#define MUTABLE_IO(p) ((IO *)MUTABLE_PTR(p))
#define MUTABLE_SV(p) ((SV *)MUTABLE_PTR(p))
-#ifdef I_STDBOOL
+#if defined(I_STDBOOL) && !defined(PERL_BOOL_AS_CHAR)
# include <stdbool.h>
# ifndef HAS_BOOL
# define HAS_BOOL 1
Andy Dougherty February 2000
*/
#ifdef __GNUG__ /* GNU g++ has bool built-in */
+# ifndef PERL_BOOL_AS_CHAR
# ifndef HAS_BOOL
# define HAS_BOOL 1
# endif
+# endif
#endif
/* The NeXT dynamic loader headers will not build with the bool macro
#endif /* NeXT || __NeXT__ */
#ifndef HAS_BOOL
-# if defined(VMS)
-# define bool int
-# else
-# define bool char
+# ifdef bool
+# undef bool
# endif
+# define bool char
# define HAS_BOOL 1
#endif
#if (defined(__STDC_VERSION__) && __STDC_VERSION__ >= 199901L) || (defined(__SUNPRO_C)) /* C99 or close enough. */
# define FUNCTION__ __func__
#else
-# if (defined(_MSC_VER) && _MSC_VER < 1300) || /* MSVC6 has neither __func__ nor __FUNCTION and no good workarounds, either. */ \
+# if (defined(USING_MSVC6)) || /* MSVC6 has neither __func__ nor __FUNCTION and no good workarounds, either. */ \
(defined(__DECC_VER)) /* Tru64 or VMS, and strict C89 being used, but not modern enough cc (in Tur64, -c99 not known, only -std1). */
# define FUNCTION__ ""
# else
# define _HIGHEST_REGCOMP_DOT_H_SYNC _CC_VERTSPACE
/* The members of the third group below do not need to be coordinated with data
- * structures in regcomp.[ch] and regexec.c */
+ * structures in regcomp.[ch] and regexec.c. But they should be added to
+ * bootstrap_ctype() */
# define _CC_IDFIRST 17
# define _CC_CHARNAME_CONT 18
# define _CC_NONLATIN1_FOLD 19
# else /* ! DOINIT */
EXTCONST U32 PL_charclass[];
# endif
+#endif /* Has perl.h */
+
+#if defined(H_PERL) && ! defined(BOOTSTRAP_CHARSET)
/* The 1U keeps Solaris from griping when shifting sets the uppermost bit */
# define _CC_mask(classnum) (1U << (classnum))
/* The _A version makes sure that both the desired bit and the ASCII bit
* are present */
-# define _generic_isCC_A(c, classnum) (FITS_IN_8_BITS(c) \
+# define _generic_isCC_A(c, classnum) (FITS_IN_8_BITS(c) \
&& ((PL_charclass[(U8) (c)] & _CC_mask_A(classnum)) \
== _CC_mask_A(classnum)))
# define isUPPER_A(c) _generic_isCC_A(c, _CC_UPPER)
# define isWORDCHAR_A(c) _generic_isCC_A(c, _CC_WORDCHAR)
# define isXDIGIT_A(c) _generic_isCC(c, _CC_XDIGIT)
-# define isIDFIRST_A(c) _generic_isCC_A(c, ( _CC_IDFIRST))
+# define isIDFIRST_A(c) _generic_isCC_A(c, _CC_IDFIRST)
+# define isALPHA_L1(c) _generic_isCC(c, _CC_ALPHA)
+# define isALPHANUMERIC_L1(c) _generic_isCC(c, _CC_ALPHANUMERIC)
+# define isBLANK_L1(c) _generic_isCC(c, _CC_BLANK)
+
+ /* continuation character for legal NAME in \N{NAME} */
+# define isCHARNAME_CONT(c) _generic_isCC(c, _CC_CHARNAME_CONT)
+
+# define isCNTRL_L1(c) _generic_isCC(c, _CC_CNTRL)
+# define isGRAPH_L1(c) _generic_isCC(c, _CC_GRAPH)
+# define isLOWER_L1(c) _generic_isCC(c, _CC_LOWER)
+# define isPRINT_L1(c) _generic_isCC(c, _CC_PRINT)
+# define isPSXSPC_L1(c) _generic_isCC(c, _CC_PSXSPC)
+# define isPUNCT_L1(c) _generic_isCC(c, _CC_PUNCT)
+# define isSPACE_L1(c) _generic_isCC(c, _CC_SPACE)
+# define isUPPER_L1(c) _generic_isCC(c, _CC_UPPER)
+# define isWORDCHAR_L1(c) _generic_isCC(c, _CC_WORDCHAR)
+# define isIDFIRST_L1(c) _generic_isCC(c, _CC_IDFIRST)
/* Either participates in a fold with a character above 255, or is a
* multi-char fold */
_generic_isCC(c, _CC_NON_FINAL_FOLD)
# define _IS_IN_SOME_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) \
_generic_isCC(c, _CC_IS_IN_SOME_FOLD)
-#else /* No perl.h. */
-# define isBLANK_A(c) ((c) == ' ' || (c) == '\t')
-# define isIDFIRST_A(c) (isALPHA_A(c) || (c) == '_')
-# define isWORDCHAR_A(c) (isALPHANUMERIC_A(c) || (c) == '_')
-# define isPSXSPC_A(c) (isSPACE_A(c) || (c) == '\v')
+#else /* Either don't have perl.h or don't want to use char_class_tab.h */
+
+ /* 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. We can also get here if we are configured to bootstrap up Perl
+ * on a non-ASCII platform that doesn't have a working Perl (currently only
+ * EBCDIC). For these we currently use the native definitions to get
+ * things going. (It should also be possible to use the translation
+ * function NATIVE_TO_LATIN1(), but that is an extra layer of dependence on
+ * Perl, so it is currently avoided for the macros where it's possible to
+ * do so.) */
# ifdef EBCDIC
- /* We could be called without perl.h, so the native functions are the
- * easiest to code these in. They likely will return false for all
+ /* Use the native functions. They likely will return false for all
* non-ASCII values, but this makes sure */
# define isALPHA_A(c) (isASCII(c) && isalpha(c))
+# define isALPHA_A(c) (isASCII(c) && isalpha(c))
# define isALPHANUMERIC_A(c) (isASCII(c) && isalnum(c))
# define isCNTRL_A(c) (isASCII(c) && iscntrl(c))
# define isDIGIT_A(c) (isASCII(c) && isdigit(c))
# define isSPACE_A(c) (isASCII(c) && isspace(c))
# define isUPPER_A(c) (isASCII(c) && isupper(c))
# define isXDIGIT_A(c) (isASCII(c) && isxdigit(c))
-# else /* ASCII platform, no perl.h */
+# else /* ASCII platform. These are coded based on first principals */
# define isALPHA_A(c) (isUPPER_A(c) || isLOWER_A(c))
# define isALPHANUMERIC_A(c) (isALPHA_A(c) || isDIGIT_A(c))
-# define isCNTRL_A(c) (isASCII(c) && ! isPRINT_A(c))
+# 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 isGRAPH_A(c) (isPRINT_A(c) && (c) != ' ')
# define isLOWER_A(c) ((c) >= 'a' && (c) <= 'z')
# define isPRINT_A(c) (((c) >= 32 && (c) < 127))
-# define isPUNCT_A(c) (isGRAPH_A(c) && ! isALPHANUMERIC(c))
-# define isSPACE_A(c) ((c) == ' ' \
- || (c) == '\t' \
- || (c) == '\n' \
- || (c) =='\r' \
+# 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 isUPPER_A(c) ((c) <= 'Z' && (c) >= 'A')
-# define isXDIGIT_A(c) (isDIGIT_A(c) \
- || ((c) >= 'a' && (c) <= 'f') \
- || ((c) <= 'F' && (c) >= 'A'))
-# endif
-#endif /* ASCII range definitions */
-
-/* Latin1 definitions */
-#ifdef H_PERL
-# define isALPHA_L1(c) _generic_isCC(c, _CC_ALPHA)
-# define isALPHANUMERIC_L1(c) _generic_isCC(c, _CC_ALPHANUMERIC)
-# define isBLANK_L1(c) _generic_isCC(c, _CC_BLANK)
-
-/* continuation character for legal NAME in \N{NAME} */
-# define isCHARNAME_CONT(c) _generic_isCC(c, _CC_CHARNAME_CONT)
-
-# define isCNTRL_L1(c) _generic_isCC(c, _CC_CNTRL)
-# define isGRAPH_L1(c) _generic_isCC(c, _CC_GRAPH)
-# define isLOWER_L1(c) _generic_isCC(c, _CC_LOWER)
-# define isPRINT_L1(c) _generic_isCC(c, _CC_PRINT)
-# define isPSXSPC_L1(c) _generic_isCC(c, _CC_PSXSPC)
-# define isPUNCT_L1(c) _generic_isCC(c, _CC_PUNCT)
-# define isSPACE_L1(c) _generic_isCC(c, _CC_SPACE)
-# define isUPPER_L1(c) _generic_isCC(c, _CC_UPPER)
-# define isWORDCHAR_L1(c) _generic_isCC(c, _CC_WORDCHAR)
-# define isIDFIRST_L1(c) _generic_isCC(c, _CC_IDFIRST)
-#else /* No access to perl.h. Only a few provided here, just in case needed
- * for backwards compatibility */
- /* ALPHAU includes Unicode semantics for latin1 characters. It has an extra
- * >= AA test to speed up ASCII-only tests at the expense of the others */
-# define isALPHA_L1(c) (isALPHA(c) || (NATIVE_TO_LATIN1((U8) c) >= 0xAA \
- && ((NATIVE_TO_LATIN1((U8) c) >= 0xC0 \
- && NATIVE_TO_LATIN1((U8) c) != 0xD7 && NATIVE_TO_LATIN1((U8) c) != 0xF7) \
- || NATIVE_TO_LATIN1((U8) c) == 0xAA \
- || NATIVE_TO_LATIN1((U8) c) == 0xB5 \
- || NATIVE_TO_LATIN1((U8) c) == 0xBA)))
-# define isCHARNAME_CONT(c) (isWORDCHAR_L1(c) \
- || (c) == ' ' \
- || (c) == '-' \
- || (c) == '(' \
- || (c) == ')' \
- || (c) == ':' \
- || NATIVE_TO_LATIN1((U8) c) == 0xA0)
-#endif
-
-# define isALPHANUMERIC(c) isALPHANUMERIC_A(c)
-# define isALPHA(c) isALPHA_A(c)
-# define isBLANK(c) isBLANK_A(c)
-# define isCNTRL(c) isCNTRL_A(c)
-# define isDIGIT(c) isDIGIT_A(c)
-# define isGRAPH(c) isGRAPH_A(c)
-# define isIDFIRST(c) isIDFIRST_A(c)
-# define isLOWER(c) isLOWER_A(c)
-# define isPRINT(c) isPRINT_A(c)
-# define isPSXSPC(c) isPSXSPC_A(c)
-# define isPUNCT(c) isPUNCT_A(c)
-# define isSPACE(c) isSPACE_A(c)
-# define isUPPER(c) isUPPER_A(c)
-# define isWORDCHAR(c) isWORDCHAR_A(c)
-# define isXDIGIT(c) isXDIGIT_A(c)
-
- /* ASCII casing. These could also be written as
- #define toLOWER(c) (isASCII(c) ? toLOWER_LATIN1(c) : (c))
- #define toUPPER(c) (isASCII(c) ? toUPPER_LATIN1_MOD(c) : (c))
- which uses table lookup and mask instead of subtraction. (This would
- work because the _MOD does not apply in the ASCII range) */
-# define toLOWER(c) (isUPPER(c) ? (c) + ('a' - 'A') : (c))
-# define toUPPER(c) (isLOWER(c) ? (c) - ('a' - 'A') : (c))
+# define isXDIGIT_A(c) (isDIGIT_A(c) \
+ || ((c) >= 'a' && (c) <= 'f') \
+ || ((c) <= 'F' && (c) >= 'A'))
+# endif /* Below are common definitions for ASCII and non-ASCII */
+# define isBLANK_A(c) ((c) == ' ' || (c) == '\t')
+# define isIDFIRST_A(c) (isALPHA_A(c) || (c) == '_')
+# define isWORDCHAR_A(c) (isALPHANUMERIC_A(c) || (c) == '_')
+
+ /* The _L1 macros may be unnecessary for both the utilities and for
+ * bootstrapping; I (khw) added them during debugging of bootstrapping, and
+ * it seems best to keep them. */
+# define isPSXSPC_A(c) isSPACE_A(c) /* XXX Assumes SPACE matches '\v' */
+# define isALPHA_L1(c) (isUPPER_L1(c) || isLOWER_L1(c))
+# define isALPHANUMERIC_L1(c) (isALPHA_L1(c) || isDIGIT_A(c))
+# define isBLANK_L1(c) (isBLANK_A(c) \
+ || (FITS_IN_8_BITS(c) \
+ && NATIVE_TO_LATIN1((U8) c) == 0xA0))
+# define isCNTRL_L1(c) (FITS_IN_8_BITS(c) && (! isPRINT_L1(c)))
+# define isGRAPH_L1(c) (isPRINT_L1(c) && (! isBLANK_L1(c)))
+# define isLOWER_L1(c) (isLOWER_A(c) \
+ || (FITS_IN_8_BITS(c) \
+ && ((NATIVE_TO_LATIN1((U8) c) >= 0xDF \
+ && NATIVE_TO_LATIN1((U8) c) != 0xF7) \
+ || NATIVE_TO_LATIN1((U8) c) == 0xAA \
+ || NATIVE_TO_LATIN1((U8) c) == 0xBA \
+ || NATIVE_TO_LATIN1((U8) c) == 0xB5)))
+# define isPRINT_L1(c) (isPRINT_A(c) \
+ || (FITS_IN_8_BITS(c) \
+ && NATIVE_TO_LATIN1((U8) c) >= 0xA0))
+# define isPSXSPC_L1(c) isSPACE_L1(c)
+# define isPUNCT_L1(c) (isPUNCT_A(c) \
+ || (FITS_IN_8_BITS(c) \
+ && (NATIVE_TO_LATIN1((U8) c) == 0xA1 \
+ || NATIVE_TO_LATIN1((U8) c) == 0xA7 \
+ || NATIVE_TO_LATIN1((U8) c) == 0xAB \
+ || NATIVE_TO_LATIN1((U8) c) == 0xB6 \
+ || NATIVE_TO_LATIN1((U8) c) == 0xB7 \
+ || NATIVE_TO_LATIN1((U8) c) == 0xBB \
+ || NATIVE_TO_LATIN1((U8) c) == 0xBF)))
+# define isSPACE_L1(c) (isSPACE_A(c) \
+ || (FITS_IN_8_BITS(c) \
+ && (NATIVE_TO_LATIN1((U8) c) == 0x85 \
+ || NATIVE_TO_LATIN1((U8) c) == 0xA0)))
+# define isUPPER_L1(c) (isUPPER_A(c) \
+ || (FITS_IN_8_BITS(c) \
+ && (NATIVE_TO_LATIN1((U8) c) >= 0xC0 \
+ && NATIVE_TO_LATIN1((U8) c) <= 0xDE \
+ && NATIVE_TO_LATIN1((U8) c) != 0xD7)))
+# define isWORDCHAR_L1(c) (isIDFIRST_L1(c) || isDIGIT_A(c))
+# define isIDFIRST_L1(c) (isALPHA_L1(c) || NATIVE_TO_LATIN1(c) == '_')
+# define isCHARNAME_CONT(c) (isWORDCHAR_L1(c) \
+ || isBLANK_L1(c) \
+ || (c) == '-' \
+ || (c) == '(' \
+ || (c) == ')')
+ /* The following are not fully accurate in the above-ASCII range. I (khw)
+ * don't think it's necessary to be so for the purposes where this gets
+ * compiled */
+# define _isQUOTEMETA(c) (FITS_IN_8_BITS(c) && ! isWORDCHAR_L1(c))
+# define _IS_IN_SOME_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) isALPHA_L1(c)
+
+ /* And these aren't accurate at all. They are useful only for above
+ * Latin1, which utilities and bootstrapping don't deal with */
+# define _IS_NON_FINAL_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c) 0
+# define _HAS_NONLATIN1_FOLD_CLOSURE_ONLY_FOR_USE_BY_REGCOMP_DOT_C_AND_REGEXEC_DOT_C(c) 0
+
+ /* Many of the macros later in this file are defined in terms of these. By
+ * implementing them with a function, which converts the class number into
+ * a call to the desired macro, all of the later ones work. However, that
+ * function won't be actually defined when building a utility program (no
+ * perl.h), and so a compiler error will be generated if one is attempted
+ * to be used. And the above-Latin1 code points require Unicode tables to
+ * be present, something unlikely to be the case when bootstrapping */
+# define _generic_isCC(c, classnum) \
+ (FITS_IN_8_BITS(c) && S_bootstrap_ctype((U8) (c), (classnum), TRUE))
+# define _generic_isCC_A(c, classnum) \
+ (FITS_IN_8_BITS(c) && S_bootstrap_ctype((U8) (c), (classnum), FALSE))
+#endif /* End of no perl.h or have BOOTSTRAP_CHARSET */
+
+#define isALPHANUMERIC(c) isALPHANUMERIC_A(c)
+#define isALPHA(c) isALPHA_A(c)
+#define isBLANK(c) isBLANK_A(c)
+#define isCNTRL(c) isCNTRL_A(c)
+#define isDIGIT(c) isDIGIT_A(c)
+#define isGRAPH(c) isGRAPH_A(c)
+#define isIDFIRST(c) isIDFIRST_A(c)
+#define isLOWER(c) isLOWER_A(c)
+#define isPRINT(c) isPRINT_A(c)
+#define isPSXSPC(c) isPSXSPC_A(c)
+#define isPUNCT(c) isPUNCT_A(c)
+#define isSPACE(c) isSPACE_A(c)
+#define isUPPER(c) isUPPER_A(c)
+#define isWORDCHAR(c) isWORDCHAR_A(c)
+#define isXDIGIT(c) isXDIGIT_A(c)
+
+/* ASCII casing. These could also be written as
+ #define toLOWER(c) (isASCII(c) ? toLOWER_LATIN1(c) : (c))
+ #define toUPPER(c) (isASCII(c) ? toUPPER_LATIN1_MOD(c) : (c))
+ which uses table lookup and mask instead of subtraction. (This would
+ work because the _MOD does not apply in the ASCII range) */
+#define toLOWER(c) (isUPPER(c) ? (c) + ('a' - 'A') : (c))
+#define toUPPER(c) (isLOWER(c) ? (c) - ('a' - 'A') : (c))
/* In the ASCII range, these are equivalent to what they're here defined to be.
* But by creating these definitions, other code doesn't have to be aware of
_CC_WORDCHAR, c)
#define isXDIGIT_LC_uvchr(c) _generic_LC_uvchr(isXDIGIT_LC, is_XDIGIT_cp_high, c)
-
#define isBLANK_LC_uni(c) isBLANK_LC_uvchr(UNI_TO_NATIVE(c))
/* Everything whose name begins with an underscore is for internal core Perl
/* Like the above, but should be used only when it is known that there are no
* characters in the range 128-255 which the class is TRUE for. Hence it can
* skip the tests for this range. 'above_latin1' should include its arguments */
-#define _generic_utf8_no_upper_latin1(classnum, p, above_latin1) \
+#define _generic_utf8_no_upper_latin1(classnum, p, above_latin1) \
(UTF8_IS_INVARIANT(*(p)) \
? _generic_isCC(*(p), classnum) \
: (UTF8_IS_ABOVE_LATIN1(*(p))) \
- ? above_latin1 \
+ ? above_latin1 \
: 0)
/* NOTE that some of these macros have very similar ones in regcharclass.h.
#define PoisonFree(d,n,t) PoisonWith(d,n,t,0xEF)
#define Poison(d,n,t) PoisonFree(d,n,t)
+#ifdef PERL_POISON
+# define PERL_POISON_EXPR(x) x
+#else
+# define PERL_POISON_EXPR(x)
+#endif
+
#ifdef USE_STRUCT_COPY
#define StructCopy(s,d,t) (*((t*)(d)) = *((t*)(s)))
#else