[Bug c++/13086] New: warning: `e' has incomplete type without location

[Bug c++/13086] New: warning: `e' has incomplete type without location

[gcc-bugzilla at gcc dot gnu dot org]

The following piece of code:

/tmp % cat -n foo.cc                                             nostromo Err 1
     1  class Exp;
     2  Exp* e;
     3  
     4  int
     5  main (void)
     6  {
     7    delete e;
     8  }

Triggers an error with 3.2, but a warning with both 3.3 and 3.4.  The warning
is:

/tmp % g++-3.2 -Wall foo.cc; echo STATUS: $?                     nostromo 17:49
foo.cc: In function `int main()':
foo.cc:2: `e' has incomplete type
foo.cc:1: forward declaration of `struct Exp'
STATUS: 1
/tmp % g++-3.3 -Wall foo.cc; echo STATUS: $?                     nostromo 17:50
foo.cc: In function `int main()':
foo.cc:2: warning: `e' has incomplete type
foo.cc:1: warning: forward declaration of `struct Exp'
STATUS: 0
/tmp % g++-3.4 -Wall foo.cc; echo STATUS: $?                     nostromo 17:50
foo.cc: In function `int main()':
foo.cc:2: AVERTISSEMENT:  e  a un type incomplet
foo.cc:1: AVERTISSEMENT: forward declaration of `struct Exp'
STATUS: 0

as you can see, the problem is that the class Exp is unknown when the
dtor is called via delete.  Nevertheless the reason for the complaints
is _never_ made explicit: it took me a while to understand that it
is the call to delete that was responsible for the warning is my Real
Life Code.  I would suggest that G++ would point to the line 7
when saying ``e' has incomplete type' instead of pointing to the
declaration of `e' (or its declaration could be added, but the current
way seem to be complaining that `e's declaration is incomplete!).

Thanks for all the good work!

Environment:
System: Linux nostromo 2.4.22 #1 SMP mer sep 17 19:49:48 CEST 2003 i686 GNU/Linux
Architecture: i686

	
host: i486-pc-linux-gnu
build: i486-pc-linux-gnu
target: i486-pc-linux-gnu
configured with: ../src/configure -v --enable-languages=c,c++,java,f77,pascal,objc,ada,treelang --prefix=/usr --mandir=/usr/share/man --infodir=/usr/share/info --with-gxx-include-dir=/usr/include/c++/3.3 --enable-shared --with-system-zlib --enable-nls --without-included-gettext --enable-__cxa_atexit --enable-clocale=gnu --enable-debug --enable-java-gc=boehm --enable-java-awt=xlib --enable-objc-gc i486-linux

How-To-Repeat:
Compile the sample above.
------- Additional Comments From akim at epita dot fr  2003-11-17 16:58 -------
Fix:
See above.

-- 
           Summary: warning: `e' has incomplete type without location
           Product: gcc
           Version: 3.3.2
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c++
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: akim at epita dot fr
                CC: gcc-bugs at gcc dot gnu dot org
 GCC build triplet: i486-pc-linux-gnu
  GCC host triplet: i486-pc-linux-gnu
GCC target triplet: i486-pc-linux-gnu


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=13086

[Bug c++/13086] [3.3/3.4 regression] no error when calling delete on incomplete type, and no location

[bangerth at dealii dot org]

------- Additional Comments From bangerth at dealii dot org  2003-11-17 17:44 -------
Confirmed. There are at least two regressions here: up to
2.95 we issued the line of the code that tried to call operator delete,
and we actually issued a hard error instead of a warning (I think
I recall that a hard error is required).

W.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
     Ever Confirmed|                            |1
           Keywords|                            |diagnostic
   Last reconfirmed|0000-00-00 00:00:00         |2003-11-17 17:44:15
               date|                            |
            Summary|warning: `e' has incomplete |[3.3/3.4 regression] no
                   |type without location       |error when calling delete on
                   |                            |incomplete type, and no
                   |                            |location
   Target Milestone|---                         |3.3.3


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=13086

[Bug c++/13086] [3.3/3.4 regression] no error when calling delete on incomplete type, and no location

[falk dot hueffner at student dot uni-tuebingen dot de]

------- Additional Comments From falk dot hueffner at student dot uni-tuebingen dot de  2003-11-17 17:51 -------
Subject: Re:  [3.3/3.4 regression] no error when calling delete on incomplete type, and no location

"bangerth at dealii dot org" <gcc-bugzilla@gcc.gnu.org> writes:

> Confirmed. There are at least two regressions here: up to
> 2.95 we issued the line of the code that tried to call operator delete,
> and we actually issued a hard error instead of a warning (I think
> I recall that a hard error is required).

I don't think it is. The standard says:

"If the object being deleted has incomplete class type at the point of
deletion and the complete class has a non-trivial destructor or a
deallocation function, the behavior is undefined."

which would be pointless if this was a hard error.



-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=13086

[Bug c++/13086] [3.3/3.4 regression] no error when calling delete on incomplete type, and no location

[bangerth at dealii dot org]

------- Additional Comments From bangerth at dealii dot org  2003-11-17 20:46 -------
I stand corrected (although I must admit that I don't understand why
the standard would choose to make something undefined behavior that
is so easy to detect; at least we warn about this).

The problem with the unspecified source line remains, though.
W.

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=13086

[Bug c++/13086] [3.3/3.4 regression] the location of the warning message is wrong when calling delete on incomplete type

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2003-11-17 21:44 -------
Changing summary based on analysis.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|[3.3/3.4 regression] no     |[3.3/3.4 regression] the
                   |error when calling delete on|location of the warning
                   |incomplete type, and no     |message is wrong when
                   |location                    |calling delete on incomplete
                   |                            |type


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=13086

[Bug c++/13086] [3.3/3.4 regression] the location of the warning message is wrong when calling delete on incomplete type

[gdr at integrable-solutions dot net]

------- Additional Comments From gdr at integrable-solutions dot net  2003-11-19 04:36 -------
Subject: Re:  [3.3/3.4 regression] no error when calling delete on incomplete type, and no location

"falk dot hueffner at student dot uni-tuebingen dot de" <gcc-bugzilla@gcc.gnu.org> writes:

| ------- Additional Comments From falk dot hueffner at student dot uni-tuebingen dot de  2003-11-17 17:51 -------
| Subject: Re:  [3.3/3.4 regression] no error when calling delete on incomplete type, and no location
| 
| "bangerth at dealii dot org" <gcc-bugzilla@gcc.gnu.org> writes:
| 
| > Confirmed. There are at least two regressions here: up to
| > 2.95 we issued the line of the code that tried to call operator delete,
| > and we actually issued a hard error instead of a warning (I think
| > I recall that a hard error is required).
| 
| I don't think it is. The standard says:
| 
| "If the object being deleted has incomplete class type at the point of
| deletion and the complete class has a non-trivial destructor or a
| deallocation function, the behavior is undefined."
| 
| which would be pointless if this was a hard error.


But GCC can decide to define that behaviour to be hard error.

I believe I discussed -- a while ago -- this specific issue with Jason, Alexandre Oliva.
I argued that the compiler should issue a hard error.  I was in minority :-(

I still think, this should be a hard error.

-- Gaby


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=13086

[Bug c++/13086] [3.3/3.4 regression] the location of the warning message is wrong when calling delete on incomplete type

[falk at debian dot org]

------- Additional Comments From falk at debian dot org  2003-11-19 09:56 -------
(In reply to comment #6)
> "falk dot hueffner at student dot uni-tuebingen dot de"
<gcc-bugzilla@gcc.gnu.org> writes:
> 
> | "If the object being deleted has incomplete class type at the point of
> | deletion and the complete class has a non-trivial destructor or a
> | deallocation function, the behavior is undefined."
> | 
> | which would be pointless if this was a hard error.
> 
> But GCC can decide to define that behaviour to be hard error.

Only if both conditions are given. The object here has incomplete class,
but it does not have a destructor (well, not shown above, but let's assume.).

> I believe I discussed -- a while ago -- this specific issue with Jason,
Alexandre Oliva.
> I argued that the compiler should issue a hard error.  I was in minority :-(
> 
> I still think, this should be a hard error.

I wouldn't mind at all if it was, because this is a completely useless
feature which could potentially do some damage, but I think it would not be
standard compliant.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=13086

[Bug c++/13086] [3.3/3.4 regression] the location of the warning message is wrong when calling delete on incomplete type

[gdr at integrable-solutions dot net]

------- Additional Comments From gdr at integrable-solutions dot net  2003-11-20 03:16 -------
Subject: Re:  [3.3/3.4 regression] the location of the warning message is wrong when calling delete on incomplete type

"falk at debian dot org" <gcc-bugzilla@gcc.gnu.org> writes:

| (In reply to comment #6)
| > "falk dot hueffner at student dot uni-tuebingen dot de"
| <gcc-bugzilla@gcc.gnu.org> writes:
| > 
| > | "If the object being deleted has incomplete class type at the point of
| > | deletion and the complete class has a non-trivial destructor or a
| > | deallocation function, the behavior is undefined."
| > | 
| > | which would be pointless if this was a hard error.
| > 
| > But GCC can decide to define that behaviour to be hard error.
| 
| Only if both conditions are given. The object here has incomplete class,
| but it does not have a destructor (well, not shown above, but let's assume.).

Any object type has a destructor.  The fact that the type is incomplete at
the point of deleting does not morally rule out that fact. 
Deleting an object with an incomplete type simply does not make sense.

| > I believe I discussed -- a while ago -- this specific issue with Jason,
| Alexandre Oliva.
| > I argued that the compiler should issue a hard error.  I was in minority :-(
| > 
| > I still think, this should be a hard error.
| 
| I wouldn't mind at all if it was, because this is a completely useless
| feature which could potentially do some damage, but I think it would not be
| standard compliant.

Certainly, it is standard conformant.  The issue here is purely a
Quality of Implementation issue. 

-- Gaby  


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=13086

[Bug c++/13086] [3.3/3.4 regression] the location of the warning message is wrong when calling delete on incomplete type

[gdr at gcc dot gnu dot org]

------- Additional Comments From gdr at gcc dot gnu dot org  2003-12-24 22:07 -------
Will investigate.

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=13086

[Bug c++/13086] [3.3/3.4 regression] the location of the warning message is wrong when calling delete on incomplete type

[pinskia at gcc dot gnu dot org]

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
           Severity|normal                      |minor


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=13086

[Bug c++/13086] [3.3/3.4 regression] the location of the warning message is wrong when calling delete on incomplete type

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-01-11 01:50 -------
Since this only about a small problem with the error message, I am moving this to 3.4.1.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|3.3.3                       |3.4.1


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=13086

[Bug c++/13086] [3.3/3.4 regression] the location of the warning message is wrong when calling delete on incomplete type

[steven at gcc dot gnu dot org]

------- Additional Comments From steven at gcc dot gnu dot org  2004-01-12 23:39 -------
Move back the target for all regressions from 3.4.1 to 3.4.0, as required by 
our bug management policy. 

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|3.4.1                       |3.4.0


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=13086

[Bug c++/13086] [3.3/3.4 regression] the location of the warning message is wrong when calling delete on incomplete type

[giovannibajo at libero dot it]

------- Additional Comments From giovannibajo at libero dot it  2004-01-15 03:00 -------
I'm testing a patch for this.

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


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=13086

[Bug c++/13086] [3.3/3.4 regression] the location of the warning message is wrong when calling delete on incomplete type

[giovannibajo at libero dot it]

------- Additional Comments From giovannibajo at libero dot it  2004-01-15 12:37 -------
Patch submitted, waiting for review:
http://gcc.gnu.org/ml/gcc-patches/2004-01/msg01457.html


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


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=13086

[Bug c++/13086] [3.3/3.4/3.5 regression] the location of the warning message is wrong when calling delete on incomplete type

[mmitchel at gcc dot gnu dot org]

------- Additional Comments From mmitchel at gcc dot gnu dot org  2004-02-04 18:42 -------
Reword this:

note: the generated code will not work, unless the class does
not have a destructor nor a delete operator.

as:

note: neither the destructor nor the class-specific operator delete will be
called, even if they are declared when the class is defined.

I believe that takes into account Gaby's and Jason's comments.

Then, please check in the patch.

Thanks.

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=13086

[Bug c++/13086] [3.3/3.4/3.5 regression] the location of the warning message is wrong when calling delete on incomplete type

[cvs-commit at gcc dot gnu dot org]

------- Additional Comments From cvs-commit at gcc dot gnu dot org  2004-02-05 02:48 -------
Subject: Bug 13086

CVSROOT:	/cvs/gcc
Module name:	gcc
Changes by:	giovannibajo@gcc.gnu.org	2004-02-05 02:48:31

Modified files:
	gcc/cp         : ChangeLog init.c 

Log message:
	PR c++/13086
	* init.c (build_delete): Emit a more informative error message in
	case of an incomplete type, and on the correct source line.

Patches:
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/cp/ChangeLog.diff?cvsroot=gcc&r1=1.3935&r2=1.3936
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/cp/init.c.diff?cvsroot=gcc&r1=1.360&r2=1.361



-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=13086

[Bug c++/13086] [3.3/3.4/3.5 regression] the location of the warning message is wrong when calling delete on incomplete type

[cvs-commit at gcc dot gnu dot org]

------- Additional Comments From cvs-commit at gcc dot gnu dot org  2004-02-05 02:49 -------
Subject: Bug 13086

CVSROOT:	/cvs/gcc
Module name:	gcc
Changes by:	giovannibajo@gcc.gnu.org	2004-02-05 02:49:32

Modified files:
	gcc/testsuite  : ChangeLog 
	gcc/testsuite/g++.dg/warn: incomplete1.C 

Log message:
	PR c++/13086
	* g++.dg/warn/incomplete1.C: Remove xfail.

Patches:
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/testsuite/ChangeLog.diff?cvsroot=gcc&r1=1.3469&r2=1.3470
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/testsuite/g++.dg/warn/incomplete1.C.diff?cvsroot=gcc&r1=1.1&r2=1.2



-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=13086

[Bug c++/13086] [3.3/3.4/3.5 regression] the location of the warning message is wrong when calling delete on incomplete type

[cvs-commit at gcc dot gnu dot org]

------- Additional Comments From cvs-commit at gcc dot gnu dot org  2004-02-05 02:51 -------
Subject: Bug 13086

CVSROOT:	/cvs/gcc
Module name:	gcc
Branch: 	gcc-3_4-branch
Changes by:	giovannibajo@gcc.gnu.org	2004-02-05 02:51:22

Modified files:
	gcc/cp         : ChangeLog init.c 

Log message:
	PR c++/13086
	* init.c (build_delete): Emit a more informative error message in
	case of an incomplete type, and on the correct source line.

Patches:
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/cp/ChangeLog.diff?cvsroot=gcc&only_with_tag=gcc-3_4-branch&r1=1.3892.2.35&r2=1.3892.2.36
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/cp/init.c.diff?cvsroot=gcc&only_with_tag=gcc-3_4-branch&r1=1.356.2.4&r2=1.356.2.5



-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=13086

[Bug c++/13086] [3.3/3.4/3.5 regression] the location of the warning message is wrong when calling delete on incomplete type

[cvs-commit at gcc dot gnu dot org]

------- Additional Comments From cvs-commit at gcc dot gnu dot org  2004-02-05 02:52 -------
Subject: Bug 13086

CVSROOT:	/cvs/gcc
Module name:	gcc
Branch: 	gcc-3_4-branch
Changes by:	giovannibajo@gcc.gnu.org	2004-02-05 02:52:39

Modified files:
	gcc/testsuite  : ChangeLog 
	gcc/testsuite/g++.dg/warn: incomplete1.C 

Log message:
	PR c++/13086
	* g++.dg/warn/incomplete1.C: Remove xfail.

Patches:
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/testsuite/ChangeLog.diff?cvsroot=gcc&only_with_tag=gcc-3_4-branch&r1=1.3389.2.51&r2=1.3389.2.52
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/testsuite/g++.dg/warn/incomplete1.C.diff?cvsroot=gcc&only_with_tag=gcc-3_4-branch&r1=1.1&r2=1.1.50.1



-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=13086

[Bug c++/13086] [3.3 Regression] the location of the warning message is wrong when calling delete on incomplete type

[giovannibajo at libero dot it]

------- Additional Comments From giovannibajo at libero dot it  2004-02-05 02:55 -------
Fixed in 3.4.0 and mainline. Now it's a regression in the 3.3 branch only.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|[3.3/3.4/3.5 regression] the|[3.3 Regression] the
                   |location of the warning     |location of the warning
                   |message is wrong when       |message is wrong when
                   |calling delete on incomplete|calling delete on incomplete
                   |type                        |type
   Target Milestone|3.4.0                       |3.3.4


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=13086

[Bug c++/13086] [3.3 Regression] the location of the warning message is wrong when calling delete on incomplete type

[gdr at integrable-solutions dot net]

------- Additional Comments From gdr at integrable-solutions dot net  2004-02-05 02:59 -------
Subject: Re:  [3.3/3.4/3.5 regression] the location of the warning message is wrong when calling delete on incomplete type

"cvs-commit at gcc dot gnu dot org" <gcc-bugzilla@gcc.gnu.org> writes:

| Branch: 	gcc-3_4-branch
| Changes by:	giovannibajo@gcc.gnu.org	2004-02-05 02:52:39

It is OK for gcc-3_3-branch too.

-- Gaby


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=13086

[Bug c++/13086] [3.3 Regression] the location of the warning message is wrong when calling delete on incomplete type

[giovannibajo at libero dot it]

------- Additional Comments From giovannibajo at libero dot it  2004-02-05 03:39 -------
Subject: Re:  [3.3 Regression] the location of the warning message is wrong when calling delete on incomplete type

> It is OK for gcc-3_3-branch too.

I just tested it there, no regressions. Is the branch still freezed, isn't it?

Giovanni Bajo




-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=13086

[Bug c++/13086] [3.3 Regression] the location of the warning message is wrong when calling delete on incomplete type

[cvs-commit at gcc dot gnu dot org]

------- Additional Comments From cvs-commit at gcc dot gnu dot org  2004-02-05 13:22 -------
Subject: Bug 13086

CVSROOT:	/cvs/gcc
Module name:	gcc
Branch: 	gcc-3_3-branch
Changes by:	giovannibajo@gcc.gnu.org	2004-02-05 13:22:11

Modified files:
	gcc/cp         : ChangeLog init.c 

Log message:
	PR c++/13086
	* init.c (build_delete): Emit a more informative error message in
	case of an incomplete type, and on the correct source line.

Patches:
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/cp/ChangeLog.diff?cvsroot=gcc&only_with_tag=gcc-3_3-branch&r1=1.3076.2.246&r2=1.3076.2.247
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/cp/init.c.diff?cvsroot=gcc&only_with_tag=gcc-3_3-branch&r1=1.299.2.13&r2=1.299.2.14



-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=13086

[Bug c++/13086] [3.3 Regression] the location of the warning message is wrong when calling delete on incomplete type

[cvs-commit at gcc dot gnu dot org]

------- Additional Comments From cvs-commit at gcc dot gnu dot org  2004-02-05 13:24 -------
Subject: Bug 13086

CVSROOT:	/cvs/gcc
Module name:	gcc
Branch: 	gcc-3_3-branch
Changes by:	giovannibajo@gcc.gnu.org	2004-02-05 13:24:07

Modified files:
	gcc/testsuite  : ChangeLog 
	gcc/testsuite/g++.dg/warn: incomplete1.C 

Log message:
	PR c++/13086
	* g++.dg/warn/incomplete1.C: Remove xfail.

Patches:
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/testsuite/ChangeLog.diff?cvsroot=gcc&only_with_tag=gcc-3_3-branch&r1=1.2261.2.355&r2=1.2261.2.356
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/testsuite/g++.dg/warn/incomplete1.C.diff?cvsroot=gcc&only_with_tag=gcc-3_3-branch&r1=1.1&r2=1.1.22.1



-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=13086

[Bug c++/13086] [3.3 Regression] the location of the warning message is wrong when calling delete on incomplete type

[giovannibajo at libero dot it]

------- Additional Comments From giovannibajo at libero dot it  2004-02-05 13:25 -------
Patch backported. This bug is now fixed in any future GCC version since 3.3.3. 
Thanks for your report!

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


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=13086

[Bug c++/13086] [3.3 Regression] the location of the warning message is wrong when calling delete on incomplete type

[giovannibajo at libero dot it]

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|3.3.4                       |3.3.3


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=13086