perl5.001 patch.1c

Configure
    Updated to dist PL53.
    Fix overlapping memcpy test.
    Add check for ld.  Use $cc instead, if on an ELF system.
    With -d, don't reuse config.sh unless $myuname matches.
    Warn more explicitly about changing compilers before reloading
        old config.sh.
    Detect Linux ELF format in nm scan.
    Better detection of d_castneg.  (ISC 4.1 was passing the test,
        but couldn't cast in an argument list.)
    Suggest -fpic for dynamic loading if you're using GNU CC under any 
        name.
    No longer test for byacc, fmod, or drem, since they are not used.

Makefile.SH
    Use $ld, not ld (only matters for SVR4)
    Silence some byacc-related harmless error messages.

README
    Suggest using -Dcc=gcc (or whatever).
    Warn about reusing old config.sh.  (The warning was already there
        in 5.001; I've just expanded it a little.)
    Warn against using GNU as and GNU ld on SunOS & Solaris.
   
config.H
config_h.SH
    Updated to match Configure.

doio.c
    Add socket includes.

ext/Fcntl/Fcntl.xs
    Fix typo:  s/SETFL/F_SETFL/;

handy.h
    Check _G_HAVE_BOOL, not just if it's defined.

hints/dynix.sh
hints/hpux_9.sh
hints/linux.sh
hints/netbsd.sh
hints/titanos.sh
    Updated.  ELF on linux should probably work.

installperl
    Install pod2html, pod2latex, and pod2man.

lib/ExtUtils/MakeMaker.pm
    Updated to 4.091.  (4.09 + a small writedoc() patch.)

myconfig
    Now includes 'ld' command.

perl.c
    Revised an #elif clause since Pyramid's cpp doesn't understand
    #elif.

perl.h
    Fix U_L, I_V, and I_32 cast macros to ensure that
    the cast_ulong(), cast_iv(), and cast_i32() functions (if used)
    are passed a double.  In particular, the FIXSTATUS macros were
    handing int's to U_L().
    Remove unnecessary HAS_FMOD testing (See util.c).

proto.h
    Remove my_fmod() prototype.  (See util.c)

sv.h
    Fix GV/CV typo.

util.c
    Simplified cast_i32() and cast_iv() to mimic what *actually*
    happens on a SPARC running SunOS 4.1.3.  (Previously, they did
    some complicated fmod() calculation.  I've since discovered that's
    not what happens on the SPARC.)

    With this change, fmod() is no longer necessary.  Hence my_fmod
    is removed.  This also means the HAS_FMOD and HAS_DREM tests are
    no longer needed in Configure, so they are gone too.

vms/config.vms
    Remove unnecessary HAS_FMOD and HAS_DREM defines.

x2p/Makefile.SH
    Silence byacc-related things.

1 files changed, 6 insertions, 31 deletions

diff --git a/util.c b/util.c
index f2bf077666..4a45c5ef99 100644
--- a/util.c
+++ b/util.c
@@ -1579,12 +1579,12 @@ I32
 cast_i32(f)
 double f;
 {
-#   define BIGDOUBLE 2147483648.0        /* Assume 32 bit int's ! */
+#   define BIGDOUBLE 2147483647.0        /* Assume 32 bit int's ! */
 #   define BIGNEGDOUBLE (-2147483648.0)
     if (f >= BIGDOUBLE)
-	return (I32)fmod(f, BIGDOUBLE);
+	return (I32) BIGDOUBLE;
     if (f <= BIGNEGDOUBLE)
-	return (I32)fmod(f, BIGNEGDOUBLE);
+	return (I32) BIGNEGDOUBLE;
     return (I32) f;
 }
 # undef BIGDOUBLE
@@ -1595,12 +1595,12 @@ cast_iv(f)
 double f;
 {
     /* XXX  This should be fixed.  It assumes 32 bit IV's. */
-#   define BIGDOUBLE 2147483648.0        /* Assume 32 bit IV's ! */
+#   define BIGDOUBLE 2147483647.0        /* Assume 32 bit IV's ! */
 #   define BIGNEGDOUBLE (-2147483648.0)
     if (f >= BIGDOUBLE)
-	return (IV)fmod(f, BIGDOUBLE);
+	return (IV) BIGDOUBLE;
     if (f <= BIGNEGDOUBLE)
-	return (IV)fmod(f, BIGNEGDOUBLE);
+	return (IV) BIGNEGDOUBLE;
     return (IV) f;
 }
 # undef BIGDOUBLE
@@ -1687,28 +1687,3 @@ I32 *retlen;
     *retlen = s - start;
     return retval;
 }
-
-/* Amazingly enough, some systems (e.g. Dynix 3) don't have fmod.
-   This is a slow, stupid, but working emulation.  (AD)
-*/
-#ifdef USE_MY_FMOD
-double
-my_fmod(x, y)
-double x, y;
-{
-    double i = 0.0;   /* Can't use int because it can overflow */
-    if ((x == 0) || (y == 0))
-       return 0;
-    /* The sign of fmod is the same as the sign of x.  */
-    if ( (x < 0 && y > 0) || (x > 0 && y < 0) )
-	y = -y;
-    if (x > 0) {
-	while (x - i*y > y)
-	    i++;
-    } else {
-	while (x - i*y < y)
-	    i++;
-    }
-    return x - i * y;
-}
-#endif