRONG TARGET FOR (IN) OUT PARAMETER - N1.
   * C41203B WRONG TARGET FOR (IN) OUT PARAMETER - N2.
   * C41203B WRONG TARGET FOR (IN) OUT PARAMETER - F2.
   * C41203B WRONG TARGET FOR (IN) OUT PARAMETER - N3.
   * C41203B WRONG TARGET FOR (IN) OUT PARAMETER - N4.
   * C41203B WRONG TARGET FOR (IN) OUT PARAMETER - C41203B.N1.
**** C41203B FAILED ****************************.

,.,. C52102B ACATS 2.5 05-04-27 01:36:33
---- C52102B CHECK THAT THE ASSIGNMENT OF OVERLAPPING SOURCE AND TARGET 
                VARIABLES (INCLUDING ARRAYS AND SLICES IN VARIOUS
                COMBINATIONS) SATISFIES THE SEMANTICS OF "COPY"
                ASSIGNMENT (PART 2: DYNAMIC BOUNDS).
   * C52102B WRONG VALUES  -  I1.
   * C52102B WRONG VALUES  -  I2.
   * C52102B WRONG VALUES  -  I3.
   * C52102B WRONG VALUES  -  I4.
   * C52102B WRONG VALUES  -  B3.
   * C52102B WRONG VALUES  -  B4.
**** C52102B FAILED ****************************.

,.,. C52102D ACATS 2.5 05-04-27 01:36:37
---- C52102D CHECK THAT THE ASSIGNMENT OF OVERLAPPING SOURCE AND TARGET 
                VARIABLES (INCLUDING ARRAYS AND SLICES IN VARIOUS
                COMBINATIONS) SATISFIES THE SEMANTICS OF "COPY"
                ASSIGNMENT WHEN INITIAL ASSIGNMENT VALUES ARE DYNAMIC 
                (PART 2: DYNAMIC BOUNDS).
   * C52102D WRONG VALUES  -  I1.
   * C52102D WRONG VALUES  -  I2.
   * C52102D WRONG VALUES  -  I3.
   * C52102D WRONG VALUES  -  I4.
   * C52102D WRONG VALUES  -  B3.
   * C52102D WRONG VALUES  -  B4.
**** C52102D FAILED ****************************.

,.,. C62002A ACATS 2.5 05-04-27 01:37:15
---- C62002A CHECK THAT COMPONENTS OF ACCESS IN PARAMETERS MAY BE USED
                IN ASSIGNMENT CONTEXTS.
   * C62002A ASSIGNMENT TO COMPONENT OF ARRAY ACCESS PARAMETER FAILED.
**** C62002A FAILED ****************************.

-- 
           Summary: ACATS wrong array copy code c41203a c41203b c52102b
                    c52102d c62002a
           Product: gcc
           Version: 4.1.0
            Status: UNCONFIRMED
          Keywords: wrong-code
          Severity: normal
          Priority: P2
         Component: ada
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: laurent at guerby dot net
                CC: gcc-bugs at gcc dot gnu dot org


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21242

[Bug ada/21242] ACATS wrong array copy code c52102b c52102d

[laurent at guerby dot net]

------- Additional Comments From laurent at guerby dot net  2005-06-11 23:16 -------
C52102B and C52102D still failing as of 4.1.0 Fri Jun 10 20:07:24 UTC 2005 on
x86_64-linux.

C41203A C41203B C62002A are no longer failing.


-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|ACATS wrong array copy code |ACATS wrong array copy code
                   |c41203a c41203b c52102b     |c52102b c52102d
                   |c52102d c62002a             |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21242

[Bug ada/21242] ACATS wrong array copy code c52102b c52102d (works with -fno-tree-dce)

[laurent at guerby dot net]

------- Additional Comments From laurent at guerby dot net  2005-06-19 11:14 -------
On x86_64 and x86-linux as of LAST_UPDATED: Sun Jun 19 07:48:06 UTC 2005, still
failing. More information:

Works at -O0, -O1, fails at -O2 but
works at -O2 -fno-tree-dce


-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|ACATS wrong array copy code |ACATS wrong array copy code
                   |c52102b c52102d             |c52102b c52102d (works with
                   |                            |-fno-tree-dce)


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21242

[Bug ada/21242] ACATS wrong array copy code c52102b c52102d (works with -fno-tree-dce)

[laurent at guerby dot net]

------- Additional Comments From laurent at guerby dot net  2005-06-19 11:46 -------
Here is a reduced test case, works at -0O, fails at -O1 works at -O1
-fno-tree-dce. First .tNN to be different is p.adb.t24.forwprop1.

--
procedure P is
   function F return Integer is
   begin
      return 1;
   end F;
   A : array (-F .. 1) of Boolean;
begin
   A := (False,True,False);
   A(-1 .. 0)  :=  A (0 .. 1);
   if A /= (True, False, False) then
      raise Program_Error;
   end if;
end P;


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21242

[Bug ada/21242] ACATS wrong array copy code c52102b c52102d (works with -fno-tree-dce)

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2005-06-19 12:33 -------
Huh:
  #   VUSE <D.486_68>;
  #   VUSE <C.26_118>;
  (*a.3_27)[-1 ...]{lb: -1 sz: 1} = D.486;

Why is there no V_MUST_DEF/V_MAY_DEF in there?

This is an aliasing bug.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |alias


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21242

[Bug ada/21242] [4.1 Regression] wrong array copy code (ACATS c52102b c52102d)

[pinskia at gcc dot gnu dot org]

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|ACATS wrong array copy code |[4.1 Regression] wrong array
                   |c52102b c52102d (works with |copy code (ACATS c52102b
                   |-fno-tree-dce)              |c52102d)
   Target Milestone|---                         |4.1.0


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21242

[Bug ada/21242] [4.1 Regression] wrong array copy code (ACATS c52102b c52102d)

[laurent at guerby dot net]

------- Additional Comments From laurent at guerby dot net  2005-06-19 13:25 -------
Note that with the following patch on the reduced testcase (supressing the
anonymous array) it works:

-   A : array (-F .. 1) of Boolean;
+   type BA is array (Integer range <>) of Boolean;
+   A : BA (-F .. 1);


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21242

[Bug ada/21242] [4.1 Regression] wrong array copy code (ACATS c52102b c52102d)

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2005-06-19 13:46 -------
Confirmed.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |dnovillo at gcc dot gnu dot
                   |                            |org
             Status|UNCONFIRMED                 |NEW
     Ever Confirmed|                            |1
   Last reconfirmed|0000-00-00 00:00:00         |2005-06-19 13:46:50
               date|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21242

[Bug ada/21242] [4.1 Regression] wrong array copy code (ACATS c52102b c52102d)

[laurent at guerby dot net]

------- Additional Comments From laurent at guerby dot net  2005-07-12 08:26 -------
Fixed between:

LAST_UPDATED: Sat Jul  9 23:06:33 UTC 2005
LAST_UPDATED: Mon Jul 11 22:29:49 UTC 2005 

on x86-linux and x86_64-linux (don't know what patch though).


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


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21242

[Bug ada/21242] [4.1 Regression] wrong array copy code (ACATS c52102b c52102d)

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2005-07-12 15:07 -------
(In reply to comment #7)
> Fixed between:
> 
> LAST_UPDATED: Sat Jul  9 23:06:33 UTC 2005
> LAST_UPDATED: Mon Jul 11 22:29:49 UTC 2005 

Known to be fixed before that:
LAST_UPDATED: Sun Jul 10 00:31:54 UTC 2005

It was most likely fixed by:
2005-07-09  Diego Novillo  <dnovillo@redhat.com>

        * Makefile.in (tree-ssa-alias.o): Depend on tree-ssa-structalias.h
        * tree-cfg.c (CHECK_OP): Only test for is_gimple_val.
        * tree-dfa.c (dump_subvars_for): New.
        (debug_subvars_for): New.
        (dump_variable): Show subvariables if VAR has them.
        * tree-flow-inline.h (get_subvar_at): New.
        (overlap_subvar): Change offset and size to unsigned HOST_WIDE_INT.
        * tree-flow.h (struct ptr_info_def): Remove field pt_malloc.
        Update all users.
        (struct subvar): Change fields offset and size to unsigned
        HOST_WIDE_INT.
        (dump_subvars_for): Declare.
        (debug_subvars_for): Declare.
        (get_subvar_at): Declare.



-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21242