Fix isASCII for EBCDIC

Prior to this commit isASCII on EBCDIC platforms was defined as the
isascii() libc function.  It turns out that this doesn't work properly.
It needed to be this way back when EBCDIC was bootstrapped onto the
target machine, but now, various header files are furnished with the
requisite definitions, so this is no longer necessary.

The problem with isascii() is that it is locale-dependent, unlike on
ASCII platforms.  This means that instead of getting a standard ASCII
definition, it returns whatever the underlying locale says, even if
there is no 'use locale' anywhere in the program.  Starting with this
commit, the isASCII definition now comes from the l1_char_class_tab.h
file which we know is accurate and not locale-dependent.

This header can be used in compilations of utility programs where perl.h
is not available.  For these, there are alternate, more complicated
definitions, should they be needed in those utility programs.  Several
of those definitions prior to this commit also used locale-sensitive
isfoo() functions.  The bulk of this commit refactors those definitions
to not use these functions as much as possible.  As noted in the
added comments in the code, the one remaining use of such a function is
only for the lesser-used control characters.  Likely these aren't used
in the utility programs.

1 files changed, 82 insertions, 33 deletions

diff --git a/handy.h b/handy.h
index 8cf4147afa..2fba1a482b 100644
--- a/handy.h
+++ b/handy.h
@@ -893,21 +893,17 @@ patched there.  The file as of this writing is cpan/Devel-PPPort/parts/inc/misc
 
 #ifdef EBCDIC
 #   ifndef _ALL_SOURCE
-        /* This returns the wrong results on at least z/OS unless this is
-         * defined. */
+        /* The native libc isascii() et.al. functions return the wrong results
+         * on at least z/OS unless this is defined. */
 #       error   _ALL_SOURCE should probably be defined
 #   endif
-
-    /* We could be called without perl.h, in which case NATIVE_TO_ASCII() is
-     * likely not defined, and so we use the native function */
-#   define isASCII(c)    cBOOL(isascii(c))
 #else
+    /* There is a simple definition of ASCII for ASCII platforms.  But the
+     * EBCDIC one isn't so simple, so is defined using table look-up like the
+     * other macros below */
 #   define isASCII(c)    ((WIDEST_UTYPE)(c) < 128)
 #endif
 
-#define isASCII_A(c)  isASCII(c)
-#define isASCII_L1(c)  isASCII(c)
-
 /* The lower 3 bits in both the ASCII and EBCDIC representations of '0' are 0,
  * and the 8 possible permutations of those bits exactly comprise the 8 octal
  * digits */
@@ -1088,6 +1084,10 @@ EXTCONST U32 PL_charclass[];
 #   define isWORDCHAR_L1(c) _generic_isCC(c, _CC_WORDCHAR)
 #   define isIDFIRST_L1(c) _generic_isCC(c, _CC_IDFIRST)
 
+#   ifdef EBCDIC
+#       define isASCII(c) _generic_isCC(c, _CC_ASCII)
+#   endif
+
     /* Participates in a single-character fold with a character above 255 */
 #   define _HAS_NONLATIN1_SIMPLE_FOLD_CLOSURE_ONLY_FOR_USE_BY_REGCOMP_DOT_C_AND_REGEXEC_DOT_C(c) ((! cBOOL(FITS_IN_8_BITS(c))) || (PL_charclass[(U8) (c)] & _CC_mask(_CC_NONLATIN1_SIMPLE_FOLD)))
 
@@ -1105,37 +1105,84 @@ EXTCONST U32 PL_charclass[];
 
     /* If we don't have perl.h, we are compiling a utility program.  Below we
      * hard-code various macro definitions that wouldn't otherwise be available
-     * to it. */
-#   ifdef EBCDIC
-        /* Use the native functions.  They likely will return false for all
-         * non-ASCII values, but this makes sure */
-#       define isLOWER_A(c)    (isASCII(c) && islower(c))
-#       define isPRINT_A(c)    (isASCII(c) && isprint(c))
-#       define isUPPER_A(c)    (isASCII(c) && isupper(c))
-#   else   /* ASCII platform.  These are coded based on first principals */
+     * to it. Most are coded based on first principals.  First some ones common
+     * to both ASCII and EBCDIC */
+#   define isDIGIT_A(c)  ((c) <= '9' && (c) >= '0')
+#   define isBLANK_A(c)  ((c) == ' ' || (c) == '\t')
+#   define isSPACE_A(c)  (isBLANK_A(c)                                       \
+                          || (c) == '\n'                                     \
+                          || (c) == '\r'                                     \
+                          || (c) == '\v'                                     \
+                          || (c) == '\f')
+#   ifdef EBCDIC    /* There are gaps between 'i' and 'j'; 'r' and 's'.  Same
+                       for uppercase.  This is ordered to exclude most things
+                       early */
+#       define isLOWER_A(c)  ((c) >= 'a' && (c) <= 'z'                       \
+                               && ((c) <= 'i'                                \
+                                   || ((c) >= 'j' && (c) <= 'r')             \
+                                   || (c) >= 's'))
+#       define isUPPER_A(c)  ((c) >= 'A' && (c) <= 'Z'                       \
+                               && ((c) <= 'I'                                \
+                                   || ((c) >= 'J' && (c) <= 'R')             \
+                                   || (c) >= 'S'))
+#   else   /* ASCII platform. */
 #       define isLOWER_A(c)  ((c) >= 'a' && (c) <= 'z')
-#       define isPRINT_A(c)  (((c) >= 32 && (c) < 127))
 #       define isUPPER_A(c)  ((c) <= 'Z' && (c) >= 'A')
-#   endif   /* Below are common definitions for ASCII and non-ASCII */
+#   endif
+
+    /* Some more ASCII, non-ASCII common definitions */
 #   define isALPHA_A(c)  (isUPPER_A(c) || isLOWER_A(c))
 #   define isALPHANUMERIC_A(c) (isALPHA_A(c) || isDIGIT_A(c))
-#   define isBLANK_A(c)      ((c) == ' ' || (c) == '\t')
-#   define isCNTRL_A(c)  (isASCII(c) && (! isPRINT_A(c)))
-#   define isDIGIT_A(c)  ((c) <= '9' && (c) >= '0')
-#   define isGRAPH_A(c)  (isPRINT_A(c) && (c) != ' ')
-#   define isIDFIRST_A(c)    (isALPHA_A(c) || (c) == '_')
-#   define isPUNCT_A(c)  (isGRAPH_A(c) && (! isALPHANUMERIC_A(c)))
-#   define isSPACE_A(c)  ((c) == ' '                                         \
-                       || (c) == '\t'                                        \
-                       || (c) == '\n'                                        \
-                       || (c) == '\r'                                        \
-                       || (c) == '\v'                                        \
-                       || (c) == '\f')
 #   define isWORDCHAR_A(c)   (isALPHANUMERIC_A(c) || (c) == '_')
-#   define isXDIGIT_A(c) (isDIGIT_A(c)                                       \
-                          || ((c) >= 'a' && (c) <= 'f')                      \
+#   define isIDFIRST_A(c)    (isALPHA_A(c) || (c) == '_')
+#   define isXDIGIT_A(c) (isDIGIT_A(c)                                      \
+                          || ((c) >= 'a' && (c) <= 'f')                     \
                           || ((c) <= 'F' && (c) >= 'A'))
 
+#   ifdef EBCDIC
+#       define isPUNCT_A(c)  ((c) == '-' || (c) == '!' || (c) == '"'        \
+                           || (c) == '#' || (c) == '$' || (c) == '%'        \
+                           || (c) == '&' || (c) == '\'' || (c) == '('       \
+                           || (c) == ')' || (c) == '*' || (c) == '+'        \
+                           || (c) == ',' || (c) == '.' || (c) == '/'        \
+                           || (c) == ':' || (c) == ';' || (c) == '<'        \
+                           || (c) == '=' || (c) == '>' || (c) == '?'        \
+                           || (c) == '@' || (c) == '[' || (c) == '\\'       \
+                           || (c) == ']' || (c) == '^' || (c) == '_'        \
+                           || (c) == '`' || (c) == '{' || (c) == '|'        \
+                           || (c) == '}' || (c) == '~')
+#       define isGRAPH_A(c)  (isALPHANUMERIC_A(c) || isPUNCT_A(c))
+#       define isPRINT_A(c)  (isGRAPH_A(c) || (c) == ' ')
+
+#       ifdef QUESTION_MARK_CTRL
+#           define _isQMC(c) ((c) == QUESTION_MARK_CTRL)
+#       else
+#           define _isQMC(c) 0
+#       endif
+
+        /* I (khw) can't think of a way to define all the ASCII controls
+         * without resorting to a libc (locale-sensitive) call.  But we know
+         * that all controls but the question-mark one are in the range 0-0x3f.
+         * This makes sure that all the controls that have names are included,
+         * and all controls that are also considered ASCII in the locale.  This
+         * may include more or fewer than what it actually should, but the
+         * wrong ones are less-important controls, so likely won't impact
+         * things (keep in mind that this is compiled only if perl.h isn't
+         * available).  The question mark control is included if available */
+#       define isCNTRL_A(c)  (((c) < 0x40 && isascii(c))                    \
+                            || (c) == '\0' || (c) == '\a' || (c) == '\b'    \
+                            || (c) == '\f' || (c) == '\n' || (c) == '\r'    \
+                            || (c) == '\t' || (c) == '\v' || _isQMC(c))
+
+#       define isASCII(c)    (isCNTRL_A(c) || isPRINT_A(c))
+#   else    /* ASCII platform; things are simpler, and  isASCII has already
+               been defined */
+#       define isGRAPH_A(c)  (((c) > ' ' && (c) < 127))
+#       define isPRINT_A(c)  (isGRAPH_A(c) || (c) == ' ')
+#       define isPUNCT_A(c)  (isGRAPH_A(c) && (! isALPHANUMERIC_A(c)))
+#       define isCNTRL_A(c)  (isASCII(c) && (! isPRINT_A(c)))
+#   endif
+
     /* The _L1 macros may be unnecessary for the utilities; I (khw) added them
      * during debugging, and it seems best to keep them.  We may be called
      * without NATIVE_TO_LATIN1 being defined.  On ASCII platforms, it doesn't
@@ -1214,6 +1261,8 @@ EXTCONST U32 PL_charclass[];
 
 #define isALPHANUMERIC(c)  isALPHANUMERIC_A(c)
 #define isALPHA(c)   isALPHA_A(c)
+#define isASCII_A(c)  isASCII(c)
+#define isASCII_L1(c)  isASCII(c)
 #define isBLANK(c)   isBLANK_A(c)
 #define isCNTRL(c)   isCNTRL_A(c)
 #define isDIGIT(c)   isDIGIT_A(c)