[Bug c++/14724] New: Destructor not called on backwards goto past initialization

[Bug c++/14724] New: Destructor not called on backwards goto past initialization

[gson at gson dot org]

When the following program is compiled with g++ 3.3.2 and executed, it
will invoke the constructor of the object "v" 10 times in a row with
no intervening calls to the destructor, as evidenced by it printing
"ctor" 10 times but "dtor" only once.

According to ISO/IEC 14882:1998(E) section 6.6, the destructor of v
should be invoked each time the goto statement is executed:
"a transfer [...] back past an initialized variable with automatic
storage duration involves the destruction of variables with automatic
storage duration that are in scope at the point transferred from but
not at the point transferred to".

The program works correctly when compiled with gcc 2.95.

#include <iostream>

template <class T>
struct C {
        C() { std::cout << "ctor\n"; }
        ~C() { std::cout << "dtor\n"; }
};

int main(int, char **) {
        int i = 0;
 again:
        C<int> v;
        if (++i < 10)
                goto again;
}

-- 
           Summary: Destructor not called on backwards goto past
                    initialization
           Product: gcc
           Version: 3.3.2
            Status: UNCONFIRMED
          Severity: critical
          Priority: P1
         Component: c++
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: gson at gson dot org
                CC: gcc-bugs at gcc dot gnu dot org
 GCC build triplet: i386--netbsdelf
  GCC host triplet: i386--netbsdelf
GCC target triplet: i386--netbsdelf


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=14724

[Bug c++/14724] Destructor not called on backwards goto past initialization

[bangerth at dealii dot org]

------- Additional Comments From bangerth at dealii dot org  2004-03-24 23:00 -------
Confirmed, a bad regression in 3.3/3.4/3.5 w.r.t. 2.95. 
 
W. 

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
     Ever Confirmed|                            |1
           Keywords|                            |wrong-code
      Known to fail|                            |3.2.3 3.3.3 3.4.0 3.5.0
      Known to work|                            |2.95
   Last reconfirmed|0000-00-00 00:00:00         |2004-03-24 23:00:49
               date|                            |
   Target Milestone|---                         |3.3.4


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=14724

[Bug c++/14724] [3.3/3.4/3.5 Regression] Destructor not called on backwards goto past initialization

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-03-25 01:25 -------
Note this is already fixed on the tree-ssa:
ctor
dtor
ctor
dtor
ctor
dtor
ctor
dtor
ctor
dtor
ctor
dtor
ctor
dtor
ctor
dtor
ctor
dtor
ctor
dtor

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
      Known to work|2.95                        |2.95 tree-ssa
            Summary|Destructor not called on    |[3.3/3.4/3.5 Regression]
                   |backwards goto past         |Destructor not called on
                   |initialization              |backwards goto past
                   |                            |initialization


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=14724

[Bug c++/14724] [3.3/3.4/3.5 Regression] Destructor not called on backwards goto past initialization

[giovannibajo at libero dot it]

------- Additional Comments From giovannibajo at libero dot it  2004-03-26 01:40 -------
For the logs, I can reproduce it with:
gcc version 3.3.4 20040218 (prerelease)
gcc version 3.4.0 20040216 (prerelease)
gcc version 3.2.2 (mingw special 20030208-1)
on i686-cygwin. 

Andrew Pinski can reproduce it with:
gcc version 3.5.0 20040325 (experimental)
on i686-linux and powerpc-apple-darwin. 

But RTH couldn't reproduce it with gcc version 3.3.3 (Debian) on alpha, or 
mainline on linux. So there must be something weird going on.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=14724

[Bug c++/14724] [3.3/3.4/3.5 Regression] Destructor not called on backwards goto past initialization

[bangerth at dealii dot org]

------- Additional Comments From bangerth at dealii dot org  2004-03-26 15:14 -------
Hm, I can reproduce this with 
gcc version 3.5.0 20040307 (experimental) 
gcc version 3.4.0 20040319 (prerelease) 
gcc version 3.3.4 20040223 (prerelease) 
 
all on x86 linux... 
 
W. 

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=14724

[Bug c++/14724] [3.3/3.4/3.5 Regression] Destructor not called on backwards goto past initialization

[mmitchel at gcc dot gnu dot org]

------- Additional Comments From mmitchel at gcc dot gnu dot org  2004-03-30 22:19 -------
Working on a fix.

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


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=14724

[Bug c++/14724] [3.3/3.4/3.5 Regression] Destructor not called on backwards goto past initialization

[cvs-commit at gcc dot gnu dot org]

------- Additional Comments From cvs-commit at gcc dot gnu dot org  2004-03-30 23:43 -------
Subject: Bug 14724

CVSROOT:	/cvs/gcc
Module name:	gcc
Branch: 	gcc-3_4-branch
Changes by:	mmitchel@gcc.gnu.org	2004-03-30 23:43:20

Modified files:
	gcc/cp         : ChangeLog decl.c pt.c 
	gcc/testsuite  : ChangeLog 
Added files:
	gcc/testsuite/g++.dg/init: goto1.C 
	gcc/testsuite/g++.dg/template: defarg4.C 

Log message:
	PR c++/14724
	* decl.c (start_decl_1): Do not decide whether or not to create a
	new cleanup level until after the type has been completed.
	
	PR c++/14763
	* pt.c (tsubst_default_argument): Clear current_function_decl.
	
	PR c++/14724
	* g++.dg/init/goto1.C: New test.
	
	PR c++/14763
	* g++.dg/template/defarg4.C: New test.

Patches:
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/cp/ChangeLog.diff?cvsroot=gcc&only_with_tag=gcc-3_4-branch&r1=1.3892.2.89&r2=1.3892.2.90
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/cp/decl.c.diff?cvsroot=gcc&only_with_tag=gcc-3_4-branch&r1=1.1174.2.17&r2=1.1174.2.18
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/cp/pt.c.diff?cvsroot=gcc&only_with_tag=gcc-3_4-branch&r1=1.816.2.21&r2=1.816.2.22
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/testsuite/ChangeLog.diff?cvsroot=gcc&only_with_tag=gcc-3_4-branch&r1=1.3389.2.158&r2=1.3389.2.159
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/testsuite/g++.dg/init/goto1.C.diff?cvsroot=gcc&only_with_tag=gcc-3_4-branch&r1=NONE&r2=1.1.2.1
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/testsuite/g++.dg/template/defarg4.C.diff?cvsroot=gcc&only_with_tag=gcc-3_4-branch&r1=NONE&r2=1.1.2.1



-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=14724

[Bug c++/14724] [3.3/3.4/3.5 Regression] Destructor not called on backwards goto past initialization

[cvs-commit at gcc dot gnu dot org]

------- Additional Comments From cvs-commit at gcc dot gnu dot org  2004-03-30 23:45 -------
Subject: Bug 14724

CVSROOT:	/cvs/gcc
Module name:	gcc
Changes by:	mmitchel@gcc.gnu.org	2004-03-30 23:45:00

Modified files:
	gcc/cp         : ChangeLog decl.c pt.c 
	gcc/testsuite  : ChangeLog 
Added files:
	gcc/testsuite/g++.dg/init: goto1.C 
	gcc/testsuite/g++.dg/template: defarg4.C 

Log message:
	PR c++/14724
	* decl.c (start_decl_1): Do not decide whether or not to create a
	new cleanup level until after the type has been completed.
	
	PR c++/14763
	* pt.c (tsubst_default_argument): Clear current_function_decl.
	
	PR c++/14724
	* g++.dg/init/goto1.C: New test.
	
	PR c++/14763
	* g++.dg/template/defarg4.C: New test.

Patches:
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/cp/ChangeLog.diff?cvsroot=gcc&r1=1.4018&r2=1.4019
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/cp/decl.c.diff?cvsroot=gcc&r1=1.1198&r2=1.1199
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/cp/pt.c.diff?cvsroot=gcc&r1=1.844&r2=1.845
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/testsuite/ChangeLog.diff?cvsroot=gcc&r1=1.3646&r2=1.3647
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/testsuite/g++.dg/init/goto1.C.diff?cvsroot=gcc&r1=1.1&r2=1.2
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/testsuite/g++.dg/template/defarg4.C.diff?cvsroot=gcc&r1=1.1&r2=1.2



-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=14724

[Bug c++/14724] [3.3 Regression] Destructor not called on backwards goto past initialization

[mmitchel at gcc dot gnu dot org]

------- Additional Comments From mmitchel at gcc dot gnu dot org  2004-03-30 23:47 -------
Fixed in GCC 3.4 and GCC 3.5.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
         AssignedTo|mark at codesourcery dot com|unassigned at gcc dot gnu
                   |                            |dot org
             Status|ASSIGNED                    |NEW
      Known to fail|3.2.3 3.3.3 3.4.0 3.5.0     |3.2.3 3.3.3
      Known to work|2.95 tree-ssa               |2.95 tree-ssa 3.4.0 3.5.0
            Summary|[3.3/3.4/3.5 Regression]    |[3.3 Regression] Destructor
                   |Destructor not called on    |not called on backwards goto
                   |backwards goto past         |past initialization
                   |initialization              |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=14724

[Bug c++/14724] [3.3 Regression] Destructor not called on backwards goto past initialization

[cvs-commit at gcc dot gnu dot org]

------- Additional Comments From cvs-commit at gcc dot gnu dot org  2004-04-13 20:42 -------
Subject: Bug 14724

CVSROOT:	/cvs/gcc
Module name:	gcc
Branch: 	gcc-3_3-rhl-branch
Changes by:	jakub@gcc.gnu.org	2004-04-13 20:41:57

Modified files:
	gcc/cp         : ChangeLog decl.c decl2.c pt.c 
	gcc/testsuite  : ChangeLog 
Added files:
	gcc/testsuite/g++.dg/init: goto1.C 
	gcc/testsuite/g++.dg/template: defarg4.C 

Log message:
	2004-04-07  Jakub Jelinek  <jakub@redhat.com>
	
	* decl2.c (mark_used): Don't segfault if cfun != NULL but
	current_function_decl == NULL.
	
	2004-03-30  Mark Mitchell  <mark@codesourcery.com>
	
	PR c++/14724
	* decl.c (start_decl_1): Do not decide whether or not to create a
	new cleanup level until after the type has been completed.
	
	PR c++/14763
	* pt.c (tsubst_default_argument): Clear current_function_decl.
	
	PR c++/14724
	* g++.dg/init/goto1.C: New test.
	
	PR c++/14763
	* g++.dg/template/defarg4.C: New test.

Patches:
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/cp/ChangeLog.diff?cvsroot=gcc&only_with_tag=gcc-3_3-rhl-branch&r1=1.3076.2.132.2.29&r2=1.3076.2.132.2.30
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/cp/decl.c.diff?cvsroot=gcc&only_with_tag=gcc-3_3-rhl-branch&r1=1.965.2.42.2.18&r2=1.965.2.42.2.19
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/cp/decl2.c.diff?cvsroot=gcc&only_with_tag=gcc-3_3-rhl-branch&r1=1.575.2.28.2.6&r2=1.575.2.28.2.7
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/cp/pt.c.diff?cvsroot=gcc&only_with_tag=gcc-3_3-rhl-branch&r1=1.635.2.22.2.14&r2=1.635.2.22.2.15
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/testsuite/ChangeLog.diff?cvsroot=gcc&only_with_tag=gcc-3_3-rhl-branch&r1=1.2261.2.170.2.57&r2=1.2261.2.170.2.58
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/testsuite/g++.dg/init/goto1.C.diff?cvsroot=gcc&only_with_tag=gcc-3_3-rhl-branch&r1=NONE&r2=1.2.10.1
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/testsuite/g++.dg/template/defarg4.C.diff?cvsroot=gcc&only_with_tag=gcc-3_3-rhl-branch&r1=NONE&r2=1.2.10.1



-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=14724

[Bug c++/14724] [3.3 Regression] Destructor not called on backwards goto past initialization

[cvs-commit at gcc dot gnu dot org]

------- Additional Comments From cvs-commit at gcc dot gnu dot org  2004-05-01 19:57 -------
Subject: Bug 14724

CVSROOT:	/cvs/gcc
Module name:	gcc
Branch: 	gcc-3_3-branch
Changes by:	gdr@gcc.gnu.org	2004-05-01 19:57:29

Modified files:
	gcc/cp         : ChangeLog decl.c 
Added files:
	gcc/testsuite/g++.dg/init: goto1.C 

Log message:
	Backport from gcc-3_4-branch:
	2004-03-30  Mark Mitchell  <mark@codesourcery.com>
	PR c++/14724
	* decl.c (start_decl_1): Do not decide whether or not to
	create a
	new cleanup level until after the type has been completed.

Patches:
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/testsuite/g++.dg/init/goto1.C.diff?cvsroot=gcc&only_with_tag=gcc-3_3-branch&r1=NONE&r2=1.2.20.1
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/cp/ChangeLog.diff?cvsroot=gcc&only_with_tag=gcc-3_3-branch&r1=1.3076.2.265&r2=1.3076.2.266
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/cp/decl.c.diff?cvsroot=gcc&only_with_tag=gcc-3_3-branch&r1=1.965.2.81&r2=1.965.2.82



-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=14724

[Bug c++/14724] [3.3 Regression] Destructor not called on backwards goto past initialization

[gdr at gcc dot gnu dot org]

------- Additional Comments From gdr at gcc dot gnu dot org  2004-05-01 19:58 -------
Fixed in 3.3.4 too.

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


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=14724