public inbox for gcc-bugs@sourceware.org
help / color / mirror / Atom feed
[parent not found: <bug-15774-4@http.gcc.gnu.org/bugzilla/>]
* [Bug c++/15774] New: Conflicting function decls not diagnosed
@ 2004-06-02 10:54 dannysmith at users dot sourceforge dot net
2004-06-02 10:59 ` [Bug c++/15774] " dannysmith at users dot sourceforge dot net
` (5 more replies)
0 siblings, 6 replies; 11+ messages in thread
From: dannysmith at users dot sourceforge dot net @ 2004-06-02 10:54 UTC (permalink / raw)
To: gcc-bugs
The C++ frontend doesn't always call targetm.compare_type_attributes
when comparing function decls in decl.c: decls_match. As a result
conflicting decls are not diagnosed. Here is a testcase.
/* stdcall.c */
extern void foo(int);
void bar() { foo(1);}
void __attribute__((stdcall)) foo(int i) {};
This should casue fatal error because of type conflict between
declaration and definition of foo.
In C it does.
D:\develop\bugs>gcc -S -Wall stdcall.c
> stdcall.c:5: error: conflicting types for 'foo'
> stdcall.c:1: error: previous declaration of 'foo' was here
In C++ it doesn't
D:\develop\bugs>gcc -xc++ -S -Wall stdcall.c
compiles without error.
But it produces bad code. This is easier to spot on windows targets because
of the target specific stdcall decoration.
.file "stdcall.c"
.text
.align 4
.align 2
.globl __Z3barv
.def __Z3barv; .scl 2; .type
32; .endef
__Z3barv:
pushl %ebp
movl %esp, %ebp
subl $8, %esp
subl $12, %esp
pushl $1
call __Z3fooi <<< undefined
addl $16, %esp <<< Should be $12.
leave
ret
.align 2
.globl __Z3fooi@4
.def __Z3fooi@4; .scl 2; .type
32; .endef
__Z3fooi@4:
pushl %ebp
movl %esp, %ebp
leave
ret $4
.def __Z3fooi; .scl 3; .type
32; .endef
The testcase fails in 3.4.0 and 3.5.0 on i386-i386-pc-mingw32
The following patch to cp/declc fixes the testcase. I don't think it is right,
but I'll run it throw bootstrap and regtest and submit proper patch if it
doesn't cause any other problems.
Danny
Index: decl.c
===================================================================
RCS file: /cvs/gcc/gcc/gcc/cp/decl.c,v
retrieving revision 1.1212
diff -c -3 -p -r1.1212 decl.c
*** decl.c 28 May 2004 21:58:15 -0000 1.1212
--- decl.c 2 Jun 2004 10:35:33 -0000
*************** decls_match (tree newdecl, tree olddecl)
*** 1021,1027 ****
}
#endif
else
! types_match = compparms (p1, p2);
}
else
types_match = 0;
--- 1021,1030 ----
}
#endif
else
! types_match = compparms (p1, p2)
! && comptypes (TREE_TYPE (newdecl),
! TREE_TYPE (olddecl),
! COMPARE_REDECLARATION);
}
else
types_match = 0;
--
Summary: Conflicting function decls not diagnosed
Product: gcc
Version: 3.4.0
Status: UNCONFIRMED
Severity: normal
Priority: P2
Component: c++
AssignedTo: unassigned at gcc dot gnu dot org
ReportedBy: dannysmith at users dot sourceforge dot net
CC: gcc-bugs at gcc dot gnu dot org
GCC build triplet: i386-pc-mingw32
GCC host triplet: i386-pc-mingw32
GCC target triplet: i386-pc-mingw32
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=15774
^ permalink raw reply [flat|nested] 11+ messages in thread
* [Bug c++/15774] Conflicting function decls not diagnosed
2004-06-02 10:54 [Bug c++/15774] New: " dannysmith at users dot sourceforge dot net
@ 2004-06-02 10:59 ` dannysmith at users dot sourceforge dot net
2004-06-02 11:37 ` pinskia at gcc dot gnu dot org
` (4 subsequent siblings)
5 siblings, 0 replies; 11+ messages in thread
From: dannysmith at users dot sourceforge dot net @ 2004-06-02 10:59 UTC (permalink / raw)
To: gcc-bugs
--
What |Removed |Added
----------------------------------------------------------------------------
Known to fail| |2.95.3 3.3.3 3.4.0 3.5.0
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=15774
^ permalink raw reply [flat|nested] 11+ messages in thread
* [Bug c++/15774] Conflicting function decls not diagnosed
2004-06-02 10:54 [Bug c++/15774] New: " dannysmith at users dot sourceforge dot net
2004-06-02 10:59 ` [Bug c++/15774] " dannysmith at users dot sourceforge dot net
@ 2004-06-02 11:37 ` pinskia at gcc dot gnu dot org
2004-06-28 1:25 ` giovannibajo at libero dot it
` (3 subsequent siblings)
5 siblings, 0 replies; 11+ messages in thread
From: pinskia at gcc dot gnu dot org @ 2004-06-02 11:37 UTC (permalink / raw)
To: gcc-bugs
------- Additional Comments From pinskia at gcc dot gnu dot org 2004-06-02 11:37 -------
Confirmed.
--
What |Removed |Added
----------------------------------------------------------------------------
Status|UNCONFIRMED |NEW
Ever Confirmed| |1
Last reconfirmed|0000-00-00 00:00:00 |2004-06-02 11:37:54
date| |
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=15774
^ permalink raw reply [flat|nested] 11+ messages in thread
* [Bug c++/15774] Conflicting function decls not diagnosed
2004-06-02 10:54 [Bug c++/15774] New: " dannysmith at users dot sourceforge dot net
2004-06-02 10:59 ` [Bug c++/15774] " dannysmith at users dot sourceforge dot net
2004-06-02 11:37 ` pinskia at gcc dot gnu dot org
@ 2004-06-28 1:25 ` giovannibajo at libero dot it
2004-06-28 1:43 ` dannysmith at users dot sourceforge dot net
` (2 subsequent siblings)
5 siblings, 0 replies; 11+ messages in thread
From: giovannibajo at libero dot it @ 2004-06-28 1:25 UTC (permalink / raw)
To: gcc-bugs
------- Additional Comments From giovannibajo at libero dot it 2004-06-28 01:04 -------
Danny,
did you test the patch? What was the outcome?
--
What |Removed |Added
----------------------------------------------------------------------------
CC| |giovannibajo at libero dot
| |it
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=15774
^ permalink raw reply [flat|nested] 11+ messages in thread
* [Bug c++/15774] Conflicting function decls not diagnosed
2004-06-02 10:54 [Bug c++/15774] New: " dannysmith at users dot sourceforge dot net
` (2 preceding siblings ...)
2004-06-28 1:25 ` giovannibajo at libero dot it
@ 2004-06-28 1:43 ` dannysmith at users dot sourceforge dot net
2004-06-28 7:50 ` giovannibajo at libero dot it
2004-06-29 3:31 ` jason at redhat dot com
5 siblings, 0 replies; 11+ messages in thread
From: dannysmith at users dot sourceforge dot net @ 2004-06-28 1:43 UTC (permalink / raw)
To: gcc-bugs
------- Additional Comments From dannysmith at users dot sourceforge dot net 2004-06-28 01:42 -------
Yes I did test it and it was wrong, because it appears C++ front end handles
attributes of class member functions differently than global functions.
Attributes of member functions are only honoured when declared within class
definition, which makes sense to me. So for something like this (g++-old-
deja/g++.ext/attrib2.C) my patch would cause an error to be emitted. Should it?
Should C++ require (as does C) that the subsequent definition also have an
explicit attribute?
// { dg-do run { target i?86-*-* } }
// Test that stdcall doesn't prevent us from using op delete.
// Contributed by Jason Merrill <jason@cygnus.com>
struct A {
void operator delete (void *) __attribute__ ((stdcall));
};
void foo()
{
A* ap = new A;
delete ap;
}
void A:: operator delete (void *) { }
int main()
{
A* ap = new A;
delete ap;
}
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=15774
^ permalink raw reply [flat|nested] 11+ messages in thread
* [Bug c++/15774] Conflicting function decls not diagnosed
2004-06-02 10:54 [Bug c++/15774] New: " dannysmith at users dot sourceforge dot net
` (3 preceding siblings ...)
2004-06-28 1:43 ` dannysmith at users dot sourceforge dot net
@ 2004-06-28 7:50 ` giovannibajo at libero dot it
2004-06-29 3:31 ` jason at redhat dot com
5 siblings, 0 replies; 11+ messages in thread
From: giovannibajo at libero dot it @ 2004-06-28 7:50 UTC (permalink / raw)
To: gcc-bugs
------- Additional Comments From giovannibajo at libero dot it 2004-06-28 07:40 -------
I think the idea is that if you *ADD* an attribute in a later redeclaration,
you should get an error (like in your testcase), but if you omit it in the
redeclaration it is ok (the attribute is implicit in that case).
// OK
void foo(void) stdcall;
void foo(void) {}
// OK
void bar(void) stdcall;
void bar(void) stdcall {}
// ERROR
void foobar(void);
void foobar(void) stdcall {}
Jason, would you kindly comment on this? We need to know which way to proceed
with this patch.
--
What |Removed |Added
----------------------------------------------------------------------------
CC| |jason at redhat dot com
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=15774
^ permalink raw reply [flat|nested] 11+ messages in thread
* [Bug c++/15774] Conflicting function decls not diagnosed
2004-06-02 10:54 [Bug c++/15774] New: " dannysmith at users dot sourceforge dot net
` (4 preceding siblings ...)
2004-06-28 7:50 ` giovannibajo at libero dot it
@ 2004-06-29 3:31 ` jason at redhat dot com
5 siblings, 0 replies; 11+ messages in thread
From: jason at redhat dot com @ 2004-06-29 3:31 UTC (permalink / raw)
To: gcc-bugs
------- Additional Comments From jason at redhat dot com 2004-06-29 03:03 -------
Subject: Re: Conflicting function decls not diagnosed
On 28 Jun 2004 07:40:43 -0000, "giovannibajo at libero dot it" <gcc-bugzilla@gcc.gnu.org> wrote:
> I think the idea is that if you *ADD* an attribute in a later
> redeclaration, you should get an error (like in your testcase), but if
> you omit it in the redeclaration it is ok (the attribute is implicit in
> that case).
> Jason, would you kindly comment on this? We need to know which way to proceed
> with this patch.
That makes sense to me as a general rule.
Jason
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=15774
^ permalink raw reply [flat|nested] 11+ messages in thread
end of thread, other threads:[~2011-02-19 21:26 UTC | newest]
Thread overview: 11+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
[not found] <bug-15774-4066@http.gcc.gnu.org/bugzilla/>
2006-06-30 3:10 ` [Bug c++/15774] Conflicting function decls not diagnosed dannysmith at users dot sourceforge dot net
2007-09-19 21:52 ` dannysmith at users dot sourceforge dot net
[not found] <bug-15774-4@http.gcc.gnu.org/bugzilla/>
2010-12-25 10:41 ` ktietz at gcc dot gnu.org
2011-01-13 21:52 ` ktietz at gcc dot gnu.org
2011-02-19 22:06 ` ktietz at gcc dot gnu.org
2004-06-02 10:54 [Bug c++/15774] New: " dannysmith at users dot sourceforge dot net
2004-06-02 10:59 ` [Bug c++/15774] " dannysmith at users dot sourceforge dot net
2004-06-02 11:37 ` pinskia at gcc dot gnu dot org
2004-06-28 1:25 ` giovannibajo at libero dot it
2004-06-28 1:43 ` dannysmith at users dot sourceforge dot net
2004-06-28 7:50 ` giovannibajo at libero dot it
2004-06-29 3:31 ` jason at redhat dot com
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).