[Bug c/17178] New: [3.5 Regression] static consts and -Wunused-variable

[Bug c/17178] New: [3.5 Regression] static consts and -Wunused-variable

[pinskia at gcc dot gnu dot org]

static const int kUnloadAndDeleteClient = 'N';
gcc -Wunused-variable -Werror t.c
t:1: warning: unused variable `kUnloadAndDeleteClient'

-- 
           Summary: [3.5 Regression] static consts and -Wunused-variable
           Product: gcc
           Version: 3.5.0
            Status: UNCONFIRMED
          Keywords: diagnostic
          Severity: critical
          Priority: P2
         Component: c
        AssignedTo: zack at gcc dot gnu dot org
        ReportedBy: pinskia at gcc dot gnu dot org
                CC: gcc-bugs at gcc dot gnu dot org


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17178

[Bug c/17178] [3.5 Regression] static consts and -Wunused-variable

[pinskia at gcc dot gnu dot org]

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |ASSIGNED
     Ever Confirmed|                            |1
   Last reconfirmed|0000-00-00 00:00:00         |2004-08-24 22:31:56
               date|                            |
   Target Milestone|---                         |3.5.0


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17178

[Bug c/17178] [3.5 Regression] static consts and -Wunused-variable

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-08-24 22:32 -------
Casued by:
        PR 12267, 12391, 12560, 13129, 14114, 14113

        * c-tree.h: Forward declare struct c_binding.  Declare
        c_override_bindings_to_false.  Update prototypes.
        ....

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17178

[Bug c/17178] [3.5 Regression] static consts and -Wunused-variable

[bangerth at dealii dot org]

------- Additional Comments From bangerth at dealii dot org  2004-08-24 22:48 -------
Andrew, is your testcase the following single one line? 
--------------------------- 
static const int kUnloadAndDeleteClient = 'N'; 
--------------------------- 
What is wrong with gcc giving a warning about this when you ask for it? Or what 
exactly is the problem you are thinking of? 
 
W. 

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17178

[Bug c/17178] [3.5 Regression] static consts and -Wunused-variable

[zack at codesourcery dot com]

------- Additional Comments From zack at codesourcery dot com  2004-08-24 22:57 -------
Subject: Re:  New: [3.5 Regression] static consts and -Wunused-variable


This was changed on purpose; please justify changing it back.

zw


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17178

[Bug c/17178] [3.5 Regression] static consts and -Wunused-variable

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-08-24 23:30 -------
Well look into PR13029, I don't see any justification for changing it in the first place.
<http://gcc.gnu.org/ml/gcc-patches/2004-03/msg01958.html>.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
  BugsThisDependsOn|                            |13029


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17178

[Bug c/17178] [3.5 Regression] static consts and -Wunused-variable

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-08-24 23:32 -------
The other thing to look into is the patch from Mark which fixed it for 3.3 and 3.4.0:
<http://gcc.gnu.org/ml/gcc-patches/2003-05/msg00468.html>

So why are you asking me for justification when I don't see any for yours.

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17178

[Bug c/17178] [3.5 Regression] static consts and -Wunused-variable

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-08-24 23:44 -------
Again also look into what Joe Buck posted which says why we should not warn here:
<http://gcc.gnu.org/ml/gcc/2003-05/msg00172.html>.

This should be the justificiation I need.

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17178

[Bug c/17178] [3.5 Regression] static consts and -Wunused-variable

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-08-24 23:51 -------
Warn whenever a local variable or non-constant static variable is unused aside from its declaration

Note the "non-constant static".

Again if this is on purpose as you said it was, why did you not change the documenation then?

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17178

[Bug c/17178] [3.5 Regression] static consts and -Wunused-variable

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-08-24 23:53 -------
Mark, could you settle  this one as Zack said he did it deliberately but did not change the 
documentation because:
[16:32] < pinskia> zwol: http://gcc.gnu.org/ml/gcc-patches/2003-05/msg00468.html
[16:44] * stevenb (~steven@70-90.ipact.nl) has left the channel
[16:45] < zwol> pinskia: you might deduce from my response that i think that change was wrong.
[16:45] < pinskia> well it is still a regression which needs to be fixed.
[16:46] < pinskia> I should note that Mac OS X has a huge number of these static const variables in the 
headers is we are going to warn about them
[16:47] < zwol> static const char*, sure, i can see the rcsid[] argument.  but static const int, that's 
different, that's someone expecting C++ const semantics, and it's very likely their program is not doing 
what they thought it was.
[16:47] < zwol> Mac OS X headers should be fixed then.
[16:48] < zwol> no, i have no sympathy.
[16:48] < zwol> why do you ask?
[16:49] < pinskia> I don't care it is a regression and that is all I am going to say
[16:50] < pinskia> Should I quote the manual  which you did not change when you committed the 
change?
[16:50] < pinskia> Warn whenever a local variable or non-constant static variable is unused aside from 
its declaration
[16:50] < pinskia> "non-constant static"
[16:50] * gadek (~gadek@209.29.internet.uqam.ca) has joined
[16:51] < zwol> I do not consider the manual to be Free and therefore (while I will fix obvious 
problems that are pointed out to me) I will not look through it for problems when I make changes.
[16:51] < zwol> and, my response to "it is a regression" is "no, it is a deliberate change, justify why it 
should be changed back."
[16:51] < zwol> and if you don't care to justify it, then i don't care.
[16:52] < pinskia> but you did not justify why it was changed in the first place which means it was not 
a deliberate change 

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |mmitchel at gcc dot gnu dot
                   |                            |org


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17178

[Bug c/17178] [3.5 Regression] static consts and -Wunused-variable

[mmitchel at gcc dot gnu dot org]

------- Additional Comments From mmitchel at gcc dot gnu dot org  2004-08-25 00:01 -------
I'm not sure I can "settle this"; my voice is no stronger than any other
maintainer on non-release issues.

My feeling is that we should indeed not warn on this case; that's why I made the
change you reference.  I see the logic in Zack's argument that we should target
rcs_id more specifically by using "const char *" -- but I think that given the
previous controversy, we should make changes here only with substantial
consensus.  So, I think that the behavior should be changed back to not warn on
this case. Then, if Zack feels strongly (and is willing to fight the battle) he
could reopen this debate on the mailing lists.  If that doesn't happen, for some
reason, I agree that we should change the manual to reflect the new reality.

Just my two cents, though.

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17178

[Bug c/17178] [4.0 Regression] static consts and -Wunused-variable

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-10-03 23:03 -------
I have a fix, the same patch which fixes PR 17820 also fixes this one, it looks like it was a mistake in 
the patch which broke this.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
         AssignedTo|zack at gcc dot gnu dot org |pinskia at gcc dot gnu dot
                   |                            |org


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17178

[Bug c/17178] [4.0 Regression] static consts and -Wunused-variable

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-10-03 23:19 -------
Patch here: <http://gcc.gnu.org/ml/gcc-patches/2004-10/msg00217.html>.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |patch


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17178

[Bug c/17178] [4.0 Regression] static consts and -Wunused-variable

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-10-04 00:10 -------
Fixed.

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


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17178

[Bug c/17178] [4.0 Regression] static consts and -Wunused-variable

[cvs-commit at gcc dot gnu dot org]

------- Additional Comments From cvs-commit at gcc dot gnu dot org  2004-10-04 00:10 -------
Subject: Bug 17178

CVSROOT:	/cvs/gcc
Module name:	gcc
Changes by:	pinskia@gcc.gnu.org	2004-10-04 00:10:11

Modified files:
	gcc/testsuite/gcc.dg: unused-4.c 
	gcc/testsuite  : ChangeLog 
	gcc            : ChangeLog c-decl.c 
Added files:
	gcc/testsuite/gcc.dg: unused-7.c 

Log message:
	2004-10-03  Andrew Pinski  <pinskia@physics.uc.edu>
	
	PR c/17178, PR c/17820
	* c-decl.c (pop_scope): Do not warn about unused static
	variables as they warned in the middle-end.
	
	2004-10-03  Andrew Pinski  <pinskia@physics.uc.edu>
	
	PR c/17178
	* gcc.dg/unused-4.c: Adjust to be correct.
	
	PR c/17820
	* gcc.dg/unused-7.c: New test.

Patches:
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/testsuite/gcc.dg/unused-7.c.diff?cvsroot=gcc&r1=NONE&r2=1.1
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/testsuite/gcc.dg/unused-4.c.diff?cvsroot=gcc&r1=1.3&r2=1.4
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/testsuite/ChangeLog.diff?cvsroot=gcc&r1=1.4376&r2=1.4377
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/ChangeLog.diff?cvsroot=gcc&r1=2.5735&r2=2.5736
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/c-decl.c.diff?cvsroot=gcc&r1=1.594&r2=1.595



-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17178