Re: c/4939: [cygwin] Failure of test gcc.dg/c99-intconst-1.c

Re: c/4939: [cygwin] Failure of test gcc.dg/c99-intconst-1.c

[neil]

Synopsis: [cygwin] Failure of test gcc.dg/c99-intconst-1.c

State-Changed-From-To: analyzed->closed
State-Changed-By: neil
State-Changed-When: Sun Jun  9 06:09:22 2002
State-Changed-Why:
    I believe this is now fixed.

http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view%20audit-trail&database=gcc&pr=4939

Re: c/4939: [cygwin] Failure of test gcc.dg/c99-intconst-1.c

[jsm28]

The following reply was made to PR c/4939; it has been noted by GNATS.

From: jsm28@gcc.gnu.org
To: David.Billinghurst@riotinto.com, gcc-bugs@gcc.gnu.org,
  gcc-gnats@gcc.gnu.org, gcc-prs@gcc.gnu.org, jsm28@cam.ac.uk,
  nobody@gcc.gnu.org, zack@gcc.gnu.org
Cc:  
Subject: Re: c/4939: [cygwin] Failure of test gcc.dg/c99-intconst-1.c
Date: 24 Nov 2001 11:00:40 -0000

 Synopsis: [cygwin] Failure of test gcc.dg/c99-intconst-1.c
 
 Responsible-Changed-From-To: unassigned->zack
 Responsible-Changed-By: jsm28
 Responsible-Changed-When: Sat Nov 24 03:00:39 2001
 Responsible-Changed-Why:
     The testcase failures are expected until Zack's patch is in.
     The deficiencies in Cygwin's <limits.h> - if GCC needs to
     use that rather than its own functional <limits.h> - are
     for the Cygwin maintainers to deal with, though
     fixincludes perhaps ought to fix the problems on existing
     versions of Cygwin with deficient headers.
 State-Changed-From-To: open->analyzed
 State-Changed-By: jsm28
 State-Changed-When: Sat Nov 24 03:00:39 2001
 State-Changed-Why:
     The testcase fails (depending on architecture).  This is expected.
     Zack asked me to write the testcases; having done so I added them.
     http://gcc.gnu.org/ml/gcc-patches/2001-08/msg01411.html
     http://gcc.gnu.org/ml/gcc-patches/2001-11/msg01370.html
 
 http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view%20audit-trail&pr=4939&database=gcc

Re: c/4939: [cygwin] Failure of test gcc.dg/c99-intconst-1.c

[jsm28]

Synopsis: [cygwin] Failure of test gcc.dg/c99-intconst-1.c

Responsible-Changed-From-To: unassigned->zack
Responsible-Changed-By: jsm28
Responsible-Changed-When: Sat Nov 24 03:00:39 2001
Responsible-Changed-Why:
    The testcase failures are expected until Zack's patch is in.
    The deficiencies in Cygwin's <limits.h> - if GCC needs to
    use that rather than its own functional <limits.h> - are
    for the Cygwin maintainers to deal with, though
    fixincludes perhaps ought to fix the problems on existing
    versions of Cygwin with deficient headers.
State-Changed-From-To: open->analyzed
State-Changed-By: jsm28
State-Changed-When: Sat Nov 24 03:00:39 2001
State-Changed-Why:
    The testcase fails (depending on architecture).  This is expected.
    Zack asked me to write the testcases; having done so I added them.
    http://gcc.gnu.org/ml/gcc-patches/2001-08/msg01411.html
    http://gcc.gnu.org/ml/gcc-patches/2001-11/msg01370.html

http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view%20audit-trail&pr=4939&database=gcc

c/4939: [cygwin] Failure of test gcc.dg/c99-intconst-1.c

[David.Billinghurst]

>Number:         4939
>Category:       c
>Synopsis:       [cygwin] Failure of test gcc.dg/c99-intconst-1.c
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    unassigned
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Fri Nov 23 21:36:00 PST 2001
>Closed-Date:
>Last-Modified:
>Originator:     David Billinghurst
>Release:        gcc-3.1
>Organization:
>Environment:
i686-pc-cygwin
>Description:
New test gcc.dg/c99-inconst-1.c fails on cygwin.  Part of the problem is missing definitions in limits.h.  With these fixed, the test still fails with repeated warnings on lines 237 and 238:

/usr/local/src/gcc/gcc/testsuite/gcc.dg/c99-intconst-1.c:237: warning: decimal constant is so large that it is unsigned

/usr/local/src/gcc/gcc/testsuite/gcc.dg/c99-intconst-1.c:237: assignment from incompatible pointer type

A reduced test case that gives the same error is

/* Origin: Joseph Myers <jsm28@cam.ac.uk>.  */
/* { dg-do compile } */
/* { dg-options "-std=iso9899:1999 -pedantic-errors" } */

#include <limits.h>

/* Assertion that constant C is of type T.  */
#define ASSERT_CONST_TYPE(C, T)			\
	do {					\
	  typedef T type;			\
	  typedef type **typepp;		\
	  typedef __typeof__((C)) ctype;	\
	  typedef ctype **ctypepp;		\
	  typepp x = 0;				\
	  ctypepp y = 0;			\
	  x = y;				\
	  y = x;				\
	} while (0)

/* (T *) if E is zero, (void *) otherwise.  */
#define type_if_not(T, E) __typeof__(0 ? (T *)0 : (void *)(E))

/* (T *) if E is nonzero, (void *) otherwise.  */
#define type_if(T, E) type_if_not(T, !(E))

/* Combine pointer types, all but one (void *).  */
#define type_comb2(T1, T2) __typeof__(0 ? (T1)0 : (T2)0)
#define type_comb3(T1, T2, T3) type_comb2(T1, type_comb2(T2, T3))

/* (T1 *) if E1, otherwise (T2 *) if E2, otherwise (T3 *) if E3.  */
#define first_of3p(T1, E1, T2, E2, T3, E3)			\
	type_comb3(type_if(T1, (E1)),				\
		   type_if(T2, (!(E1) && (E2))),		\
		   type_if(T3, (!(E1) && !(E2) && (E3))))

#define first_of3(T1, E1, T2, E2, T3, E3)				\
	__typeof__(*((first_of3p(T1, (E1), T2, (E2), T3, (E3)))0))

/* Types of constants according to the C99 rules.  */
#define C99_UNSUF_DEC_TYPE(C)				\
	first_of3(int, (C) <= INT_MAX,			\
		  long int, (C) <= LONG_MAX,		\
		  long long int, (C) <= LLONG_MAX)

/* Checks that constants have correct type.  */
#define CHECK_UNSUF_DEC_TYPE(C) ASSERT_CONST_TYPE((C), C99_UNSUF_DEC_TYPE((C)))

void
foo (void)
{
  CHECK_UNSUF_DEC_TYPE(2147483648);
}


The preprocessed source from this is attached:


>How-To-Repeat:

>Fix:

>Release-Note:
>Audit-Trail:
>Unformatted:
----gnatsweb-attachment----
Content-Type: text/plain; name="l.i"
Content-Disposition: inline; filename="l.i"

# 1 "l.c"
# 1 "<builtin>"
# 1 "<command line>"
# 1 "l.c"






# 1 "/usr/local/obj/gcc/gcc/include/limits.h" 1 3
# 11 "/usr/local/obj/gcc/gcc/include/limits.h" 3
# 1 "/usr/local/obj/gcc/gcc/include/syslimits.h" 1 3






# 1 "/usr/local/obj/gcc/gcc/include/limits.h" 1 3
# 132 "/usr/local/obj/gcc/gcc/include/limits.h" 3
# 1 "/usr/local/lib/gcc-lib/i686-pc-cygwin/3.1/include/limits.h" 1 3
# 132 "/usr/local/lib/gcc-lib/i686-pc-cygwin/3.1/include/limits.h" 3
# 1 "/usr/include/limits.h" 1 3
# 133 "/usr/local/lib/gcc-lib/i686-pc-cygwin/3.1/include/limits.h" 2 3
# 133 "/usr/local/obj/gcc/gcc/include/limits.h" 2 3
# 8 "/usr/local/obj/gcc/gcc/include/syslimits.h" 2 3
# 12 "/usr/local/obj/gcc/gcc/include/limits.h" 2 3
# 8 "l.c" 2
# 50 "l.c"
void
foo (void)
{
  int int_max = 2147483647;
  do { typedef __typeof__(*((__typeof__(0 ? (__typeof__(0 ? (int *)0 : (void *)(!(((((2147483648)) <= 2147483647))))))0 : (__typeof__(0 ? (__typeof__(0 ? (long int *)0 : (void *)(!((!((((2147483648)) <= 2147483647)) && ((((2147483648)) <= 2147483647L)))))))0 : (__typeof__(0 ? (long long int *)0 : (void *)(!((!((((2147483648)) <= 2147483647)) && !((((2147483648)) <= 2147483647L)) && ((((2147483648)) <= 9223372036854775807LL)))))))0))0))0)) type; typedef type **typepp; typedef __typeof__(((2147483648))) ctype; typedef ctype **ctypepp; typepp x = 0; ctypepp y = 0; x = y; y = x; } while (0);
}