[Bug c++/44108] New: [4.6 Regression] -Wunused-but-set-variable does not consider array sizing use of a const variable

[Bug c++/44108] New: [4.6 Regression] -Wunused-but-set-variable does not consider array sizing use of a const variable

[rubidium at openttd dot org]

Like PR c/43981, but still happening with 4.6.0 20100513 (Last Changed Rev:
159356
), which is why I copied and editted the title.

In case a const variable is used for array sizing it is not considered to be
read whereas a non-const variable would be considered to be read. It does NOT
happen when using gcc on a .c file, but it happens when using g++ on a .cpp
file, as such component c++.

See the attached test case which shows that only the const variables are
affected by this.


-- 
           Summary: [4.6 Regression] -Wunused-but-set-variable does not
                    consider array sizing use of a const variable
           Product: gcc
           Version: 4.6.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c++
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: rubidium at openttd dot org
 GCC build triplet: x86_64-unknown-linux-gnu
  GCC host triplet: x86_64-unknown-linux-gnu
GCC target triplet: x86_64-unknown-linux-gnu


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44108

[Bug c++/44108] [4.6 Regression] -Wunused-but-set-variable does not consider array sizing use of a const variable

[rubidium at openttd dot org]

------- Comment #1 from rubidium at openttd dot org  2010-05-13 11:04 -------
Created an attachment (id=20655)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=20655&action=view)
Simple testcase

Compile with g++ -Wunused-but-set-variable testcase.cpp


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44108

[Bug c++/44108] [4.6 Regression] -Wunused-but-set-variable does not consider array sizing use of a const variable

[rguenth at gcc dot gnu dot org]

-- 

rguenth at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |4.6.0


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44108

[Bug c++/44108] [4.6 Regression] -Wunused-but-set-variable does not consider array sizing use of a const variable

[dodji at gcc dot gnu dot org]

-- 

dodji at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         AssignedTo|unassigned at gcc dot gnu   |dodji at gcc dot gnu dot org
                   |dot org                     |
             Status|UNCONFIRMED                 |ASSIGNED
     Ever Confirmed|0                           |1
   Last reconfirmed|0000-00-00 00:00:00         |2010-05-14 06:42:26
               date|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44108

[Bug c++/44108] [4.6 Regression] -Wunused-but-set-variable does not consider array sizing use of a const variable

[jakub at gcc dot gnu dot org]

------- Comment #2 from jakub at gcc dot gnu dot org  2010-05-14 07:25 -------
Perhaps we just shouldn't ever warn for TREE_CONSTANT decls with
-Wunused-but-set-* ?
Such variables can't be ever set (the set is just the definition),  
therefore all sets of TREE_USED for them are reads. 
That said, it would be good to find out why we don't warn in that  
case; for [M_CONST + 6] we do warn. 
Or for DECL_INTEGRAL_CONSTANT_VAR_P decls, or for decls for which  
integral_constant_value (decl) != decl not sure what is best.
In any case, compute_array_index_type (and probably many other spots)
calls integral_constant_value and thus
doesn't see the passed in decl anymore afterwards.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44108

[Bug c++/44108] [4.6 Regression] -Wunused-but-set-variable does not consider array sizing use of a const variable

[dodji at gcc dot gnu dot org]

------- Comment #3 from dodji at gcc dot gnu dot org  2010-05-17 12:30 -------
Created an attachment (id=20678)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=20678&action=view)
Candidate patch

On Fri, May 14, 2010 at 07:25:18AM -0000, jakub at gcc dot gnu dot org wrote:
> That said, it would be good to find out why we don't warn in that
> case; for [M_CONST + 6] we do warn.

I think we don't warn in that case because M_CONST in the expression
M_CONST + 6, M_CONST is an rvalue use and we recognize it as such via
default_conversion that should be called on the arguments of the + expression.

We fail to recognize the use of [M_CONST], because M_CONST gets folded
into it's value without us marking that folding as an rvalue use. I
am thus testing the attached patch. The reason why the patch seems
convoluted is that I don't want to mark a dependant value as a rvalue
use. I believe only full instantiations should be marked.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44108

[Bug c++/44108] [4.6 Regression] -Wunused-but-set-variable does not consider array sizing use of a const variable

[dodji at gcc dot gnu dot org]

------- Comment #4 from dodji at gcc dot gnu dot org  2010-05-17 13:54 -------
Created an attachment (id=20681)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=20681&action=view)
Don't warn for integral constants

Jakub, this one line patch implements the first idea which is to not warn for
integral constants, as far as Wunused_but_set_variable is concerned.

I have tested it only for the reported error case. I haven't regtested it yet.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44108

[Bug c++/44108] [4.6 Regression] -Wunused-but-set-variable does not consider array sizing use of a const variable

[jakub at gcc dot gnu dot org]

------- Comment #5 from jakub at gcc dot gnu dot org  2010-05-17 17:25 -------
Subject: Bug 44108

Author: jakub
Date: Mon May 17 17:24:32 2010
New Revision: 159497

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=159497
Log:
        PR c++/44108
        * decl.c (compute_array_index_type): Call mark_rvalue_use.

        * c-c++-common/Wunused-var-8.c: New test.

Added:
    trunk/gcc/testsuite/c-c++-common/Wunused-var-8.c
Modified:
    trunk/gcc/cp/ChangeLog
    trunk/gcc/cp/decl.c
    trunk/gcc/testsuite/ChangeLog


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44108

[Bug c++/44108] [4.6 Regression] -Wunused-but-set-variable does not consider array sizing use of a const variable

[dodji at gcc dot gnu dot org]

------- Comment #6 from dodji at gcc dot gnu dot org  2010-05-17 19:45 -------
This should be fixed in trunk (4.6)


-- 

dodji at gcc dot gnu dot org changed:

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


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44108