c/6968: functions that shouldn't accept arguments accept infinite arguments (eg; test() )

c/6968: functions that shouldn't accept arguments accept infinite arguments (eg; test() )

[tmorelan]

>Number:         6968
>Category:       c
>Synopsis:       functions that shouldn't accept arguments accept infinite arguments (eg; test() )
>Confidential:   no
>Severity:       non-critical
>Priority:       low
>Responsible:    unassigned
>State:          open
>Class:          accepts-illegal
>Submitter-Id:   net
>Arrival-Date:   Fri Jun 07 18:16:00 PDT 2002
>Closed-Date:
>Last-Modified:
>Originator:     Terry Moreland
>Release:        2.95, 2.96, 3.1
>Organization:
>Environment:
Debian Linux (Potato)
Redhat Linux 7.2 (base install) glibc 2.2.2-10
Redhat Linux 7.3 (base install) glibc 2.2.5
>Description:
a function declared as taking no arguments accepts infinite arguments without complaint under -Wall -ansi -pedantic or any other standard 

eg;
void test( ) {
  printf("shouldn't compile\n");
}

int main(void) {
  test( 1, 2, 3, "Hello", "World", "this", "shouldn't", "compile" );
  return(0);
}

above code compiles and runs without error
>How-To-Repeat:
sample code:

#include <stdio.h>

void test( ) {
  printf("shouldn't compile\n" );
}

int main(void) {
  test( 1, 2, 3, "Hello", "World" );
  return(0);
}
>Fix:
change prototype of function that shouldn't accept arguments
to use void as argument

eg;
void test(void);
>Release-Note:
>Audit-Trail:
>Unformatted:

Re: c/6968: functions that shouldn't accept arguments accept infinite arguments (eg; test() )

[Terry Moreland]

The following reply was made to PR c/6968; it has been noted by GNATS.

From: Terry Moreland <tmorelan@q.cis.uoguelph.ca>
To: "Joseph S. Myers" <jsm28@cam.ac.uk>
Cc: zack@gcc.gnu.org, <gcc-bugs@gcc.gnu.org>, <gcc-gnats@gcc.gnu.org>
Subject: Re: c/6968: functions that shouldn't accept arguments accept infinite
 arguments (eg; test() )
Date: Sat, 8 Jun 2002 15:29:18 -0400 (EDT)

 ok, I'll accept that, but still passing arguments to a function defined in this
 way should still produce a warning since the arguments are inaccessible from
 within the function and it just looks wrong
 
 Terry
 
 On Sat, 8 Jun 2002, Joseph S. Myers wrote:
 
 > On Sat, 8 Jun 2002, Terry Moreland wrote:
 > 
 > > mean that in c99 void test() is a function that takes no arguments since 
 > > enmpty identifier list is part of the function definition and as such the 
 > > void test() example should produce an error in c99, but doesn't in gcc 3.1 using
 > > -std=c99 
 > > 
 > > please let me know if I am wrong, I have slept in a while and the iso c99 spec
 > > can be very wordy at times
 > 
 > It specifies, for the purposes of the definition, that it has no 
 > parameters.  It does not, for the purposes of subsequent calls to the 
 > function, give it a type that includes a prototype, so the subsequent 
 > calls are not checked.
 > 
 > 
 
 -- 
 --------------------------------------------------------
 Terry Moreland       |  Computer and Information Science
 tmorelan@uoguelph.ca |              University of Guelph
 
 If you were plowing a field, which would you rather use?  
 Two strong oxen or 1024 chickens?
     - Seymour Cray (1925-1996), father of supercomputing
 --------------------------------------------------------
 

Re: c/6968: functions that shouldn't accept arguments accept infinite arguments (eg; test() )

[Joseph S. Myers]

The following reply was made to PR c/6968; it has been noted by GNATS.

From: "Joseph S. Myers" <jsm28@cam.ac.uk>
To: Terry Moreland <tmorelan@q.cis.uoguelph.ca>
Cc: <zack@gcc.gnu.org>,  <gcc-bugs@gcc.gnu.org>,  <gcc-gnats@gcc.gnu.org>
Subject: Re: c/6968: functions that shouldn't accept arguments accept infinite
 arguments (eg; test() )
Date: Sat, 8 Jun 2002 10:41:30 +0100 (BST)

 On Sat, 8 Jun 2002, Terry Moreland wrote:
 
 > mean that in c99 void test() is a function that takes no arguments since 
 > enmpty identifier list is part of the function definition and as such the 
 > void test() example should produce an error in c99, but doesn't in gcc 3.1 using
 > -std=c99 
 > 
 > please let me know if I am wrong, I have slept in a while and the iso c99 spec
 > can be very wordy at times
 
 It specifies, for the purposes of the definition, that it has no 
 parameters.  It does not, for the purposes of subsequent calls to the 
 function, give it a type that includes a prototype, so the subsequent 
 calls are not checked.
 
 -- 
 Joseph S. Myers
 jsm28@cam.ac.uk
 

Re: c/6968: functions that shouldn't accept arguments accept infinite arguments (eg; test() )

[Terry Moreland]

The following reply was made to PR c/6968; it has been noted by GNATS.

From: Terry Moreland <tmorelan@q.cis.uoguelph.ca>
To: zack@gcc.gnu.org, <gcc-bugs@gcc.gnu.org>, <gcc-prs@gcc.gnu.org>,
   <nobody@gcc.gnu.org>, <gcc-gnats@gcc.gnu.org>
Cc:  
Subject: Re: c/6968: functions that shouldn't accept arguments accept infinite
 arguments (eg; test() )
Date: Sat, 8 Jun 2002 02:44:36 -0400 (EDT)

 I might be reading this wrong but does not:
 
 An identifier list declares only the identifiers of the parameters of the function. An empty
 list in a function declarator that is part of a definition of that function specifies that the
 function has no parameters. The empty list in a function declarator that is not part of a
 definition of that function specifies that no information about the number or types of the
 parameters is supplied.124)
 
 ( ISO 9899:1999 Section 6.7.5.3 Paragraph 14 )
 
 mean that in c99 void test() is a function that takes no arguments since 
 enmpty identifier list is part of the function definition and as such the 
 void test() example should produce an error in c99, but doesn't in gcc 3.1 using
 -std=c99 
 
 please let me know if I am wrong, I have slept in a while and the iso c99 spec
 can be very wordy at times
 
 the reference 124 leads to 6.11.6 which mentions:
 The use of function declarators with empty parentheses (not prototype-format parameter
 type declarators) is an obsolescent feature.
 
 Terry
 
 On 8 Jun 2002 zack@gcc.gnu.org wrote:
 
 > Synopsis: functions that shouldn't accept arguments accept infinite arguments (eg; test() )
 > 
 > State-Changed-From-To: open->closed
 > State-Changed-By: zack
 > State-Changed-When: Fri Jun  7 22:56:53 2002
 > State-Changed-Why:
 >     This is not a bug.  The C standard specifies that a
 >     function declared with an empty parameter list (such
 >     as your void test ( ) ) accepts an indefinite number
 >     of arguments, not zero arguments.  You must write
 >     void test ( void ) if you want to declare a function
 >     accepting zero arguments.
 >     
 >     (You may have C confused with C++.  In C++ the rule
 >     is as you expected.)
 > 
 > http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view%20audit-trail&database=gcc&pr=6968
 > 
 
 -- 
 --------------------------------------------------------
 Terry Moreland       |  Computer and Information Science
 tmorelan@uoguelph.ca |              University of Guelph
 
 If you were plowing a field, which would you rather use?  
 Two strong oxen or 1024 chickens?
     - Seymour Cray (1925-1996), father of supercomputing
 --------------------------------------------------------
 

Re: c/6968: functions that shouldn't accept arguments accept infinite arguments (eg; test() )

[zack]

Synopsis: functions that shouldn't accept arguments accept infinite arguments (eg; test() )

State-Changed-From-To: open->closed
State-Changed-By: zack
State-Changed-When: Fri Jun  7 22:56:53 2002
State-Changed-Why:
    This is not a bug.  The C standard specifies that a
    function declared with an empty parameter list (such
    as your void test ( ) ) accepts an indefinite number
    of arguments, not zero arguments.  You must write
    void test ( void ) if you want to declare a function
    accepting zero arguments.
    
    (You may have C confused with C++.  In C++ the rule
    is as you expected.)

http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view%20audit-trail&database=gcc&pr=6968