preprocessor/4902: no macro redef warnings

preprocessor/4902: no macro redef warnings

[Golubev I. N.]

>Number:         4902
>Category:       preprocessor
>Synopsis:       no macro redef warnings
>Confidential:   no
>Severity:       non-critical
>Priority:       medium
>Responsible:    unassigned
>State:          open
>Class:          doc-bug
>Submitter-Id:   net
>Arrival-Date:   Mon Nov 19 10:56:00 PST 2001
>Closed-Date:
>Last-Modified:
>Originator:     Golubev I. N.
>Release:        3.0.2
>Organization:
>Environment:
System: 


host: 
build: 
target: 
configured with:  
>Description:

Neil wrote in <http://gcc.gnu.org/ml/gcc-bugs/2001-11/msg00263.html>:

> you get a warning iff -pedantic

`(cpp)Undefining and Redefining Macros' does not say so.  Neither does
it explain (or refer to explanation) why that change of requiring
`-pedantic' was made (making `gcc' different from many other compilers
which warn by default).

Perhaps some people think that redefn warnings `trigger frequently on
harmless code' (as `-pedantic' description in `(cpp)Invocation' says),
but I disagree.  My experience shows that more often than not they
detect real bugs.

>How-To-Repeat:

#define a 1
#define a 2

I expect: warnings like this:

2:warning: `a' redefined
1:warning: this is the location of the previous definition

I get: no warnings.

>Fix:
	
1. Add more granularity to cpp warning options.  Allow redefn warnings
to be enabled without requesting all `-pedantic' diagnostics.

2. Whether it done or not, document in `(cpp)Undefining and Redefining
Macros' option required to get redefn warnings (or to suppress them,
if they are re-enabled by default).
>Release-Note:
>Audit-Trail:
>Unformatted:

Re: preprocessor/4902: no macro redef warnings

[neil]

The following reply was made to PR preprocessor/4902; it has been noted by GNATS.

From: neil@gcc.gnu.org
To: gcc-bugs@gcc.gnu.org, gcc-gnats@gcc.gnu.org, gcc-prs@gcc.gnu.org,
  gin@mo.msk.ru, nobody@gcc.gnu.org
Cc:  
Subject: Re: preprocessor/4902: no macro redef warnings
Date: 23 Nov 2001 08:45:32 -0000

 Synopsis: no macro redef warnings
 
 State-Changed-From-To: open->closed
 State-Changed-By: neil
 State-Changed-When: Fri Nov 23 00:45:31 2001
 State-Changed-Why:
     Fixed for 3.1; I'll get something in for 3.0.3 too, soon.
 
 http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view%20audit-trail&pr=4902&database=gcc

Re: preprocessor/4902: no macro redef warnings

[neil]

Synopsis: no macro redef warnings

State-Changed-From-To: open->closed
State-Changed-By: neil
State-Changed-When: Fri Nov 23 00:45:31 2001
State-Changed-Why:
    Fixed for 3.1; I'll get something in for 3.0.3 too, soon.

http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view%20audit-trail&pr=4902&database=gcc

Re: preprocessor/4902: no macro redef warnings

[Neil Booth]

The following reply was made to PR preprocessor/4902; it has been noted by GNATS.

From: Neil Booth <neil@daikokuya.demon.co.uk>
To: gcc-patches@gcc.gnu.org
Cc: gcc-gnats@gcc.gnu.org
Subject: Re: preprocessor/4902: no macro redef warnings
Date: Fri, 23 Nov 2001 07:52:50 +0000

 This fixes the complaint of PR 4902 where we would only warn of
 non-trivial macro redefinitions if -pedantic.  Now we always warn,
 which is what we used to do.  It's not clear why this changed.
 
 I'll put a similar thing in the branch soon.
 
 Neil.
 
 	* cppmacro.c (warn_of_redefinition): Warn regardless of -pedantic.
 	* gcc.dg/cpp/mi6.c, gcc.dg/cpp/redef2.c: Update.
 
 Index: cppmacro.c
 ===================================================================
 RCS file: /cvs/gcc/gcc/gcc/cppmacro.c,v
 retrieving revision 1.85
 diff -u -p -r1.85 cppmacro.c
 --- cppmacro.c	2001/11/19 21:04:48	1.85
 +++ cppmacro.c	2001/11/23 07:44:42
 @@ -1124,9 +1124,6 @@ warn_of_redefinition (pfile, node, macro
    if (node->flags & NODE_WARN)
      return 1;
  
 -  if (! CPP_PEDANTIC (pfile))
 -    return 0;
 -
    /* Redefinition of a macro is allowed if and only if the old and new
       definitions are the same.  (6.10.3 paragraph 2).  */
    macro1 = node->value.macro;
 Index: testsuite/gcc.dg/cpp/redef2.c
 ===================================================================
 RCS file: /cvs/gcc/gcc/gcc/testsuite/gcc.dg/cpp/redef2.c,v
 retrieving revision 1.4
 diff -u -p -r1.4 redef2.c
 --- redef2.c	2001/09/30 10:03:11	1.4
 +++ redef2.c	2001/11/23 07:44:42
 @@ -1,7 +1,7 @@
  /* Test for redefining macros with significant differences.  */
  
  /* { dg-do preprocess }
 -   { dg-options "-ansi -pedantic -Wall -fno-show-column" } */
 +   { dg-options "-ansi -Wall -fno-show-column" } */
  
  #define mac(a, b) (a) + (b)
  #define mac(a, b) (a) * (b)
 @@ -28,6 +28,4 @@
     { dg-warning "previous"  "prev def mac"  { target *-*-* } 8  }
     { dg-warning "previous"  "prev def ro"   { target *-*-* } 11 }
     { dg-warning "previous"  "prev def va"   { target *-*-* } 14 }
 -
 -   { dg-warning "named var" "named variadic"  { target *-*-* } 14 }
 -   { dg-warning "anonymous var" "anon variadic" { target *-*-* } 15 } */ 
 +*/
 Index: testsuite/gcc.dg/cpp/mi6.c
 ===================================================================
 RCS file: /cvs/gcc/gcc/gcc/testsuite/gcc.dg/cpp/mi6.c,v
 retrieving revision 1.2
 diff -u -p -r1.2 mi6.c
 --- mi6.c	2001/10/01 10:04:36	1.2
 +++ mi6.c	2001/11/23 07:44:42
 @@ -29,10 +29,15 @@
  /* Define the macro guard, and redefine the macros to something that
     forces compilation of the conditional blocks.  */
  #define guard
 +#undef EMPTYL
  #define EMPTYL 1 ||
 +#undef EMPTYR
  #define EMPTYR || 1
 +#undef NOT
  #define NOT
 +#undef DEFINED
  #define DEFINED 0
 +#undef NOT_DEFINED
  #define NOT_DEFINED 1
  
  #define VAR five

Re: preprocessor/4902: no macro redef warnings

[Neil Booth]

The following reply was made to PR preprocessor/4902; it has been noted by GNATS.

From: Neil Booth <neil@daikokuya.demon.co.uk>
To: Zack Weinberg <zack@codesourcery.com>
Cc: "Golubev I. N." <gin@mo.msk.ru>, gcc-gnats@gcc.gnu.org
Subject: Re: preprocessor/4902: no macro redef warnings
Date: Wed, 21 Nov 2001 18:06:34 +0000

 Zack Weinberg wrote:-
 
 > This warning was on by default in 2.95, and I agree with the reporter
 > that non-trivial redefinitions do tend to represent bugs.
 > warn_of_redefinition is your code, so I must presume you meant to
 > change it - why?
 
 I don't remember; I don't think I did change it.  I suspect the
 semantics were already there.
 
 In fact, I seem to remember us discussing this last year, and you
 having some reason for wanting it -pedantic.  Or was that something
 else?
 
 Anyway, we seem to be agreed.  I'll make it unconditional.
 
 Neil.

Re: preprocessor/4902: no macro redef warnings

[Zack Weinberg]

The following reply was made to PR preprocessor/4902; it has been noted by GNATS.

From: Zack Weinberg <zack@codesourcery.com>
To: Neil Booth <neil@daikokuya.demon.co.uk>
Cc: "Golubev I. N." <gin@mo.msk.ru>, gcc-gnats@gcc.gnu.org
Subject: Re: preprocessor/4902: no macro redef warnings
Date: Tue, 20 Nov 2001 00:08:53 -0800

 On Tue, Nov 20, 2001 at 07:51:43AM +0000, Neil Booth wrote:
 > > Neil wrote in <http://gcc.gnu.org/ml/gcc-bugs/2001-11/msg00263.html>:
 > > 
 > > > you get a warning iff -pedantic
 > > 
 > > `(cpp)Undefining and Redefining Macros' does not say so.  Neither does
 > > it explain (or refer to explanation) why that change of requiring
 > > `-pedantic' was made (making `gcc' different from many other compilers
 > > which warn by default).
 > > 
 > > Perhaps some people think that redefn warnings `trigger frequently on
 > > harmless code' (as `-pedantic' description in `(cpp)Invocation' says),
 > > but I disagree.  My experience shows that more often than not they
 > > detect real bugs.
 > 
 > What do you think, Zack?
 
 This warning was on by default in 2.95, and I agree with the reporter
 that non-trivial redefinitions do tend to represent bugs.
 warn_of_redefinition is your code, so I must presume you meant to
 change it - why?
 
 zw

Re: preprocessor/4902: no macro redef warnings

[Neil Booth]

The following reply was made to PR preprocessor/4902; it has been noted by GNATS.

From: Neil Booth <neil@daikokuya.demon.co.uk>
To: "Golubev I. N." <gin@mo.msk.ru>
Cc: gcc-gnats@gcc.gnu.org, Zack Weinberg <zack@codesourcery.com>
Subject: Re: preprocessor/4902: no macro redef warnings
Date: Tue, 20 Nov 2001 07:51:43 +0000

 > Neil wrote in <http://gcc.gnu.org/ml/gcc-bugs/2001-11/msg00263.html>:
 > 
 > > you get a warning iff -pedantic
 > 
 > `(cpp)Undefining and Redefining Macros' does not say so.  Neither does
 > it explain (or refer to explanation) why that change of requiring
 > `-pedantic' was made (making `gcc' different from many other compilers
 > which warn by default).
 > 
 > Perhaps some people think that redefn warnings `trigger frequently on
 > harmless code' (as `-pedantic' description in `(cpp)Invocation' says),
 > but I disagree.  My experience shows that more often than not they
 > detect real bugs.
 
 What do you think, Zack?
 
 Neil.