#define MUTABLE_IO(p) ((IO *)MUTABLE_PTR(p))
#define MUTABLE_SV(p) ((SV *)MUTABLE_PTR(p))
-/* XXX Configure ought to have a test for a boolean type, if I can
- just figure out all the headers such a test needs.
- Andy Dougherty August 1996
-*/
+#ifdef I_STDBOOL
+# include <stdbool.h>
+# ifndef HAS_BOOL
+# define HAS_BOOL 1
+# endif
+#endif
+
/* bool is built-in for g++-2.6.3 and later, which might be used
for extensions. <_G_config.h> defines _G_HAVE_BOOL, but we can't
be sure _G_config.h will be included before this file. _G_config.h
# endif
#endif
-/* HMB H.Merijn Brand - a placeholder for preparing Configure patches:
- *
- * USE_DTRACE HAS_PSEUDOFORK HAS_TIMEGM LOCALTIME_R_NEEDS_TZSET
- * GMTIME_MAX GMTIME_MIN LOCALTIME_MAX LOCALTIME_MIN
- * HAS_CTIME64 HAS_LOCALTIME64 HAS_GMTIME64 HAS_DIFFTIME64
- * HAS_MKTIME64 HAS_ASCTIME64 HAS_GETADDRINFO HAS_GETNAMEINFO
- * HAS_INETNTOP HAS_INETPTON CHARBITS HAS_PRCTL
- * Not (yet) used at top level, but mention them for metaconfig
- */
-
-/* Mention I8SIZE, U8SIZE, I16SIZE, U16SIZE, I32SIZE, U32SIZE,
- I64SIZE, and U64SIZE here so that metaconfig pulls them in. */
-
#if defined(UINT8_MAX) && defined(INT16_MAX) && defined(INT32_MAX)
/* I8_MAX and I8_MIN constants are not defined, as I8 is an ambiguous type.
Like C<newSVpvn_share>, but takes a literal string instead of a string/length
pair and omits the hash parameter.
+=for apidoc Am|void|sv_catpvs_flags|SV* sv|const char* s|I32 flags
+Like C<sv_catpvn_flags>, but takes a literal string instead of a
+string/length pair.
+
+=for apidoc Am|void|sv_catpvs_nomg|SV* sv|const char* s
+Like C<sv_catpvn_nomg>, but takes a literal string instead of a
+string/length pair.
+
=for apidoc Am|void|sv_catpvs|SV* sv|const char* s
Like C<sv_catpvn>, but takes a literal string instead of a string/length pair.
+=for apidoc Am|void|sv_catpvs_mg|SV* sv|const char* s
+Like C<sv_catpvn_mg>, but takes a literal string instead of a
+string/length pair.
+
=for apidoc Am|void|sv_setpvs|SV* sv|const char* s
Like C<sv_setpvn>, but takes a literal string instead of a string/length pair.
+=for apidoc Am|void|sv_setpvs_mg|SV* sv|const char* s
+Like C<sv_setpvn_mg>, but takes a literal string instead of a
+string/length pair.
+
+=for apidoc Am|SV *|sv_setref_pvs|const char* s
+Like C<sv_setref_pvn>, but takes a literal string instead of a
+string/length pair.
+
=head1 Memory Management
=for apidoc Ama|char*|savepvs|const char* s
Like C<savepvn>, but takes a literal string instead of a string/length pair.
+=for apidoc Ama|char*|savesharedpvs|const char* s
+A version of C<savepvs()> which allocates the duplicate string in memory
+which is shared between threads.
+
=head1 GV Functions
=for apidoc Am|HV*|gv_stashpvs|const char* name|I32 create
#define newSVpvs_flags(str,flags) \
Perl_newSVpvn_flags(aTHX_ STR_WITH_LEN(str), flags)
#define newSVpvs_share(str) Perl_newSVpvn_share(aTHX_ STR_WITH_LEN(str), 0)
-#define sv_catpvs(sv, str) Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), SV_GMAGIC)
+#define sv_catpvs_flags(sv, str, flags) \
+ Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), flags)
+#define sv_catpvs_nomg(sv, str) \
+ Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), 0)
+#define sv_catpvs(sv, str) \
+ Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), SV_GMAGIC)
+#define sv_catpvs_mg(sv, str) \
+ Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), SV_GMAGIC|SV_SMAGIC)
#define sv_setpvs(sv, str) Perl_sv_setpvn(aTHX_ sv, STR_WITH_LEN(str))
+#define sv_setpvs_mg(sv, str) Perl_sv_setpvn_mg(aTHX_ sv, STR_WITH_LEN(str))
+#define sv_setref_pvs(rv, classname, str) \
+ Perl_sv_setref_pvn(aTHX_ rv, classname, STR_WITH_LEN(str))
#define savepvs(str) Perl_savepvn(aTHX_ STR_WITH_LEN(str))
-#define gv_stashpvs(str, create) Perl_gv_stashpvn(aTHX_ STR_WITH_LEN(str), create)
-#define gv_fetchpvs(namebeg, add, sv_type) Perl_gv_fetchpvn_flags(aTHX_ STR_WITH_LEN(namebeg), add, sv_type)
+#define savesharedpvs(str) Perl_savesharedpvn(aTHX_ STR_WITH_LEN(str))
+#define gv_stashpvs(str, create) \
+ Perl_gv_stashpvn(aTHX_ STR_WITH_LEN(str), create)
+#define gv_fetchpvs(namebeg, add, sv_type) \
+ Perl_gv_fetchpvn_flags(aTHX_ STR_WITH_LEN(namebeg), add, sv_type)
+#define gv_fetchpvn(namebeg, len, add, sv_type) \
+ Perl_gv_fetchpvn_flags(aTHX_ namebeg, len, add, sv_type)
+#define sv_catxmlpvs(dsv, str, utf8) \
+ Perl_sv_catxmlpvn(aTHX_ dsv, STR_WITH_LEN(str), utf8)
#define hv_fetchs(hv,key,lval) \
((SV **)Perl_hv_common(aTHX_ (hv), NULL, STR_WITH_LEN(key), 0, \
(lval) ? (HV_FETCH_JUST_SV | HV_FETCH_LVALUE) \
/*
=head1 Character classes
-The functions in this section operate using the character set of the platform
-Perl is running on, and are unaffected by locale. For ASCII platforms, they
-will all return false for characters outside the ASCII range. For EBCDIC
-platforms, they use the code page of the platform. The code pages that Perl
-knows about all have 8-bit characters, so most of these functions will return
-true for more characters than on ASCII platforms.
+There are three variants for all the functions in this section. The base ones
+operate using the character set of the platform Perl is running on. The ones
+with an C<_A> suffix operate on the ASCII character set, and the ones with an
+C<_L1> suffix operate on the full Latin1 character set. All are unaffected by
+locale and by C<use bytes>.
+
+For ASCII platforms, the base function with no suffix and the one with the
+C<_A> suffix are identical. The function with the C<_L1> suffix imposes the
+Latin-1 character set onto the platform. That is, the code points that are
+ASCII are unaffected, since ASCII is a subset of Latin-1. But the non-ASCII
+code points are treated as if they are Latin-1 characters. For example,
+C<isSPACE_L1()> will return true when called with the code point 0xA0, which is
+the Latin-1 NO-BREAK SPACE.
+
+For EBCDIC platforms, the base function with no suffix and the one with the
+C<_L1> suffix should be identical, since, as of this writing, the EBCDIC code
+pages that Perl knows about all are equivalent to Latin-1. The function that
+ends in an C<_A> suffix will not return true unless the specified character also
+has an ASCII equivalent.
=for apidoc Am|bool|isALPHA|char ch
Returns a boolean indicating whether the specified character is an
alphabetic character in the platform's native character set.
+See the L<top of this section|/Character classes> for an explanation of variants
+C<isALPHA_A> and C<isALPHA_L1>.
+
+=for apidoc Am|bool|isASCII|char ch
+Returns a boolean indicating whether the specified character is one of the 128
+characters in the ASCII character set. On non-ASCII platforms, it is if this
+character corresponds to an ASCII character. Variants C<isASCII_A()> and
+C<isASCII_L1()> are identical to C<isASCII()>.
=for apidoc Am|bool|isDIGIT|char ch
Returns a boolean indicating whether the specified character is a
digit in the platform's native character set.
+Variants C<isDIGIT_A> and C<isDIGIT_L1> are identical to C<isDIGIT>.
=for apidoc Am|bool|isLOWER|char ch
Returns a boolean indicating whether the specified character is a
lowercase character in the platform's native character set.
+See the L<top of this section|/Character classes> for an explanation of variants
+C<isLOWER_A> and C<isLOWER_L1>.
=for apidoc Am|bool|isOCTAL|char ch
Returns a boolean indicating whether the specified character is an
octal digit, [0-7] in the platform's native character set.
+Variants C<isOCTAL_A> and C<isOCTAL_L1> are identical to C<isOCTAL>.
=for apidoc Am|bool|isSPACE|char ch
Returns a boolean indicating whether the specified character is a
-whitespace character in the platform's native character set.
+whitespace character in the platform's native character set. This is the same
+as what C<\s> matches in a regular expression.
+See the L<top of this section|/Character classes> for an explanation of variants
+C<isSPACE_A> and C<isSPACE_L1>.
=for apidoc Am|bool|isUPPER|char ch
Returns a boolean indicating whether the specified character is an
uppercase character in the platform's native character set.
+See the L<top of this section|/Character classes> for an explanation of variants
+C<isUPPER_A> and C<isUPPER_L1>.
=for apidoc Am|bool|isWORDCHAR|char ch
Returns a boolean indicating whether the specified character is a
C<isALNUM()> is a synonym provided for backward compatibility. Note that it
does not have the standard C language meaning of alphanumeric, since it matches
an underscore and the standard meaning does not.
+See the L<top of this section|/Character classes> for an explanation of variants
+C<isWORDCHAR_A> and C<isWORDCHAR_L1>.
+
+=for apidoc Am|bool|isXDIGIT|char ch
+Returns a boolean indicating whether the specified character is a hexadecimal
+digit, [0-9A-Fa-f]. Variants C<isXDIGIT_A()> and C<isXDIGIT_L1()> are
+identical to C<isXDIGIT()>.
=head1 Character case changing
=cut
-NOTE: Since some of these are macros, there is no check in those that the
-parameter is a char or U8. This means that if called with a larger width
-parameter, casts can silently truncate and yield wrong results.
-
-Also note that these macros are repeated in Devel::PPPort, so should also be
+Note that these macros are repeated in Devel::PPPort, so should also be
patched there. The file as of this writing is cpan/Devel-PPPort/parts/inc/misc
*/
-/* FITS_IN_8_BITS(c) returns true if c occupies no more than 8 bits. It is
- * designed to be hopefully bomb-proof, making sure that no bits of
- * information are lost even on a 64-bit machine, but to get the compiler to
- * optimize it out if possible. This is because Configure makes sure that the
- * machine has an 8-bit byte, so if c is stored in a byte, the sizeof()
- * guarantees that this evaluates to a constant true at compile time. The use
- * of the mask instead of '< 256' keeps gcc from complaining that it is alway
- * true, when c's storage class is a byte. Use U64TYPE because U64 is known
- * only in the perl core, and this macro can be called from outside that */
+/* Specify the widest unsigned type on the platform. Use U64TYPE because U64
+ * is known only in the perl core, and this macro can be called from outside
+ * that */
#ifdef HAS_QUAD
-# define FITS_IN_8_BITS(c) ((sizeof(c) == 1) || (((U64TYPE)(c) & 0xFF) == (U64TYPE)(c)))
+# define WIDEST_UTYPE U64TYPE
#else
-# define FITS_IN_8_BITS(c) ((sizeof(c) == 1) || (((U32)(c) & 0xFF) == (U32)(c)))
+# define WIDEST_UTYPE U32
#endif
-#define isASCII(c) (FITS_IN_8_BITS(c) ? NATIVE_TO_UNI((U8) c) <= 127 : 0)
-#define isASCII_A(c) isASCII(c)
+/* FITS_IN_8_BITS(c) returns true if c doesn't have a bit set other than in
+ * the lower 8. It is designed to be hopefully bomb-proof, making sure that no
+ * bits of information are lost even on a 64-bit machine, but to get the
+ * compiler to optimize it out if possible. This is because Configure makes
+ * sure that the machine has an 8-bit byte, so if c is stored in a byte, the
+ * sizeof() guarantees that this evaluates to a constant true at compile time.
+ */
+#define FITS_IN_8_BITS(c) ((sizeof(c) == 1) || !(((WIDEST_UTYPE)(c)) & ~0xFF))
-/* include the Latin1 lookup table */
-#include "l1_char_class_tab.h"
+#ifdef EBCDIC
+# define isASCII(c) (FITS_IN_8_BITS(c) && (NATIVE_TO_UNI((U8) (c)) < 128))
+#else
+# define isASCII(c) ((WIDEST_UTYPE)(c) < 128)
+#endif
+
+#define isASCII_A(c) isASCII(c)
+#define isASCII_L1(c) isASCII(c)
/* ASCII range only */
-#define isALNUMC_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_ALNUMC_A))
-#define isALPHA_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_ALPHA_A))
-#define isBLANK_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_BLANK_A))
-#define isCNTRL_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_CNTRL_A))
-#define isDIGIT_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_DIGIT_A))
-#define isGRAPH_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_GRAPH_A))
-#define isLOWER_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_LOWER_A))
-#define isPRINT_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_PRINT_A))
-#define isPSXSPC_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_PSXSPC_A))
-#define isPUNCT_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_PUNCT_A))
-#define isSPACE_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_SPACE_A))
-#define isUPPER_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_UPPER_A))
-#define isWORDCHAR_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_WORDCHAR_A))
-#define isXDIGIT_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_XDIGIT_A))
+#ifdef H_PERL /* If have access to perl.h, lookup in its table */
+/* Bits for PL_charclass[]. These use names used in l1_char_class_tab.h but
+ * their actual definitions are here. If that has a name not used here, it
+ * won't compile. */
+# define _CC_ALNUMC_A (1<<0)
+# define _CC_ALNUMC_L1 (1<<1)
+# define _CC_ALPHA_A (1<<2)
+# define _CC_ALPHA_L1 (1<<3)
+# define _CC_BLANK_A (1<<4)
+# define _CC_BLANK_L1 (1<<5)
+# define _CC_CHARNAME_CONT (1<<6)
+# define _CC_CNTRL_A (1<<7)
+# define _CC_CNTRL_L1 (1<<8)
+# define _CC_DIGIT_A (1<<9)
+# define _CC_GRAPH_A (1<<10)
+# define _CC_GRAPH_L1 (1<<11)
+# define _CC_IDFIRST_A (1<<12)
+# define _CC_IDFIRST_L1 (1<<13)
+# define _CC_LOWER_A (1<<14)
+# define _CC_LOWER_L1 (1<<15)
+# define _CC_OCTAL_A (1<<16)
+# define _CC_PRINT_A (1<<17)
+# define _CC_PRINT_L1 (1<<18)
+# define _CC_PSXSPC_A (1<<19)
+# define _CC_PSXSPC_L1 (1<<20)
+# define _CC_PUNCT_A (1<<21)
+# define _CC_PUNCT_L1 (1<<22)
+# define _CC_SPACE_A (1<<23)
+# define _CC_SPACE_L1 (1<<24)
+# define _CC_UPPER_A (1<<25)
+# define _CC_UPPER_L1 (1<<26)
+# define _CC_WORDCHAR_A (1<<27)
+# define _CC_WORDCHAR_L1 (1<<28)
+# define _CC_XDIGIT_A (1<<29)
+# define _CC_NONLATIN1_FOLD (1<<30)
+/* Unused
+ * (1<<31)
+ * If more are needed, can give up some of the above. The first ones to go
+ * would be those that require just two tests to verify, either there are two
+ * code points, like BLANK_A, or occupy a single range like OCTAL_A, DIGIT_A,
+ * UPPER_A, and LOWER_A.
+ */
+
+# ifdef DOINIT
+EXTCONST U32 PL_charclass[] = {
+# include "l1_char_class_tab.h"
+};
+
+# else /* ! DOINIT */
+EXTCONST U32 PL_charclass[];
+# endif
+
+# define isALNUMC_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_ALNUMC_A))
+# define isALPHA_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_ALPHA_A))
+# define isBLANK_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_BLANK_A))
+# define isCNTRL_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_CNTRL_A))
+# define isDIGIT_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_DIGIT_A))
+# define isGRAPH_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_GRAPH_A))
+# define isIDFIRST_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_IDFIRST_A))
+# define isLOWER_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_LOWER_A))
+# define isOCTAL_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_OCTAL_A))
+# define isPRINT_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_PRINT_A))
+# define isPSXSPC_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_PSXSPC_A))
+# define isPUNCT_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_PUNCT_A))
+# define isSPACE_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_SPACE_A))
+# define isUPPER_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_UPPER_A))
+# define isWORDCHAR_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_WORDCHAR_A))
+# define isXDIGIT_A(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_XDIGIT_A))
+ /* Either participates in a fold with a character above 255, or is a
+ * multi-char fold */
+# define _HAS_NONLATIN1_FOLD_CLOSURE_ONLY_FOR_USE_BY_REGCOMP_DOT_C_AND_REGEXEC_DOT_C(c) ((! cBOOL(FITS_IN_8_BITS(c))) || (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_NONLATIN1_FOLD))
+#else /* No perl.h. */
+# define isOCTAL_A(c) ((c) <= '7' && (c) >= '0')
+# ifdef EBCDIC
+# define isALNUMC_A(c) (isASCII(c) && isALNUMC(c))
+# define isALPHA_A(c) (isASCII(c) && isALPHA(c))
+# define isBLANK_A(c) (isASCII(c) && isBLANK(c))
+# define isCNTRL_A(c) (isASCII(c) && isCNTRL(c))
+# define isDIGIT_A(c) (isASCII(c) && isDIGIT(c))
+# define isGRAPH_A(c) (isASCII(c) && isGRAPH(c))
+# define isIDFIRST_A(c) (isASCII(c) && isIDFIRST(c))
+# define isLOWER_A(c) (isASCII(c) && isLOWER(c))
+# define isPRINT_A(c) (isASCII(c) && isPRINT(c))
+# define isPSXSPC_A(c) (isASCII(c) && isPSXSPC(c))
+# define isPUNCT_A(c) (isASCII(c) && isPUNCT(c))
+# define isSPACE_A(c) (isASCII(c) && isSPACE(c))
+# define isUPPER_A(c) (isASCII(c) && isUPPER(c))
+# define isWORDCHAR_A(c) (isASCII(c) && isWORDCHAR(c))
+# define isXDIGIT_A(c) (isASCII(c) && isXDIGIT(c))
+# else /* ASCII platform, no perl.h */
+# define isALNUMC_A(c) (isALPHA_A(c) || isDIGIT_A(c))
+# define isALPHA_A(c) (isUPPER_A(c) || isLOWER_A(c))
+# define isBLANK_A(c) ((c) == ' ' || (c) == '\t')
+# define isCNTRL_A(c) (FITS_IN_8_BITS(c) && ((U8) (c) < ' ' || (c) == 127))
+# define isDIGIT_A(c) ((c) <= '9' && (c) >= '0')
+# define isGRAPH_A(c) (isWORDCHAR_A(c) || isPUNCT_A(c))
+# define isIDFIRST_A(c) (isALPHA_A(c) || (c) == '_')
+# define isLOWER_A(c) ((c) >= 'a' && (c) <= 'z')
+# define isPRINT_A(c) (((c) >= 32 && (c) < 127))
+# define isPSXSPC_A(c) (isSPACE_A(c) || (c) == '\v')
+# define isPUNCT_A(c) (((c) >= 33 && (c) <= 47) || ((c) >= 58 && (c) <= 64) || ((c) >= 91 && (c) <= 96) || ((c) >= 123 && (c) <= 126))
+# define isSPACE_A(c) ((c) == ' ' || (c) == '\t' || (c) == '\n' || (c) =='\r' || (c) == '\f')
+# define isUPPER_A(c) ((c) <= 'Z' && (c) >= 'A')
+# define isWORDCHAR_A(c) (isALPHA_A(c) || isDIGIT_A(c) || (c) == '_')
+# define isXDIGIT_A(c) (isDIGIT_A(c) || ((c) >= 'a' && (c) <= 'f') || ((c) <= 'F' && (c) >= 'A'))
+# endif
+#endif /* ASCII range definitions */
/* Latin1 definitions */
-/* 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 */
-/* XXX decide whether to document the ALPHAU, ALNUMU and isSPACE_L1 functions.
- * Most of these should be implemented as table lookup for speed */
-#define isALPHAU(c) (isALPHA_A(c) || (NATIVE_TO_UNI((U8) c) >= 0xAA \
- && ((NATIVE_TO_UNI((U8) c) >= 0xC0 \
- && NATIVE_TO_UNI((U8) c) != 0xD7 && NATIVE_TO_UNI((U8) c) != 0xF7) \
- || NATIVE_TO_UNI((U8) c) == 0xAA \
- || NATIVE_TO_UNI((U8) c) == 0xB5 \
- || NATIVE_TO_UNI((U8) c) == 0xBA)))
-#define isSPACE_L1(c) (isSPACE(c) \
- || (NATIVE_TO_UNI(c) == 0x85 || NATIVE_TO_UNI(c) == 0xA0))
-#define isWORDCHAR_L1(c) (isDIGIT(c) || isALPHAU(c) || (c) == '_')
-
-/* Same macro in non-EBCDIC and EBCDIC. Called macros may evaluate
- * differently between the two */
+#ifdef H_PERL
+# define isALNUMC_L1(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_ALNUMC_L1))
+# define isALPHA_L1(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_ALPHA_L1))
+# define isBLANK_L1(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_BLANK_L1))
+/* continuation character for legal NAME in \N{NAME} */
+# define isCHARNAME_CONT(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_CHARNAME_CONT))
+# define isCNTRL_L1(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_CNTRL_L1))
+# define isGRAPH_L1(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_GRAPH_L1))
+# define isIDFIRST_L1(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_IDFIRST_L1))
+# define isLOWER_L1(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_LOWER_L1))
+# define isPRINT_L1(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_PRINT_L1))
+# define isPSXSPC_L1(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_PSXSPC_L1))
+# define isPUNCT_L1(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_PUNCT_L1))
+# define isSPACE_L1(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_SPACE_L1))
+# define isUPPER_L1(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_UPPER_L1))
+# define isWORDCHAR_L1(c) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & _CC_WORDCHAR_L1))
+#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_UNI((U8) c) >= 0xAA \
+ && ((NATIVE_TO_UNI((U8) c) >= 0xC0 \
+ && NATIVE_TO_UNI((U8) c) != 0xD7 && NATIVE_TO_UNI((U8) c) != 0xF7) \
+ || NATIVE_TO_UNI((U8) c) == 0xAA \
+ || NATIVE_TO_UNI((U8) c) == 0xB5 \
+ || NATIVE_TO_UNI((U8) c) == 0xBA)))
+# define isCHARNAME_CONT(c) (isALNUM_L1(c) || (c) == ' ' || (c) == '-' || (c) == '(' || (c) == ')' || (c) == ':' || NATIVE_TO_UNI((U8) c) == 0xA0)
+#endif
+
+/* Macros for backwards compatibility and for completeness when the ASCII and
+ * Latin1 values are identical */
#define isALNUM(c) isWORDCHAR(c)
#define isALNUMU(c) isWORDCHAR_L1(c)
-/* continuation character for legal NAME in \N{NAME} */
-#define isCHARNAME_CONT(c) (isWORDCHAR_L1(c) || (c) == ' ' || (c) == '-' || (c) == '(' || (c) == ')' || (c) == ':' || NATIVE_TO_UNI((U8) c) == 0xA0)
-#define isIDFIRST(c) (isALPHA(c) || (c) == '_')
-#define isOCTAL_A(c) ((c) >= '0' && (c) <= '7')
-#define isOCTAL(c) isOCTAL_A(c)
-#define isWORDCHAR(c) (isALPHA(c) || isDIGIT(c) || (c) == '_')
-
+#define isALPHAU(c) isALPHA_L1(c)
+#define isDIGIT_L1(c) isDIGIT_A(c)
+#define isOCTAL(c) isOCTAL_A(c)
+#define isOCTAL_L1(c) isOCTAL_A(c)
+#define isXDIGIT_L1(c) isXDIGIT_A(c)
+
+/* Macros that differ between EBCDIC and ASCII. Where C89 defines a function,
+ * that is used in the EBCDIC form, because in EBCDIC we do not do locales:
+ * therefore can use native functions. For those where C89 doesn't define a
+ * function, use our function, assuming that the EBCDIC code page is isomorphic
+ * with Latin1, which the three currently recognized by Perl are. Some libc's
+ * have an isblank(), but it's not guaranteed. */
#ifdef EBCDIC
- /* In EBCDIC we do not do locales: therefore can use native functions */
# define isALNUMC(c) isalnum(c)
# define isALPHA(c) isalpha(c)
# define isBLANK(c) ((c) == ' ' || (c) == '\t' || NATIVE_TO_UNI(c) == 0xA0)
# define isCNTRL(c) iscntrl(c)
# define isDIGIT(c) isdigit(c)
# define isGRAPH(c) isgraph(c)
+# define isIDFIRST(c) (isALPHA(c) || (c) == '_')
# define isLOWER(c) islower(c)
# define isPRINT(c) isprint(c)
# define isPSXSPC(c) isspace(c)
# define isSPACE(c) (isPSXSPC(c) && (c) != '\v')
# define isUPPER(c) isupper(c)
# define isXDIGIT(c) isxdigit(c)
+# define isWORDCHAR(c) (isalnum(c) || (c) == '_')
# define toLOWER(c) tolower(c)
# define toUPPER(c) toupper(c)
#else /* Not EBCDIC: ASCII-only matching */
# 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. */
+ /* 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))
#endif
#define isPSXSPC_LC(c) (isSPACE_LC(c) || (c) == '\v')
#define isBLANK_LC(c) isBLANK(c) /* could be wrong */
-#define isALNUM_uni(c) is_uni_alnum(c)
-#define isIDFIRST_uni(c) is_uni_idfirst(c)
-#define isALPHA_uni(c) is_uni_alpha(c)
-#define isSPACE_uni(c) is_uni_space(c)
-#define isDIGIT_uni(c) is_uni_digit(c)
-#define isUPPER_uni(c) is_uni_upper(c)
-#define isLOWER_uni(c) is_uni_lower(c)
-#define isASCII_uni(c) is_uni_ascii(c)
-#define isCNTRL_uni(c) is_uni_cntrl(c)
-#define isGRAPH_uni(c) is_uni_graph(c)
-#define isPRINT_uni(c) is_uni_print(c)
-#define isPUNCT_uni(c) is_uni_punct(c)
-#define isXDIGIT_uni(c) is_uni_xdigit(c)
+/* For use in the macros just below. If the input is Latin1, use the Latin1
+ * (_L1) version of the macro; otherwise use the function. Won't compile if
+ * 'c' isn't unsigned, as won't match function prototype. The macros do bounds
+ * checking, so have duplicate checks here, so could create versions of the
+ * macros that don't, but experiments show that gcc optimizes them out anyway.
+ */
+#define generic_uni(macro, function, c) ((c) < 256 \
+ ? CAT2(macro, _L1)(c) \
+ : function(c))
+/* Note that all ignore 'use bytes' */
+
+#define isALNUM_uni(c) generic_uni(isWORDCHAR, is_uni_alnum, c)
+#define isIDFIRST_uni(c) generic_uni(isIDFIRST, is_uni_idfirst, c)
+#define isALPHA_uni(c) generic_uni(isALPHA, is_uni_alpha, c)
+#define isSPACE_uni(c) generic_uni(isSPACE, is_uni_space, c)
+#define isDIGIT_uni(c) generic_uni(isDIGIT, is_uni_digit, c)
+#define isUPPER_uni(c) generic_uni(isUPPER, is_uni_upper, c)
+#define isLOWER_uni(c) generic_uni(isLOWER, is_uni_lower, c)
+#define isASCII_uni(c) isASCII(c)
+/* All controls are in Latin1 */
+#define isCNTRL_uni(c) ((c) < 256 && isCNTRL_L1(c))
+#define isGRAPH_uni(c) generic_uni(isGRAPH, is_uni_graph, c)
+#define isPRINT_uni(c) generic_uni(isPRINT, is_uni_print, c)
+#define isPUNCT_uni(c) generic_uni(isPUNCT, is_uni_punct, c)
+#define isXDIGIT_uni(c) generic_uni(isXDIGIT, is_uni_xdigit, c)
#define toUPPER_uni(c,s,l) to_uni_upper(c,s,l)
#define toTITLE_uni(c,s,l) to_uni_title(c,s,l)
#define toLOWER_uni(c,s,l) to_uni_lower(c,s,l)
#define toFOLD_uni(c,s,l) to_uni_fold(c,s,l)
-#define isPSXSPC_uni(c) (isSPACE_uni(c) ||(c) == '\f')
+/* Posix and regular space differ only in U+000B, which is in Latin1 */
+#define isPSXSPC_uni(c) ((c) < 256 ? isPSXSPC_L1(c) : isSPACE_uni(c))
#define isBLANK_uni(c) isBLANK(c) /* could be wrong */
#define isALNUM_LC_uvchr(c) (c < 256 ? isALNUM_LC(c) : is_uni_alnum_lc(c))
#define isPSXSPC_LC_uni(c) (isSPACE_LC_uni(c) ||(c) == '\f')
#define isBLANK_LC_uni(c) isBLANK(c) /* could be wrong */
-#define isALNUM_utf8(p) is_utf8_alnum(p)
-/* The ID_Start of Unicode is quite limiting: it assumes a L-class
- * character (meaning that you cannot have, say, a CJK character).
- * Instead, let's allow ID_Continue but not digits. */
-#define isIDFIRST_utf8(p) (is_utf8_idcont(p) && !is_utf8_digit(p))
-#define isALPHA_utf8(p) is_utf8_alpha(p)
-#define isSPACE_utf8(p) is_utf8_space(p)
-#define isDIGIT_utf8(p) is_utf8_digit(p)
-#define isUPPER_utf8(p) is_utf8_upper(p)
-#define isLOWER_utf8(p) is_utf8_lower(p)
-#define isASCII_utf8(p) is_utf8_ascii(p)
-#define isCNTRL_utf8(p) is_utf8_cntrl(p)
-#define isGRAPH_utf8(p) is_utf8_graph(p)
-#define isPRINT_utf8(p) is_utf8_print(p)
-#define isPUNCT_utf8(p) is_utf8_punct(p)
-#define isXDIGIT_utf8(p) is_utf8_xdigit(p)
+/* For use in the macros just below. If the input is ASCII, use the ASCII (_A)
+ * version of the macro; if the input is in the upper Latin1 range, use the
+ * Latin1 (_L1) version of the macro, after converting from utf8; otherwise use
+ * the function. This relies on the fact that ASCII characters have the same
+ * representation whether utf8 or not */
+#define generic_utf8(macro, function, p) (isASCII(*(p)) \
+ ? CAT2(CAT2(macro,_),A)(*(p)) \
+ : (UTF8_IS_DOWNGRADEABLE_START(*(p))) \
+ ? CAT2(macro, _L1) \
+ (TWO_BYTE_UTF8_TO_UNI(*(p), \
+ *((p)+1))) \
+ : function(p))
+
+/* Note that all ignore 'use bytes' */
+
+#define isALNUM_utf8(p) generic_utf8(isWORDCHAR, is_utf8_alnum, p)
+/* To prevent S_scan_word in toke.c from hanging, we have to make sure that
+ * IDFIRST is an alnum. See
+ * http://rt.perl.org/rt3/Ticket/Display.html?id=74022 for more detail than you
+ * ever wanted to know about. XXX It is unclear if this should extend to
+ * isIDFIRST_uni() which it hasn't so far. (In the ASCII range, there isn't a
+ * difference.) This used to be not the XID version, but we decided to go with
+ * the more modern Unicode definition */
+#define isIDFIRST_utf8(p) (isASCII(*(p)) \
+ ? isIDFIRST_A(*(p)) \
+ : (UTF8_IS_DOWNGRADEABLE_START(*(p))) \
+ ? isIDFIRST_L1(TWO_BYTE_UTF8_TO_UNI(*(p), \
+ *((p)+1)))\
+ : Perl__is_utf8__perl_idstart(aTHX_ p))
+#define isIDCONT_utf8(p) generic_utf8(isWORDCHAR, is_utf8_xidcont, p)
+#define isALPHA_utf8(p) generic_utf8(isALPHA, is_utf8_alpha, p)
+#define isSPACE_utf8(p) generic_utf8(isSPACE, is_utf8_space, p)
+#define isDIGIT_utf8(p) generic_utf8(isDIGIT, is_utf8_digit, p)
+#define isUPPER_utf8(p) generic_utf8(isUPPER, is_utf8_upper, p)
+#define isLOWER_utf8(p) generic_utf8(isLOWER, is_utf8_lower, p)
+/* Because ASCII is invariant under utf8, the non-utf8 macro works */
+#define isASCII_utf8(p) isASCII(p)
+#define isCNTRL_utf8(p) generic_utf8(isCNTRL, is_utf8_cntrl, p)
+#define isGRAPH_utf8(p) generic_utf8(isGRAPH, is_utf8_graph, p)
+#define isPRINT_utf8(p) generic_utf8(isPRINT, is_utf8_print, p)
+#define isPUNCT_utf8(p) generic_utf8(isPUNCT, is_utf8_punct, p)
+#define isXDIGIT_utf8(p) generic_utf8(isXDIGIT, is_utf8_xdigit, p)
#define toUPPER_utf8(p,s,l) to_utf8_upper(p,s,l)
#define toTITLE_utf8(p,s,l) to_utf8_title(p,s,l)
#define toLOWER_utf8(p,s,l) to_utf8_lower(p,s,l)
-#define isPSXSPC_utf8(c) (isSPACE_utf8(c) ||(c) == '\f')
+/* Posix and regular space differ only in U+000B, which is in ASCII (and hence
+ * Latin1 */
+#define isPSXSPC_utf8(p) ((isASCII(*(p))) \
+ ? isPSXSPC_A(*(p)) \
+ : (UTF8_IS_DOWNGRADEABLE_START(*(p)) \
+ ? isPSXSPC_L1(TWO_BYTE_UTF8_TO_UNI(*(p), \
+ *((p)+1)))\
+ : isSPACE_utf8(p)))
#define isBLANK_utf8(c) isBLANK(c) /* could be wrong */
#define isALNUM_LC_utf8(p) isALNUM_LC_uvchr(utf8_to_uvchr(p, 0))
In 5.9.3, Newx() and friends replace the older New() API, and drops
the first parameter, I<x>, a debug aid which allowed callers to identify
themselves. This aid has been superseded by a new build option,
-PERL_MEM_LOG (see L<perlhack/PERL_MEM_LOG>). The older API is still
+PERL_MEM_LOG (see L<perlhacktips/PERL_MEM_LOG>). The older API is still
there for use in XS modules supporting older perls.
=for apidoc Am|void|Newxc|void* ptr|int nitems|type|cast