/*
=for apidoc_section $SV
-=for apidoc Am|void *|MUTABLE_PTR|void * p
-=for apidoc_item |AV *|MUTABLE_AV|AV * p
-=for apidoc_item |CV *|MUTABLE_CV|CV * p
-=for apidoc_item |GV *|MUTABLE_GV|GV * p
-=for apidoc_item |HV *|MUTABLE_HV|HV * p
-=for apidoc_item |IO *|MUTABLE_IO|IO * p
-=for apidoc_item |SV *|MUTABLE_SV|SV * p
+=for apidoc Am |AV * |MUTABLE_AV |AV * p
+=for apidoc_item |CV * |MUTABLE_CV |CV * p
+=for apidoc_item |GV * |MUTABLE_GV |GV * p
+=for apidoc_item |HV * |MUTABLE_HV |HV * p
+=for apidoc_item |IO * |MUTABLE_IO |IO * p
+=for apidoc_item |void *|MUTABLE_PTR|void * p
+=for apidoc_item |SV * |MUTABLE_SV |SV * p
The C<MUTABLE_I<*>>() macros cast pointers to the types shown, in such a way
(compiler permitting) that casting away const-ness will give a warning;
already-built-in ones return pointers to what their names indicate.
=cut
- */
+The brace group version will raise a diagnostic if 'p' is const; the other
+blindly casts away const.
+ */
#if defined(PERL_USE_GCC_BRACE_GROUPS)
# define MUTABLE_PTR(p) ({ void *p_ = (p); p_; })
#else
=cut
*/
-# ifdef I_STDINT
+#ifdef I_STDINT
typedef int_fast8_t PERL_INT_FAST8_T;
typedef uint_fast8_t PERL_UINT_FAST8_T;
typedef int_fast16_t PERL_INT_FAST16_T;
typedef uint_fast16_t PERL_UINT_FAST16_T;
-# else
+#else
typedef int PERL_INT_FAST8_T;
typedef unsigned int PERL_UINT_FAST8_T;
typedef int PERL_INT_FAST16_T;
typedef unsigned int PERL_UINT_FAST16_T;
-# endif
+#endif
/* log(2) (i.e., log base 10 of 2) is pretty close to 0.30103, just in case
* anyone is grepping for it. So BIT_DIGITS gives the number of decimal digits
/*
=for apidoc_section $versioning
=for apidoc AmR|bool|PERL_VERSION_EQ|const U8 major|const U8 minor|const U8 patch
-=for apidoc_item PERL_VERSION_NE
-=for apidoc_item PERL_VERSION_LT
-=for apidoc_item PERL_VERSION_LE
-=for apidoc_item PERL_VERSION_GT
=for apidoc_item PERL_VERSION_GE
+=for apidoc_item PERL_VERSION_GT
+=for apidoc_item PERL_VERSION_LE
+=for apidoc_item PERL_VERSION_LT
+=for apidoc_item PERL_VERSION_NE
Returns whether or not the perl currently being compiled has the specified
relationship to the perl given by the parameters. For example,
* it comes to /\w+/ with tainting enabled, we *must* be able
* to trust our character classes.
*
- * Therefore, the default tests in the text of Perl will be
- * independent of locale. Any code that wants to depend on
- * the current locale will use the tests that begin with "lc".
+ * Therefore, the default tests in the text of Perl will be independent of
+ * locale. Any code that wants to depend on the current locale will use the
+ * macros that contain _LC in their names
*/
-#ifdef USE_LOCALE
+#ifdef USE_LOCALE_CTYPE
# ifndef CTYPE256
# define CTYPE256
# endif
=for apidoc Am|bool|isALPHA|UV ch
=for apidoc_item ||isALPHA_A|UV ch
-=for apidoc_item ||isALPHA_L1|UV ch
-=for apidoc_item ||isALPHA_uvchr|UV ch
-=for apidoc_item ||isALPHA_utf8_safe|U8 * s|U8 * end
-=for apidoc_item ||isALPHA_utf8|U8 * s|U8 * end
=for apidoc_item ||isALPHA_LC|UV ch
-=for apidoc_item ||isALPHA_LC_uvchr|UV ch
=for apidoc_item ||isALPHA_LC_utf8_safe|U8 * s| U8 *end
+=for apidoc_item ||isALPHA_LC_uvchr|UV ch
+=for apidoc_item ||isALPHA_L1|UV ch
+=for apidoc_item ||isALPHA_utf8|U8 * s|U8 * end
+=for apidoc_item ||isALPHA_utf8_safe|U8 * s|U8 * end
+=for apidoc_item ||isALPHA_uvchr|UV ch
Returns a boolean indicating whether the specified input is one of C<[A-Za-z]>,
analogous to C<m/[[:alpha:]]/>.
See the L<top of this section|/Character classification> for an explanation of
=for apidoc Am|bool|isALPHANUMERIC|UV ch
=for apidoc_item ||isALPHANUMERIC_A|UV ch
-=for apidoc_item ||isALPHANUMERIC_L1|UV ch
-=for apidoc_item ||isALPHANUMERIC_uvchr|UV ch
-=for apidoc_item ||isALPHANUMERIC_utf8_safe|U8 * s|U8 * end
-=for apidoc_item ||isALPHANUMERIC_utf8|U8 * s|U8 * end
=for apidoc_item ||isALPHANUMERIC_LC|UV ch
-=for apidoc_item ||isALPHANUMERIC_LC_uvchr|UV ch
=for apidoc_item ||isALPHANUMERIC_LC_utf8_safe|U8 * s| U8 *end
-=for apidoc_item ||isALNUMC|UV ch
-=for apidoc_item ||isALNUMC_A|UV ch
-=for apidoc_item ||isALNUMC_L1|UV ch
-=for apidoc_item ||isALNUMC_LC|UV ch
-=for apidoc_item ||isALNUMC_LC_uvchr|UV ch
+=for apidoc_item ||isALPHANUMERIC_LC_uvchr|UV ch
+=for apidoc_item ||isALPHANUMERIC_L1|UV ch
+=for apidoc_item ||isALPHANUMERIC_utf8|U8 * s|U8 * end
+=for apidoc_item ||isALPHANUMERIC_utf8_safe|U8 * s|U8 * end
+=for apidoc_item ||isALPHANUMERIC_uvchr|UV ch
Returns a boolean indicating whether the specified character is one of
C<[A-Za-z0-9]>, analogous to C<m/[[:alnum:]]/>.
See the L<top of this section|/Character classification> for an explanation of
the variants.
-A (discouraged from use) synonym is C<isALNUMC> (where the C<C> suffix means
-this corresponds to the C language alphanumeric definition). Also
-there are the variants
-C<isALNUMC_A>, C<isALNUMC_L1>
-C<isALNUMC_LC>, and C<isALNUMC_LC_uvchr>.
+=for apidoc Am|bool|isALNUMC|UV ch
+=for apidoc_item ||isALNUMC_A|UV ch
+=for apidoc_item ||isALNUMC_LC|UV ch
+=for apidoc_item ||isALNUMC_LC_uvchr|UV ch
+=for apidoc_item ||isALNUMC_L1|UV ch
+These are discouraged, backward compatibility macros for L</C<isALPHANUMERIC>>.
+That is, each returns a boolean indicating whether the specified character is
+one of C<[A-Za-z0-9]>, analogous to C<m/[[:alnum:]]/>.
+
+The C<C> suffix in the names was meant to indicate that they correspond to the
+C language L<C<isalnum(3)>>.
=for apidoc Am|bool|isASCII|UV ch
=for apidoc_item ||isASCII_A|UV ch
-=for apidoc_item ||isASCII_L1|UV ch
-=for apidoc_item ||isASCII_uvchr|UV ch
-=for apidoc_item ||isASCII_utf8_safe|U8 * s|U8 * end
-=for apidoc_item ||isASCII_utf8|U8 * s|U8 * end
=for apidoc_item ||isASCII_LC|UV ch
-=for apidoc_item ||isASCII_LC_uvchr|UV ch
=for apidoc_item ||isASCII_LC_utf8_safe|U8 * s| U8 *end
+=for apidoc_item ||isASCII_LC_uvchr|UV ch
+=for apidoc_item ||isASCII_L1|UV ch
+=for apidoc_item ||isASCII_utf8|U8 * s|U8 * end
+=for apidoc_item ||isASCII_utf8_safe|U8 * s|U8 * end
+=for apidoc_item ||isASCII_uvchr|UV ch
Returns a boolean indicating whether the specified character is one of the 128
characters in the ASCII character set, analogous to C<m/[[:ascii:]]/>.
On non-ASCII platforms, it returns TRUE iff this
=for apidoc Am|bool|isBLANK|UV ch
=for apidoc_item ||isBLANK_A|UV ch
-=for apidoc_item ||isBLANK_L1|UV ch
-=for apidoc_item ||isBLANK_uvchr|UV ch
-=for apidoc_item ||isBLANK_utf8_safe|U8 * s|U8 * end
-=for apidoc_item ||isBLANK_utf8|U8 * s|U8 * end
=for apidoc_item ||isBLANK_LC|UV ch
-=for apidoc_item ||isBLANK_LC_uvchr|UV ch
=for apidoc_item ||isBLANK_LC_utf8_safe|U8 * s| U8 *end
+=for apidoc_item ||isBLANK_LC_uvchr|UV ch
+=for apidoc_item ||isBLANK_L1|UV ch
+=for apidoc_item ||isBLANK_utf8|U8 * s|U8 * end
+=for apidoc_item ||isBLANK_utf8_safe|U8 * s|U8 * end
+=for apidoc_item ||isBLANK_uvchr|UV ch
Returns a boolean indicating whether the specified character is a
character considered to be a blank, analogous to C<m/[[:blank:]]/>.
See the L<top of this section|/Character classification> for an explanation of
=for apidoc Am|bool|isCNTRL|UV ch
=for apidoc_item ||isCNTRL_A|UV ch
-=for apidoc_item ||isCNTRL_L1|UV ch
-=for apidoc_item ||isCNTRL_uvchr|UV ch
-=for apidoc_item ||isCNTRL_utf8_safe|U8 * s|U8 * end
-=for apidoc_item ||isCNTRL_utf8|U8 * s|U8 * end
=for apidoc_item ||isCNTRL_LC|UV ch
-=for apidoc_item ||isCNTRL_LC_uvchr|UV ch
=for apidoc_item ||isCNTRL_LC_utf8_safe|U8 * s| U8 *end
+=for apidoc_item ||isCNTRL_LC_uvchr|UV ch
+=for apidoc_item ||isCNTRL_L1|UV ch
+=for apidoc_item ||isCNTRL_utf8|U8 * s|U8 * end
+=for apidoc_item ||isCNTRL_utf8_safe|U8 * s|U8 * end
+=for apidoc_item ||isCNTRL_uvchr|UV ch
Returns a boolean indicating whether the specified character is a
control character, analogous to C<m/[[:cntrl:]]/>.
=for apidoc Am|bool|isDIGIT|UV ch
=for apidoc_item ||isDIGIT_A|UV ch
-=for apidoc_item ||isDIGIT_L1|UV ch
-=for apidoc_item ||isDIGIT_uvchr|UV ch
-=for apidoc_item ||isDIGIT_utf8_safe|U8 * s|U8 * end
-=for apidoc_item ||isDIGIT_utf8|U8 * s|U8 * end
=for apidoc_item ||isDIGIT_LC|UV ch
-=for apidoc_item ||isDIGIT_LC_uvchr|UV ch
=for apidoc_item ||isDIGIT_LC_utf8_safe|U8 * s| U8 *end
+=for apidoc_item ||isDIGIT_LC_uvchr|UV ch
+=for apidoc_item ||isDIGIT_L1|UV ch
+=for apidoc_item ||isDIGIT_utf8|U8 * s|U8 * end
+=for apidoc_item ||isDIGIT_utf8_safe|U8 * s|U8 * end
+=for apidoc_item ||isDIGIT_uvchr|UV ch
Returns a boolean indicating whether the specified character is a
digit, analogous to C<m/[[:digit:]]/>.
=for apidoc Am|bool|isGRAPH|UV ch
=for apidoc_item ||isGRAPH_A|UV ch
-=for apidoc_item ||isGRAPH_L1|UV ch
-=for apidoc_item ||isGRAPH_uvchr|UV ch
-=for apidoc_item ||isGRAPH_utf8_safe|U8 * s|U8 * end
-=for apidoc_item ||isGRAPH_utf8|U8 * s|U8 * end
=for apidoc_item ||isGRAPH_LC|UV ch
-=for apidoc_item ||isGRAPH_LC_uvchr|UV ch
=for apidoc_item ||isGRAPH_LC_utf8_safe|U8 * s| U8 *end
+=for apidoc_item ||isGRAPH_LC_uvchr|UV ch
+=for apidoc_item ||isGRAPH_L1|UV ch
+=for apidoc_item ||isGRAPH_utf8|U8 * s|U8 * end
+=for apidoc_item ||isGRAPH_utf8_safe|U8 * s|U8 * end
+=for apidoc_item ||isGRAPH_uvchr|UV ch
Returns a boolean indicating whether the specified character is a
graphic character, analogous to C<m/[[:graph:]]/>.
See the L<top of this section|/Character classification> for an explanation of
=for apidoc Am|bool|isLOWER|UV ch
=for apidoc_item ||isLOWER_A|UV ch
-=for apidoc_item ||isLOWER_L1|UV ch
-=for apidoc_item ||isLOWER_uvchr|UV ch
-=for apidoc_item ||isLOWER_utf8_safe|U8 * s|U8 * end
-=for apidoc_item ||isLOWER_utf8|U8 * s|U8 * end
=for apidoc_item ||isLOWER_LC|UV ch
-=for apidoc_item ||isLOWER_LC_uvchr|UV ch
=for apidoc_item ||isLOWER_LC_utf8_safe|U8 * s| U8 *end
+=for apidoc_item ||isLOWER_LC_uvchr|UV ch
+=for apidoc_item ||isLOWER_L1|UV ch
+=for apidoc_item ||isLOWER_utf8|U8 * s|U8 * end
+=for apidoc_item ||isLOWER_utf8_safe|U8 * s|U8 * end
+=for apidoc_item ||isLOWER_uvchr|UV ch
Returns a boolean indicating whether the specified character is a
lowercase character, analogous to C<m/[[:lower:]]/>.
See the L<top of this section|/Character classification> for an explanation of
=for apidoc Am|bool|isPUNCT|UV ch
=for apidoc_item ||isPUNCT_A|UV ch
-=for apidoc_item ||isPUNCT_L1|UV ch
-=for apidoc_item ||isPUNCT_uvchr|UV ch
-=for apidoc_item ||isPUNCT_utf8_safe|U8 * s|U8 * end
-=for apidoc_item ||isPUNCT_utf8|U8 * s|U8 * end
=for apidoc_item ||isPUNCT_LC|UV ch
-=for apidoc_item ||isPUNCT_LC_uvchr|UV ch
=for apidoc_item ||isPUNCT_LC_utf8_safe|U8 * s| U8 *end
+=for apidoc_item ||isPUNCT_LC_uvchr|UV ch
+=for apidoc_item ||isPUNCT_L1|UV ch
+=for apidoc_item ||isPUNCT_utf8|U8 * s|U8 * end
+=for apidoc_item ||isPUNCT_utf8_safe|U8 * s|U8 * end
+=for apidoc_item ||isPUNCT_uvchr|UV ch
Returns a boolean indicating whether the specified character is a
punctuation character, analogous to C<m/[[:punct:]]/>.
Note that the definition of what is punctuation isn't as
=for apidoc Am|bool|isSPACE|UV ch
=for apidoc_item ||isSPACE_A|UV ch
-=for apidoc_item ||isSPACE_L1|UV ch
-=for apidoc_item ||isSPACE_uvchr|UV ch
-=for apidoc_item ||isSPACE_utf8_safe|U8 * s|U8 * end
-=for apidoc_item ||isSPACE_utf8|U8 * s|U8 * end
=for apidoc_item ||isSPACE_LC|UV ch
-=for apidoc_item ||isSPACE_LC_uvchr|UV ch
=for apidoc_item ||isSPACE_LC_utf8_safe|U8 * s| U8 *end
+=for apidoc_item ||isSPACE_LC_uvchr|UV ch
+=for apidoc_item ||isSPACE_L1|UV ch
+=for apidoc_item ||isSPACE_utf8|U8 * s|U8 * end
+=for apidoc_item ||isSPACE_utf8_safe|U8 * s|U8 * end
+=for apidoc_item ||isSPACE_uvchr|UV ch
Returns a boolean indicating whether the specified character is a
whitespace character. This is analogous
to what C<m/\s/> matches in a regular expression. Starting in Perl 5.18
=for apidoc Am|bool|isPSXSPC|UV ch
=for apidoc_item ||isPSXSPC_A|UV ch
-=for apidoc_item ||isPSXSPC_L1|UV ch
-=for apidoc_item ||isPSXSPC_uvchr|UV ch
-=for apidoc_item ||isPSXSPC_utf8_safe|U8 * s|U8 * end
-=for apidoc_item ||isPSXSPC_utf8|U8 * s|U8 * end
=for apidoc_item ||isPSXSPC_LC|UV ch
-=for apidoc_item ||isPSXSPC_LC_uvchr|UV ch
=for apidoc_item ||isPSXSPC_LC_utf8_safe|U8 * s| U8 *end
+=for apidoc_item ||isPSXSPC_LC_uvchr|UV ch
+=for apidoc_item ||isPSXSPC_L1|UV ch
+=for apidoc_item ||isPSXSPC_utf8|U8 * s|U8 * end
+=for apidoc_item ||isPSXSPC_utf8_safe|U8 * s|U8 * end
+=for apidoc_item ||isPSXSPC_uvchr|UV ch
(short for Posix Space)
Starting in 5.18, this is identical in all its forms to the
corresponding C<isSPACE()> macros.
=for apidoc Am|bool|isUPPER|UV ch
=for apidoc_item ||isUPPER_A|UV ch
-=for apidoc_item ||isUPPER_L1|UV ch
-=for apidoc_item ||isUPPER_uvchr|UV ch
-=for apidoc_item ||isUPPER_utf8_safe|U8 * s|U8 * end
-=for apidoc_item ||isUPPER_utf8|U8 * s|U8 * end
=for apidoc_item ||isUPPER_LC|UV ch
-=for apidoc_item ||isUPPER_LC_uvchr|UV ch
=for apidoc_item ||isUPPER_LC_utf8_safe|U8 * s| U8 *end
+=for apidoc_item ||isUPPER_LC_uvchr|UV ch
+=for apidoc_item ||isUPPER_L1|UV ch
+=for apidoc_item ||isUPPER_utf8|U8 * s|U8 * end
+=for apidoc_item ||isUPPER_utf8_safe|U8 * s|U8 * end
+=for apidoc_item ||isUPPER_uvchr|UV ch
Returns a boolean indicating whether the specified character is an
uppercase character, analogous to C<m/[[:upper:]]/>.
See the L<top of this section|/Character classification> for an explanation of
=for apidoc Am|bool|isPRINT|UV ch
=for apidoc_item ||isPRINT_A|UV ch
-=for apidoc_item ||isPRINT_L1|UV ch
-=for apidoc_item ||isPRINT_uvchr|UV ch
-=for apidoc_item ||isPRINT_utf8_safe|U8 * s|U8 * end
-=for apidoc_item ||isPRINT_utf8|U8 * s|U8 * end
=for apidoc_item ||isPRINT_LC|UV ch
-=for apidoc_item ||isPRINT_LC_uvchr|UV ch
=for apidoc_item ||isPRINT_LC_utf8_safe|U8 * s| U8 *end
+=for apidoc_item ||isPRINT_LC_uvchr|UV ch
+=for apidoc_item ||isPRINT_L1|UV ch
+=for apidoc_item ||isPRINT_utf8|U8 * s|U8 * end
+=for apidoc_item ||isPRINT_utf8_safe|U8 * s|U8 * end
+=for apidoc_item ||isPRINT_uvchr|UV ch
Returns a boolean indicating whether the specified character is a
printable character, analogous to C<m/[[:print:]]/>.
See the L<top of this section|/Character classification> for an explanation of
=for apidoc Am|bool|isWORDCHAR|UV ch
=for apidoc_item ||isWORDCHAR_A|UV ch
-=for apidoc_item ||isWORDCHAR_L1|UV ch
-=for apidoc_item ||isWORDCHAR_uvchr|UV ch
-=for apidoc_item ||isWORDCHAR_utf8_safe|U8 * s|U8 * end
-=for apidoc_item ||isWORDCHAR_utf8|U8 * s|U8 * end
=for apidoc_item ||isWORDCHAR_LC|UV ch
-=for apidoc_item ||isWORDCHAR_LC_uvchr|UV ch
=for apidoc_item ||isWORDCHAR_LC_utf8_safe|U8 * s| U8 *end
-=for apidoc_item ||isALNUM|UV ch
-=for apidoc_item ||isALNUM_A|UV ch
-=for apidoc_item ||isALNUM_LC|UV ch
-=for apidoc_item ||isALNUM_LC_uvchr|UV ch
+=for apidoc_item ||isWORDCHAR_LC_uvchr|UV ch
+=for apidoc_item ||isWORDCHAR_L1|UV ch
+=for apidoc_item ||isWORDCHAR_utf8|U8 * s|U8 * end
+=for apidoc_item ||isWORDCHAR_utf8_safe|U8 * s|U8 * end
+=for apidoc_item ||isWORDCHAR_uvchr|UV ch
Returns a boolean indicating whether the specified character is a character
that is a word character, analogous to what C<m/\w/> and C<m/[[:word:]]/> match
in a regular expression. A word character is an alphabetic character, a
decimal digit, a connecting punctuation character (such as an underscore), or
a "mark" character that attaches to one of those (like some sort of accent).
-C<isALNUM()> is a synonym provided for backward compatibility, even though a
-word character includes more than the standard C language meaning of
-alphanumeric.
+
See the L<top of this section|/Character classification> for an explanation of
the variants.
+
C<isWORDCHAR_A>, C<isWORDCHAR_L1>, C<isWORDCHAR_uvchr>,
C<isWORDCHAR_LC>, C<isWORDCHAR_LC_uvchr>, C<isWORDCHAR_LC_utf8>, and
C<isWORDCHAR_LC_utf8_safe> are also as described there, but additionally
include the platform's native underscore.
+=for apidoc Am|bool|isALNUM |UV ch
+=for apidoc_item ||isALNUM_A |UV ch
+=for apidoc_item ||isALNUM_LC |UV ch
+=for apidoc_item ||isALNUM_LC_uvchr|UV ch
+These are each a synonym for their respectively named L</C<isWORDCHAR>>
+variant.
+
+They are provided for backward compatibility, even though a word character
+includes more than the standard C language meaning of alphanumeric.
+To get the C language definition, use the corresponding L</C<isALPHANUMERIC>>
+variant.
+
=for apidoc Am|bool|isXDIGIT|UV ch
=for apidoc_item ||isXDIGIT_A|UV ch
-=for apidoc_item ||isXDIGIT_L1|UV ch
-=for apidoc_item ||isXDIGIT_uvchr|UV ch
-=for apidoc_item ||isXDIGIT_utf8_safe|U8 * s|U8 * end
-=for apidoc_item ||isXDIGIT_utf8|U8 * s|U8 * end
=for apidoc_item ||isXDIGIT_LC|UV ch
-=for apidoc_item ||isXDIGIT_LC_uvchr|UV ch
=for apidoc_item ||isXDIGIT_LC_utf8_safe|U8 * s| U8 *end
+=for apidoc_item ||isXDIGIT_LC_uvchr|UV ch
+=for apidoc_item ||isXDIGIT_L1|UV ch
+=for apidoc_item ||isXDIGIT_utf8|U8 * s|U8 * end
+=for apidoc_item ||isXDIGIT_utf8_safe|U8 * s|U8 * end
+=for apidoc_item ||isXDIGIT_uvchr|UV ch
Returns a boolean indicating whether the specified character is a hexadecimal
digit. In the ASCII range these are C<[0-9A-Fa-f]>. Variants C<isXDIGIT_A()>
and C<isXDIGIT_L1()> are identical to C<isXDIGIT()>.
=for apidoc Am|bool|isIDFIRST|UV ch
=for apidoc_item ||isIDFIRST_A|UV ch
-=for apidoc_item ||isIDFIRST_L1|UV ch
-=for apidoc_item ||isIDFIRST_uvchr|UV ch
-=for apidoc_item ||isIDFIRST_utf8_safe|U8 * s|U8 * end
-=for apidoc_item ||isIDFIRST_utf8|U8 * s|U8 * end
=for apidoc_item ||isIDFIRST_LC|UV ch
-=for apidoc_item ||isIDFIRST_LC_uvchr|UV ch
=for apidoc_item ||isIDFIRST_LC_utf8_safe|U8 * s| U8 *end
+=for apidoc_item ||isIDFIRST_LC_uvchr|UV ch
+=for apidoc_item ||isIDFIRST_L1|UV ch
+=for apidoc_item ||isIDFIRST_utf8|U8 * s|U8 * end
+=for apidoc_item ||isIDFIRST_utf8_safe|U8 * s|U8 * end
+=for apidoc_item ||isIDFIRST_uvchr|UV ch
Returns a boolean indicating whether the specified character can be the first
character of an identifier. This is very close to, but not quite the same as
the official Unicode property C<XID_Start>. The difference is that this
=for apidoc Am|bool|isIDCONT|UV ch
=for apidoc_item ||isIDCONT_A|UV ch
-=for apidoc_item ||isIDCONT_L1|UV ch
-=for apidoc_item ||isIDCONT_uvchr|UV ch
-=for apidoc_item ||isIDCONT_utf8_safe|U8 * s|U8 * end
-=for apidoc_item ||isIDCONT_utf8|U8 * s|U8 * end
=for apidoc_item ||isIDCONT_LC|UV ch
-=for apidoc_item ||isIDCONT_LC_uvchr|UV ch
=for apidoc_item ||isIDCONT_LC_utf8_safe|U8 * s| U8 *end
+=for apidoc_item ||isIDCONT_LC_uvchr|UV ch
+=for apidoc_item ||isIDCONT_L1|UV ch
+=for apidoc_item ||isIDCONT_utf8|U8 * s|U8 * end
+=for apidoc_item ||isIDCONT_utf8_safe|U8 * s|U8 * end
+=for apidoc_item ||isIDCONT_uvchr|UV ch
Returns a boolean indicating whether the specified character can be the
second or succeeding character of an identifier. This is very close to, but
not quite the same as the official Unicode property C<XID_Continue>. The
=for apidoc Am|UV|toUPPER|UV cp
=for apidoc_item |UV|toUPPER_A|UV cp
-=for apidoc_item |UV|toUPPER_uvchr|UV cp|U8* s|STRLEN* lenp
=for apidoc_item |UV|toUPPER_utf8|U8* p|U8* e|U8* s|STRLEN* lenp
=for apidoc_item |UV|toUPPER_utf8_safe|U8* p|U8* e|U8* s|STRLEN* lenp
+=for apidoc_item |UV|toUPPER_uvchr|UV cp|U8* s|STRLEN* lenp
These all return the uppercase of a character. The differences are what domain
they operate on, and whether the input is specified as a code point (those
=for apidoc Am|UV|toFOLD|UV cp
=for apidoc_item |UV|toFOLD_A|UV cp
-=for apidoc_item |UV|toFOLD_uvchr|UV cp|U8* s|STRLEN* lenp
=for apidoc_item |UV|toFOLD_utf8|U8* p|U8* e|U8* s|STRLEN* lenp
=for apidoc_item |UV|toFOLD_utf8_safe|U8* p|U8* e|U8* s|STRLEN* lenp
+=for apidoc_item |UV|toFOLD_uvchr|UV cp|U8* s|STRLEN* lenp
These all return the foldcase of a character. "foldcase" is an internal case
for C</i> pattern matching. If the foldcase of character A and the foldcase of
=for apidoc Am|UV|toLOWER|UV cp
=for apidoc_item |UV|toLOWER_A|UV cp
-=for apidoc_item |UV|toLOWER_L1|UV cp
=for apidoc_item |UV|toLOWER_LATIN1|UV cp
=for apidoc_item |UV|toLOWER_LC|UV cp
-=for apidoc_item |UV|toLOWER_uvchr|UV cp|U8* s|STRLEN* lenp
+=for apidoc_item |UV|toLOWER_L1|UV cp
=for apidoc_item |UV|toLOWER_utf8|U8* p|U8* e|U8* s|STRLEN* lenp
=for apidoc_item |UV|toLOWER_utf8_safe|U8* p|U8* e|U8* s|STRLEN* lenp
+=for apidoc_item |UV|toLOWER_uvchr|UV cp|U8* s|STRLEN* lenp
These all return the lowercase of a character. The differences are what domain
they operate on, and whether the input is specified as a code point (those
=for apidoc Am|UV|toTITLE|UV cp
=for apidoc_item |UV|toTITLE_A|UV cp
-=for apidoc_item |UV|toTITLE_uvchr|UV cp|U8* s|STRLEN* lenp
=for apidoc_item |UV|toTITLE_utf8|U8* p|U8* e|U8* s|STRLEN* lenp
=for apidoc_item |UV|toTITLE_utf8_safe|U8* p|U8* e|U8* s|STRLEN* lenp
+=for apidoc_item |UV|toTITLE_uvchr|UV cp|U8* s|STRLEN* lenp
These all return the titlecase of a character. The differences are what domain
they operate on, and whether the input is specified as a code point (those
XXX Still undocumented isVERTWS_uvchr and _utf8; it's unclear what their names
really should be. Also toUPPER_LC and toFOLD_LC, which are subject to change,
and aren't general purpose as they don't work on U+DF, and assert against that.
+and isCASED_LC, as it really is more of an internal thing.
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
=cut
*/
-#ifdef QUADKIND
-# define WIDEST_UTYPE U64
-#else
-# define WIDEST_UTYPE U32
-#endif
+#define WIDEST_UTYPE PERL_UINTMAX_T
/* Where there could be some confusion, use this as a static assert in macros
* to make sure that a parameter isn't a pointer. But some compilers can't
/* Character class numbers. For internal core Perl use only. The ones less
* than 32 are used in PL_charclass[] and the ones up through the one that
- * corresponds to <_HIGHEST_REGCOMP_DOT_H_SYNC> are used by regcomp.h and
+ * corresponds to <HIGHEST_REGCOMP_DOT_H_SYNC_> are used by regcomp.h and
* related files. PL_charclass ones use names used in l1_char_class_tab.h but
* their actual definitions are here. If that file has a name not used here,
* it won't compile.
* to group these which have no members that match above Latin1, (or above
* ASCII in the latter case) */
-# define _CC_WORDCHAR 0 /* \w and [:word:] */
-# define _CC_DIGIT 1 /* \d and [:digit:] */
-# define _CC_ALPHA 2 /* [:alpha:] */
-# define _CC_LOWER 3 /* [:lower:] */
-# define _CC_UPPER 4 /* [:upper:] */
-# define _CC_PUNCT 5 /* [:punct:] */
-# define _CC_PRINT 6 /* [:print:] */
-# define _CC_ALPHANUMERIC 7 /* [:alnum:] */
-# define _CC_GRAPH 8 /* [:graph:] */
-# define _CC_CASED 9 /* [:lower:] or [:upper:] under /i */
-# define _CC_SPACE 10 /* \s, [:space:] */
-# define _CC_BLANK 11 /* [:blank:] */
-# define _CC_XDIGIT 12 /* [:xdigit:] */
-# define _CC_CNTRL 13 /* [:cntrl:] */
-# define _CC_ASCII 14 /* [:ascii:] */
-# define _CC_VERTSPACE 15 /* \v */
-
-# define _HIGHEST_REGCOMP_DOT_H_SYNC _CC_VERTSPACE
+# define CC_WORDCHAR_ 0 /* \w and [:word:] */
+# define CC_DIGIT_ 1 /* \d and [:digit:] */
+# define CC_ALPHA_ 2 /* [:alpha:] */
+# define CC_LOWER_ 3 /* [:lower:] */
+# define CC_UPPER_ 4 /* [:upper:] */
+# define CC_PUNCT_ 5 /* [:punct:] */
+# define CC_PRINT_ 6 /* [:print:] */
+# define CC_ALPHANUMERIC_ 7 /* [:alnum:] */
+# define CC_GRAPH_ 8 /* [:graph:] */
+# define CC_CASED_ 9 /* [:lower:] or [:upper:] under /i */
+# define CC_SPACE_ 10 /* \s, [:space:] */
+# define CC_BLANK_ 11 /* [:blank:] */
+# define CC_XDIGIT_ 12 /* [:xdigit:] */
+# define CC_CNTRL_ 13 /* [:cntrl:] */
+# define CC_ASCII_ 14 /* [:ascii:] */
+# define CC_VERTSPACE_ 15 /* \v */
+
+# 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. */
-# define _CC_IDFIRST 16
-# define _CC_CHARNAME_CONT 17
-# define _CC_NONLATIN1_FOLD 18
-# define _CC_NONLATIN1_SIMPLE_FOLD 19
-# define _CC_QUOTEMETA 20
-# define _CC_NON_FINAL_FOLD 21
-# define _CC_IS_IN_SOME_FOLD 22
-# define _CC_BINDIGIT 23
-# define _CC_OCTDIGIT 24
-# define _CC_MNEMONIC_CNTRL 25
+# define CC_IDFIRST_ 16
+# define CC_CHARNAME_CONT_ 17
+# define CC_NONLATIN1_FOLD_ 18
+# define CC_NONLATIN1_SIMPLE_FOLD_ 19
+# define CC_QUOTEMETA_ 20
+# define CC_NON_FINAL_FOLD_ 21
+# define CC_IS_IN_SOME_FOLD_ 22
+# define CC_BINDIGIT_ 23
+# define CC_OCTDIGIT_ 24
+# define CC_MNEMONIC_CNTRL_ 25
/* Unused: 26-31
* If more bits are needed, one could add a second word for non-64bit
/* An enum version of the character class numbers, to help compilers
* optimize */
typedef enum {
- _CC_ENUM_ALPHA = _CC_ALPHA,
- _CC_ENUM_ALPHANUMERIC = _CC_ALPHANUMERIC,
- _CC_ENUM_ASCII = _CC_ASCII,
- _CC_ENUM_BLANK = _CC_BLANK,
- _CC_ENUM_CASED = _CC_CASED,
- _CC_ENUM_CNTRL = _CC_CNTRL,
- _CC_ENUM_DIGIT = _CC_DIGIT,
- _CC_ENUM_GRAPH = _CC_GRAPH,
- _CC_ENUM_LOWER = _CC_LOWER,
- _CC_ENUM_PRINT = _CC_PRINT,
- _CC_ENUM_PUNCT = _CC_PUNCT,
- _CC_ENUM_SPACE = _CC_SPACE,
- _CC_ENUM_UPPER = _CC_UPPER,
- _CC_ENUM_VERTSPACE = _CC_VERTSPACE,
- _CC_ENUM_WORDCHAR = _CC_WORDCHAR,
- _CC_ENUM_XDIGIT = _CC_XDIGIT
-} _char_class_number;
+ CC_ENUM_ALPHA_ = CC_ALPHA_,
+ CC_ENUM_ALPHANUMERIC_ = CC_ALPHANUMERIC_,
+ CC_ENUM_ASCII_ = CC_ASCII_,
+ CC_ENUM_BLANK_ = CC_BLANK_,
+ CC_ENUM_CASED_ = CC_CASED_,
+ CC_ENUM_CNTRL_ = CC_CNTRL_,
+ CC_ENUM_DIGIT_ = CC_DIGIT_,
+ CC_ENUM_GRAPH_ = CC_GRAPH_,
+ CC_ENUM_LOWER_ = CC_LOWER_,
+ CC_ENUM_PRINT_ = CC_PRINT_,
+ CC_ENUM_PUNCT_ = CC_PUNCT_,
+ CC_ENUM_SPACE_ = CC_SPACE_,
+ CC_ENUM_UPPER_ = CC_UPPER_,
+ CC_ENUM_VERTSPACE_ = CC_VERTSPACE_,
+ CC_ENUM_WORDCHAR_ = CC_WORDCHAR_,
+ CC_ENUM_XDIGIT_ = CC_XDIGIT_
+} char_class_number_;
#endif
-#define POSIX_CC_COUNT (_HIGHEST_REGCOMP_DOT_H_SYNC + 1)
+#define POSIX_CC_COUNT (HIGHEST_REGCOMP_DOT_H_SYNC_ + 1)
START_EXTERN_C
# ifdef DOINIT
END_EXTERN_C
/* The 1U keeps Solaris from griping when shifting sets the uppermost bit */
-# define _CC_mask(classnum) (1U << (classnum))
+# define CC_mask_(classnum) (1U << (classnum))
/* For internal core Perl use only: the base macro for defining macros like
* isALPHA */
-# define _generic_isCC(c, classnum) cBOOL(FITS_IN_8_BITS(c) \
- && (PL_charclass[(U8) (c)] & _CC_mask(classnum)))
+# define generic_isCC_(c, classnum) cBOOL(FITS_IN_8_BITS(c) \
+ && (PL_charclass[(U8) (c)] & CC_mask_(classnum)))
/* The mask for the _A versions of the macros; it just adds in the bit for
* ASCII. */
-# define _CC_mask_A(classnum) (_CC_mask(classnum) | _CC_mask(_CC_ASCII))
+# define CC_mask_A_(classnum) (CC_mask_(classnum) | CC_mask_(CC_ASCII_))
/* For internal core Perl use only: the base macro for defining macros like
* isALPHA_A. The foo_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) \
- && ((PL_charclass[(U8) (c)] & _CC_mask_A(classnum)) \
- == _CC_mask_A(classnum)))
+# define generic_isCC_A_(c, classnum) (FITS_IN_8_BITS(c) \
+ && ((PL_charclass[(U8) (c)] & CC_mask_A_(classnum)) \
+ == CC_mask_A_(classnum)))
/* On ASCII platforms certain classes form a single range. It's faster to
* special case these. isDIGIT is a single range on all platforms */
# ifdef EBCDIC
-# define isALPHA_A(c) _generic_isCC_A(c, _CC_ALPHA)
-# define isGRAPH_A(c) _generic_isCC_A(c, _CC_GRAPH)
-# define isLOWER_A(c) _generic_isCC_A(c, _CC_LOWER)
-# define isPRINT_A(c) _generic_isCC_A(c, _CC_PRINT)
-# define isUPPER_A(c) _generic_isCC_A(c, _CC_UPPER)
+# define isALPHA_A(c) generic_isCC_A_(c, CC_ALPHA_)
+# define isGRAPH_A(c) generic_isCC_A_(c, CC_GRAPH_)
+# define isLOWER_A(c) generic_isCC_A_(c, CC_LOWER_)
+# define isPRINT_A(c) generic_isCC_A_(c, CC_PRINT_)
+# define isUPPER_A(c) generic_isCC_A_(c, CC_UPPER_)
# else
/* By folding the upper and lowercase, we can use a single range */
# define isALPHA_A(c) inRANGE((~('A' ^ 'a') & (c)), 'A', 'Z')
# define isPRINT_A(c) inRANGE(c, ' ', 0x7e)
# define isUPPER_A(c) inRANGE(c, 'A', 'Z')
# endif
-# define isALPHANUMERIC_A(c) _generic_isCC_A(c, _CC_ALPHANUMERIC)
-# define isBLANK_A(c) _generic_isCC_A(c, _CC_BLANK)
-# define isCNTRL_A(c) _generic_isCC_A(c, _CC_CNTRL)
+# define isALPHANUMERIC_A(c) generic_isCC_A_(c, CC_ALPHANUMERIC_)
+# define isBLANK_A(c) generic_isCC_A_(c, CC_BLANK_)
+# define isCNTRL_A(c) generic_isCC_A_(c, CC_CNTRL_)
# define isDIGIT_A(c) inRANGE(c, '0', '9')
-# define isPUNCT_A(c) _generic_isCC_A(c, _CC_PUNCT)
-# define isSPACE_A(c) _generic_isCC_A(c, _CC_SPACE)
-# define isWORDCHAR_A(c) _generic_isCC_A(c, _CC_WORDCHAR)
-# define isXDIGIT_A(c) _generic_isCC(c, _CC_XDIGIT) /* No non-ASCII xdigits
+# define isPUNCT_A(c) generic_isCC_A_(c, CC_PUNCT_)
+# define isSPACE_A(c) generic_isCC_A_(c, CC_SPACE_)
+# define isWORDCHAR_A(c) generic_isCC_A_(c, CC_WORDCHAR_)
+# define isXDIGIT_A(c) generic_isCC_(c, CC_XDIGIT_) /* No non-ASCII xdigits
*/
-# 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)
+# 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 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 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) isSPACE_L1(c)
-# 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)
+# 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_)
# ifdef EBCDIC
-# define isASCII(c) _generic_isCC(c, _CC_ASCII)
+# define isASCII(c) generic_isCC_(c, CC_ASCII_)
# endif
/* Participates in a single-character fold with a character above 255 */
# if defined(PERL_IN_REGCOMP_C) || defined(PERL_IN_REGEXEC_C)
-# define HAS_NONLATIN1_SIMPLE_FOLD_CLOSURE(c) \
+# define HAS_NONLATIN1_SIMPLE_FOLD_CLOSURE(c) \
(( ! cBOOL(FITS_IN_8_BITS(c))) \
- || (PL_charclass[(U8) (c)] & _CC_mask(_CC_NONLATIN1_SIMPLE_FOLD)))
+ || (PL_charclass[(U8) (c)] & CC_mask_(CC_NONLATIN1_SIMPLE_FOLD_)))
-# define IS_NON_FINAL_FOLD(c) _generic_isCC(c, _CC_NON_FINAL_FOLD)
-# define IS_IN_SOME_FOLD_L1(c) _generic_isCC(c, _CC_IS_IN_SOME_FOLD)
+# define IS_NON_FINAL_FOLD(c) generic_isCC_(c, CC_NON_FINAL_FOLD_)
+# define IS_IN_SOME_FOLD_L1(c) generic_isCC_(c, CC_IS_IN_SOME_FOLD_)
# endif
/* Like the above, but also can be part of a multi-char fold */
# define HAS_NONLATIN1_FOLD_CLOSURE(c) \
( (! cBOOL(FITS_IN_8_BITS(c))) \
- || (PL_charclass[(U8) (c)] & _CC_mask(_CC_NONLATIN1_FOLD)))
+ || (PL_charclass[(U8) (c)] & CC_mask_(CC_NONLATIN1_FOLD_)))
-# define _isQUOTEMETA(c) _generic_isCC(c, _CC_QUOTEMETA)
+# define _isQUOTEMETA(c) generic_isCC_(c, CC_QUOTEMETA_)
/* is c a control character for which we have a mnemonic? */
# if defined(PERL_CORE) || defined(PERL_EXT)
-# define isMNEMONIC_CNTRL(c) _generic_isCC(c, _CC_MNEMONIC_CNTRL)
+# define isMNEMONIC_CNTRL(c) generic_isCC_(c, CC_MNEMONIC_CNTRL_)
# endif
#else /* else we don't have perl.h H_PERL */
/* 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 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
* 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) \
+# define generic_isCC_(c, classnum) \
(FITS_IN_8_BITS(c) && S_bootstrap_ctype((U8) (c), (classnum), TRUE))
-# define _generic_isCC_A(c, classnum) \
+# define generic_isCC_A_(c, classnum) \
(FITS_IN_8_BITS(c) && S_bootstrap_ctype((U8) (c), (classnum), FALSE))
#endif /* End of no perl.h H_PERL */
/* For internal core Perl use only: the base macro for defining macros like
* isALPHA_LC, which uses the current LC_CTYPE locale. 'c' is the code point
* (0-255) to check. In a UTF-8 locale, the result is the same as calling
- * isFOO_L1(); the 'utf8_locale_classnum' parameter is something like
- * _CC_UPPER, which gives the class number for doing this. For non-UTF-8
- * locales, the code to actually do the test this is passed in 'non_utf8'. If
- * 'c' is above 255, 0 is returned. For accessing the full range of possible
- * code points under locale rules, use the macros based on _generic_LC_uvchr
- * instead of this. */
-#define _generic_LC_base(c, utf8_locale_classnum, non_utf8) \
- (! FITS_IN_8_BITS(c) \
- ? 0 \
- : IN_UTF8_CTYPE_LOCALE \
- ? cBOOL(PL_charclass[(U8) (c)] & _CC_mask(utf8_locale_classnum)) \
- : cBOOL(non_utf8))
-
-/* For internal core Perl use only: a helper macro for defining macros like
- * isALPHA_LC. 'c' is the code point (0-255) to check. The function name to
- * actually do this test is passed in 'non_utf8_func', which is called on 'c',
- * casting 'c' to the macro _LC_CAST, which should not be parenthesized. See
- * _generic_LC_base for more info */
-#define _generic_LC(c, utf8_locale_classnum, non_utf8_func) \
- _generic_LC_base(c,utf8_locale_classnum, \
- non_utf8_func( (_LC_CAST) (c)))
-
-/* For internal core Perl use only: like _generic_LC, but also returns TRUE if
- * 'c' is the platform's native underscore character */
-#define _generic_LC_underscore(c,utf8_locale_classnum,non_utf8_func) \
- _generic_LC_base(c, utf8_locale_classnum, \
- (non_utf8_func( (_LC_CAST) (c)) \
- || (char)(c) == '_'))
-
-/* These next three are also for internal core Perl use only: case-change
- * helper macros. The reason for using the PL_latin arrays is in case the
- * system function is defective; it ensures uniform results that conform to the
- * Unicod standard. It does not handle the anomalies in UTF-8 Turkic locales */
-#define _generic_toLOWER_LC(c, function, cast) (! FITS_IN_8_BITS(c) \
- ? (c) \
- : (IN_UTF8_CTYPE_LOCALE) \
- ? PL_latin1_lc[ (U8) (c) ] \
- : (cast)function((cast)(c)))
-
-/* Note that the result can be larger than a byte in a UTF-8 locale. It
- * returns a single value, so can't adequately return the upper case of LATIN
- * SMALL LETTER SHARP S in a UTF-8 locale (which should be a string of two
- * values "SS"); instead it asserts against that under DEBUGGING, and
- * otherwise returns its input. It does not handle the anomalies in UTF-8
- * Turkic locales. */
-#define _generic_toUPPER_LC(c, function, cast) \
- (! FITS_IN_8_BITS(c) \
- ? (c) \
- : ((! IN_UTF8_CTYPE_LOCALE) \
- ? (cast)function((cast)(c)) \
- : ((((U8)(c)) == MICRO_SIGN) \
- ? GREEK_CAPITAL_LETTER_MU \
- : ((((U8)(c)) == LATIN_SMALL_LETTER_Y_WITH_DIAERESIS) \
- ? LATIN_CAPITAL_LETTER_Y_WITH_DIAERESIS \
- : ((((U8)(c)) == LATIN_SMALL_LETTER_SHARP_S) \
- ? (__ASSERT_(0) (c)) \
- : PL_mod_latin1_uc[ (U8) (c) ])))))
-
-/* Note that the result can be larger than a byte in a UTF-8 locale. It
- * returns a single value, so can't adequately return the fold case of LATIN
- * SMALL LETTER SHARP S in a UTF-8 locale (which should be a string of two
- * values "ss"); instead it asserts against that under DEBUGGING, and
- * otherwise returns its input. It does not handle the anomalies in UTF-8
- * Turkic locales */
-#define _generic_toFOLD_LC(c, function, cast) \
- ((UNLIKELY((c) == MICRO_SIGN) && IN_UTF8_CTYPE_LOCALE) \
- ? GREEK_SMALL_LETTER_MU \
- : (__ASSERT_(! IN_UTF8_CTYPE_LOCALE \
- || (c) != LATIN_SMALL_LETTER_SHARP_S) \
- _generic_toLOWER_LC(c, function, cast)))
-
-/* Use the libc versions for these if available. */
+ * isFOO_L1(); 'classnum' is something like CC_UPPER_, which gives the class
+ * number for doing this. For non-UTF-8 locales, the code to actually do the
+ * test this is passed in 'non_utf8'. If 'c' is above 255, 0 is returned. For
+ * accessing the full range of possible code points under locale rules, use the
+ * macros based on generic_LC_uvchr_ instead of this. */
+#define generic_LC_base_(c, classnum, non_utf8_func) \
+ (! FITS_IN_8_BITS(c) \
+ ? 0 \
+ : IN_UTF8_CTYPE_LOCALE \
+ ? cBOOL(PL_charclass[(U8) (c)] & CC_mask_(classnum)) \
+ : cBOOL(non_utf8_func(c)))
+
+/* A helper macro for defining macros like isALPHA_LC. On systems without
+ * proper locales, these reduce to, e.g., isALPHA_A */
+#ifdef CTYPE256
+# define generic_LC_(c, classnum, non_utf8_func) \
+ generic_LC_base_(c, classnum, non_utf8_func)
+#else
+# define generic_LC_(c, classnum, non_utf8_func) \
+ generic_isCC_A_(c, classnum)
+#endif
+
+/* Below are the definitions for the locale-sensitive character classification
+ * macros whose input domain is a byte, and the locale isn't UTF-8. These are
+ * as close as possible to the bare versions on the platform and still yield
+ * POSIX Standard-compliant results.
+ *
+ * There is currently only one place these definitions should be used, in
+ * certain function calls like Perl_iswordchar_() in inline.h.
+ *
+ * Most likely you want to use the macros a ways below with names like
+ * isALPHA_LC(). Rarely, you may want isU8_ALPHA_LC(), somewhat below.
+ *
+ * The first two aren't in C89, so the fallback is to use the non-locale
+ * sensitive versions; these are the same for all platforms */
#if defined(HAS_ISASCII)
-# define isASCII_LC(c) (FITS_IN_8_BITS(c) && isascii( (U8) (c)))
+# define is_base_ASCII(c) isascii((U8) (c))
#else
-# define isASCII_LC(c) isASCII(c)
+# define is_base_ASCII(c) isASCII(c)
#endif
#if defined(HAS_ISBLANK)
-# define isBLANK_LC(c) _generic_LC(c, _CC_BLANK, isblank)
-#else /* Unlike isASCII, varies if in a UTF-8 locale */
-# define isBLANK_LC(c) ((IN_UTF8_CTYPE_LOCALE) ? isBLANK_L1(c) : isBLANK(c))
+# define is_base_BLANK(c) isblank((U8) (c))
+#else
+# define is_base_BLANK(c) isBLANK(c)
#endif
-#define _LC_CAST U8
+/* The next few are the same in all platforms. */
+#define is_base_CNTRL(c) iscntrl((U8) (c))
+#define is_base_IDFIRST(c) (UNLIKELY((c) == '_') || is_base_ALPHA(c))
+#define is_base_SPACE(c) isspace((U8) (c))
+#define is_base_WORDCHAR(c) (UNLIKELY((c) == '_') || is_base_ALPHANUMERIC(c))
+
+/* The base-level case changing macros are also the same in all platforms */
+#define to_base_LOWER(c) tolower((U8) (c))
+#define to_base_UPPER(c) toupper((U8) (c))
+#define to_base_FOLD(c) to_base_LOWER(c)
#ifdef WIN32
- /* The Windows functions don't bother to follow the POSIX standard, which
- * for example says that something can't both be a printable and a control.
- * But Windows treats the \t control as a printable, and does such things
- * as making superscripts into both digits and punctuation. This tames
- * these flaws by assuming that the definitions of both controls and space
- * are correct, and then making sure that other definitions don't have
- * weirdnesses, by making sure that isalnum() isn't also ispunct(), etc.
- * Not all possible weirdnesses are checked for, just the ones that were
- * detected on actual Microsoft code pages */
-
-# define isCNTRL_LC(c) _generic_LC(c, _CC_CNTRL, iscntrl)
-# define isSPACE_LC(c) _generic_LC(c, _CC_SPACE, isspace)
-
-# define isALPHA_LC(c) (_generic_LC(c, _CC_ALPHA, isalpha) \
- && isALPHANUMERIC_LC(c))
-# define isALPHANUMERIC_LC(c) (_generic_LC(c, _CC_ALPHANUMERIC, isalnum) && \
- ! isPUNCT_LC(c))
-# define isDIGIT_LC(c) (_generic_LC(c, _CC_DIGIT, isdigit) && \
- isALPHANUMERIC_LC(c))
-# define isGRAPH_LC(c) (_generic_LC(c, _CC_GRAPH, isgraph) && isPRINT_LC(c))
-# define isIDFIRST_LC(c) (((c) == '_') \
- || (_generic_LC(c, _CC_IDFIRST, isalpha) && ! isPUNCT_LC(c)))
-# define isLOWER_LC(c) (_generic_LC(c, _CC_LOWER, islower) && isALPHA_LC(c))
-# define isPRINT_LC(c) (_generic_LC(c, _CC_PRINT, isprint) && ! isCNTRL_LC(c))
-# define isPUNCT_LC(c) (_generic_LC(c, _CC_PUNCT, ispunct) && ! isCNTRL_LC(c))
-# define isUPPER_LC(c) (_generic_LC(c, _CC_UPPER, isupper) && isALPHA_LC(c))
-# define isWORDCHAR_LC(c) (((c) == '_') || isALPHANUMERIC_LC(c))
-# define isXDIGIT_LC(c) (_generic_LC(c, _CC_XDIGIT, isxdigit) \
- && isALPHANUMERIC_LC(c))
-
-# define toLOWER_LC(c) _generic_toLOWER_LC((c), tolower, U8)
-# define toUPPER_LC(c) _generic_toUPPER_LC((c), toupper, U8)
-# define toFOLD_LC(c) _generic_toFOLD_LC((c), tolower, U8)
-
-#elif defined(CTYPE256) || (!defined(isascii) && !defined(HAS_ISASCII))
- /* For most other platforms */
-
-# define isALPHA_LC(c) _generic_LC(c, _CC_ALPHA, isalpha)
-# define isALPHANUMERIC_LC(c) _generic_LC(c, _CC_ALPHANUMERIC, isalnum)
-# define isCNTRL_LC(c) _generic_LC(c, _CC_CNTRL, iscntrl)
-# define isDIGIT_LC(c) _generic_LC(c, _CC_DIGIT, isdigit)
-# ifdef OS390 /* This system considers NBSP to be a graph */
-# define isGRAPH_LC(c) _generic_LC(c, _CC_GRAPH, isgraph) \
- && ! isSPACE_LC(c)
+
+/* The Windows functions don't bother to follow the POSIX standard, which for
+ * example says that something can't both be a printable and a control. But
+ * Windows treats \t as both a control and a printable, and does such things as
+ * making superscripts into both digits and punctuation. These #defines tame
+ * these flaws by assuming that the definitions of controls (and the other few
+ * ones defined above) are correct, and then making sure that other definitions
+ * don't have weirdnesses, by adding a check that \w and its subsets aren't
+ * ispunct(), and things that are \W, like ispunct(), arent't controls. Not
+ * all possible weirdnesses are checked for, just ones that were detected on
+ * actual Microsoft code pages */
+# define is_base_ALPHA(c) \
+ (isalpha((U8) (c)) && ! is_base_PUNCT(c))
+# define is_base_ALPHANUMERIC(c) \
+ (isalnum((U8) (c)) && ! is_base_PUNCT(c))
+# define is_base_CASED(c) \
+ ((isupper((U8) (c)) || islower((U8) (c))) && ! is_base_PUNCT(c))
+# define is_base_DIGIT(c) \
+ (isdigit((U8) (c)) && ! is_base_PUNCT(c))
+# define is_base_GRAPH(c) \
+ (isgraph((U8) (c)) && ! is_base_CNTRL(c))
+# define is_base_LOWER(c) \
+ (islower((U8) (c)) && ! is_base_PUNCT(c))
+# define is_base_PRINT(c) \
+ (isprint((U8) (c)) && ! is_base_CNTRL(c))
+# define is_base_PUNCT(c) \
+ (ispunct((U8) (c)) && ! is_base_CNTRL(c))
+# define is_base_UPPER(c) \
+ (isupper((U8) (c)) && ! is_base_PUNCT(c))
+# define is_base_XDIGIT(c) \
+ (isxdigit((U8) (c)) && ! is_base_PUNCT(c))
+#else
+
+/* For all other platforms, as far as we know, isdigit(), etc. work sanely
+ * enough */
+# define is_base_ALPHA(c) isalpha((U8) (c))
+# define is_base_ALPHANUMERIC(c) isalnum((U8) (c))
+# define is_base_CASED(c) (islower((U8) (c)) || isupper((U8) (c)))
+# define is_base_DIGIT(c) isdigit((U8) (c))
+
+ /* ... But it seems that IBM products treat NBSP as both a space and a
+ * graphic; these are the two platforms that we have active test beds for.
+ */
+# if defined(OS390) || defined(_AIX)
+# define is_base_GRAPH(c) (isgraph((U8) (c)) && ! isspace((U8) (c)))
# else
-# define isGRAPH_LC(c) _generic_LC(c, _CC_GRAPH, isgraph)
+# define is_base_GRAPH(c) isgraph((U8) (c))
# endif
-# define isIDFIRST_LC(c) _generic_LC_underscore(c, _CC_IDFIRST, isalpha)
-# define isLOWER_LC(c) _generic_LC(c, _CC_LOWER, islower)
-# define isPRINT_LC(c) _generic_LC(c, _CC_PRINT, isprint)
-# define isPUNCT_LC(c) _generic_LC(c, _CC_PUNCT, ispunct)
-# define isSPACE_LC(c) _generic_LC(c, _CC_SPACE, isspace)
-# define isUPPER_LC(c) _generic_LC(c, _CC_UPPER, isupper)
-# define isWORDCHAR_LC(c) _generic_LC_underscore(c, _CC_WORDCHAR, isalnum)
-# define isXDIGIT_LC(c) _generic_LC(c, _CC_XDIGIT, isxdigit)
-
-
-# define toLOWER_LC(c) _generic_toLOWER_LC((c), tolower, U8)
-# define toUPPER_LC(c) _generic_toUPPER_LC((c), toupper, U8)
-# define toFOLD_LC(c) _generic_toFOLD_LC((c), tolower, U8)
-
-#else /* The final fallback position */
-
-# define isALPHA_LC(c) (isascii(c) && isalpha(c))
-# define isALPHANUMERIC_LC(c) (isascii(c) && isalnum(c))
-# define isCNTRL_LC(c) (isascii(c) && iscntrl(c))
-# define isDIGIT_LC(c) (isascii(c) && isdigit(c))
-# define isGRAPH_LC(c) (isascii(c) && isgraph(c))
-# define isIDFIRST_LC(c) (isascii(c) && (isalpha(c) || (c) == '_'))
-# define isLOWER_LC(c) (isascii(c) && islower(c))
-# define isPRINT_LC(c) (isascii(c) && isprint(c))
-# define isPUNCT_LC(c) (isascii(c) && ispunct(c))
-# define isSPACE_LC(c) (isascii(c) && isspace(c))
-# define isUPPER_LC(c) (isascii(c) && isupper(c))
-# define isWORDCHAR_LC(c) (isascii(c) && (isalnum(c) || (c) == '_'))
-# define isXDIGIT_LC(c) (isascii(c) && isxdigit(c))
-
-# define toLOWER_LC(c) (isascii(c) ? tolower(c) : (c))
-# define toUPPER_LC(c) (isascii(c) ? toupper(c) : (c))
-# define toFOLD_LC(c) (isascii(c) ? tolower(c) : (c))
+# define is_base_LOWER(c) islower((U8) (c))
+# define is_base_PRINT(c) isprint((U8) (c))
+# define is_base_PUNCT(c) ispunct((U8) (c))
+# define is_base_UPPER(c) isupper((U8) (c))
+# define is_base_XDIGIT(c) isxdigit((U8) (c))
+#endif
+/* Below is the next level up, which currently expands to nothing more
+ * than the previous layer. These are the macros to use if you really need
+ * something whose input domain is a byte, and the locale isn't UTF-8; that is,
+ * where you normally would have to use things like bare isalnum().
+ *
+ * But most likely you should instead use the layer defined further below which
+ * has names like isALPHA_LC. They deal with larger-than-byte inputs, and
+ * UTF-8 locales.
+ *
+ * (Note, proper general operation of the bare libc functons requires you to
+ * cast to U8. These do that for you automatically.) */
+
+# define WRAP_U8_LC_(c, classnum, base) base(c)
+
+#define isU8_ALPHANUMERIC_LC(c) \
+ WRAP_U8_LC_((c), CC_ALPHANUMERIC_, is_base_ALPHANUMERIC)
+#define isU8_ALPHA_LC(c) WRAP_U8_LC_((c), CC_ALPHA_, is_base_ALPHA)
+#define isU8_ASCII_LC(c) WRAP_U8_LC_((c), CC_ASCII_, is_base_ASCII)
+#define isU8_BLANK_LC(c) WRAP_U8_LC_((c), CC_BLANK_, is_base_BLANK)
+#define isU8_CASED_LC(c) WRAP_U8_LC_((c), CC_CASED_, is_base_CASED)
+#define isU8_CNTRL_LC(c) WRAP_U8_LC_((c), CC_CNTRL_, is_base_CNTRL)
+#define isU8_DIGIT_LC(c) WRAP_U8_LC_((c), CC_DIGIT_, is_base_DIGIT)
+#define isU8_GRAPH_LC(c) WRAP_U8_LC_((c), CC_GRAPH_, is_base_GRAPH)
+#define isU8_IDFIRST_LC(c) WRAP_U8_LC_((c), CC_IDFIRST_, is_base_IDFIRST)
+#define isU8_LOWER_LC(c) WRAP_U8_LC_((c), CC_LOWER_, is_base_LOWER)
+#define isU8_PRINT_LC(c) WRAP_U8_LC_((c), CC_PRINT_, is_base_PRINT)
+#define isU8_PUNCT_LC(c) WRAP_U8_LC_((c), CC_PUNCT_, is_base_PUNCT)
+#define isU8_SPACE_LC(c) WRAP_U8_LC_((c), CC_SPACE_, is_base_SPACE)
+#define isU8_UPPER_LC(c) WRAP_U8_LC_((c), CC_UPPER_, is_base_UPPER)
+#define isU8_WORDCHAR_LC(c) WRAP_U8_LC_((c), CC_WORDCHAR_, is_base_WORDCHAR)
+#define isU8_XDIGIT_LC(c) WRAP_U8_LC_((c), CC_XDIGIT_, is_base_XDIGIT)
+
+#define toU8_LOWER_LC(c) WRAP_U8_LC_((c), CC_TOLOWER_, to_base_LOWER)
+#define toU8_UPPER_LC(c) WRAP_U8_LC_((c), CC_TOUPPER_, to_base_UPPER)
+#define toU8_FOLD_LC(c) toU8_LOWER_LC(c)
+
+/* The definitions below use the ones above to create versions in which the
+ * input domain isn't restricted to bytes (though always returning false if the
+ * input doesn't fit in a byte), and to behave properly should the locale be
+ * UTF-8. These are the documented ones, suitable for general use (though
+ * toUPPER_LC and toFOLD_LC aren't documented because they need special
+ * handling to deal with SHARP S expanding to two characters). */
+
+#define isASCII_LC(c) (FITS_IN_8_BITS(c) && isU8_ASCII_LC(c))
+#define isALPHA_LC(c) generic_LC_(c, CC_ALPHA_, isU8_ALPHA_LC)
+#define isALPHANUMERIC_LC(c) \
+ generic_LC_(c, CC_ALPHANUMERIC_, isU8_ALPHANUMERIC_LC)
+#define isBLANK_LC(c) generic_LC_(c, CC_BLANK_, isU8_BLANK_LC)
+#define isCASED_LC(c) generic_LC_(c, CC_CASED_, isU8_CASED_LC)
+#define isCNTRL_LC(c) generic_LC_(c, CC_CNTRL_, isU8_CNTRL_LC)
+#define isDIGIT_LC(c) generic_LC_(c, CC_DIGIT_, isU8_DIGIT_LC)
+#define isGRAPH_LC(c) generic_LC_(c, CC_GRAPH_, isU8_GRAPH_LC)
+#define isIDFIRST_LC(c) generic_LC_(c, CC_IDFIRST_, isU8_IDFIRST_LC)
+#define isLOWER_LC(c) generic_LC_(c, CC_LOWER_, isU8_LOWER_LC)
+#define isPRINT_LC(c) generic_LC_(c, CC_PRINT_, isU8_PRINT_LC)
+#define isPUNCT_LC(c) generic_LC_(c, CC_PUNCT_, isU8_PUNCT_LC)
+#define isSPACE_LC(c) generic_LC_(c, CC_SPACE_, isU8_SPACE_LC)
+#define isUPPER_LC(c) generic_LC_(c, CC_UPPER_, isU8_UPPER_LC)
+#define isWORDCHAR_LC(c) generic_LC_(c, CC_WORDCHAR_, isU8_WORDCHAR_LC)
+#define isXDIGIT_LC(c) generic_LC_(c, CC_XDIGIT_, isU8_XDIGIT_LC)
+
+#ifndef CTYPE256
+# define toLOWER_LC(c) toLOWER_A(c)
+# define toUPPER_LC(c) toUPPER_A(c)
+# define toFOLD_LC(c) toFOLD_A(c)
+#else
+
+/* In the next three macros, the reason for using the PL_latin arrays is in
+ * case the system function is defective; it ensures uniform results that
+ * conform to the Unicode standard. */
+
+/* This does not handle the anomalies in UTF-8 Turkic locales. */
+# define toLOWER_LC(c) ((! FITS_IN_8_BITS(c)) \
+ ? (c) \
+ : ((IN_UTF8_CTYPE_LOCALE) \
+ ? PL_latin1_lc[ (U8) (c) ] \
+ : ((U8) toU8_LOWER_LC(c))))
+
+/* In this macro, note that the result can be larger than a byte in a UTF-8
+ * locale. It returns a single value, so can't adequately return the upper
+ * case of LATIN SMALL LETTER SHARP S in a UTF-8 locale (which should be a
+ * string of two values "SS"); instead it asserts against that under
+ * DEBUGGING, and otherwise returns its input. It does not handle the
+ * anomalies in UTF-8 Turkic locales. */
+# define toUPPER_LC(c) \
+ ((! FITS_IN_8_BITS(c)) \
+ ? (c) \
+ : ((! IN_UTF8_CTYPE_LOCALE) \
+ ? ((U8) toU8_UPPER_LC(c)) \
+ : (UNLIKELY(((U8)(c)) == MICRO_SIGN) \
+ ? GREEK_CAPITAL_LETTER_MU \
+ : ((UNLIKELY(((U8) (c)) == LATIN_SMALL_LETTER_Y_WITH_DIAERESIS) \
+ ? LATIN_CAPITAL_LETTER_Y_WITH_DIAERESIS \
+ : (UNLIKELY(((U8)(c)) == LATIN_SMALL_LETTER_SHARP_S) \
+ ? (__ASSERT_(0) (c)) /* Fail on Sharp S in DEBUGGING */ \
+ : PL_mod_latin1_uc[ (U8) (c) ]))))))
+
+/* In this macro, note that the result can be larger than a byte in a UTF-8
+ * locale. It returns a single value, so can't adequately return the fold case
+ * of LATIN SMALL LETTER SHARP S in a UTF-8 locale (which should be a string of
+ * two values "ss"); instead it asserts against that under DEBUGGING, and
+ * otherwise returns its input. It does not handle the anomalies in UTF-8
+ * Turkic locales */
+# define toFOLD_LC(c) \
+ ((UNLIKELY((c) == MICRO_SIGN) && IN_UTF8_CTYPE_LOCALE) \
+ ? GREEK_SMALL_LETTER_MU \
+ : (__ASSERT_( ! IN_UTF8_CTYPE_LOCALE \
+ || LIKELY((c) != LATIN_SMALL_LETTER_SHARP_S)) \
+ toLOWER_LC(c)))
#endif
#define isIDCONT(c) isWORDCHAR(c)
/* For internal core Perl use only: the base macros for defining macros like
* isALPHA_uvchr. 'c' is the code point to check. 'classnum' is the POSIX class
- * number defined earlier in this file. _generic_uvchr() is used for POSIX
+ * number defined earlier in this file. generic_uvchr_() is used for POSIX
* classes where there is a macro or function 'above_latin1' that takes the
* single argument 'c' and returns the desired value. These exist for those
* classes which have simple definitions, avoiding the overhead of an inversion
- * list binary search. _generic_invlist_uvchr() can be used
+ * list binary search. generic_invlist_uvchr_() can be used
* for classes where that overhead is faster than a direct lookup.
- * _generic_uvchr() won't compile if 'c' isn't unsigned, as it won't match the
- * 'above_latin1' prototype. _generic_isCC() macro does bounds checking, so
+ * generic_uvchr_() won't compile if 'c' isn't unsigned, as it won't match the
+ * 'above_latin1' prototype. generic_isCC_() macro does 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. */
/* Note that all ignore 'use bytes' */
-#define _generic_uvchr(classnum, above_latin1, c) ((c) < 256 \
- ? _generic_isCC(c, classnum) \
+#define generic_uvchr_(classnum, above_latin1, c) ((c) < 256 \
+ ? generic_isCC_(c, classnum) \
: above_latin1(c))
-#define _generic_invlist_uvchr(classnum, c) ((c) < 256 \
- ? _generic_isCC(c, classnum) \
+#define generic_invlist_uvchr_(classnum, c) ((c) < 256 \
+ ? generic_isCC_(c, classnum) \
: _is_uni_FOO(classnum, c))
-#define isALPHA_uvchr(c) _generic_invlist_uvchr(_CC_ALPHA, c)
-#define isALPHANUMERIC_uvchr(c) _generic_invlist_uvchr(_CC_ALPHANUMERIC, c)
+#define isALPHA_uvchr(c) generic_invlist_uvchr_(CC_ALPHA_, c)
+#define isALPHANUMERIC_uvchr(c) generic_invlist_uvchr_(CC_ALPHANUMERIC_, c)
#define isASCII_uvchr(c) isASCII(c)
-#define isBLANK_uvchr(c) _generic_uvchr(_CC_BLANK, is_HORIZWS_cp_high, c)
+#define isBLANK_uvchr(c) generic_uvchr_(CC_BLANK_, is_HORIZWS_cp_high, c)
#define isCNTRL_uvchr(c) isCNTRL_L1(c) /* All controls are in Latin1 */
-#define isDIGIT_uvchr(c) _generic_invlist_uvchr(_CC_DIGIT, c)
-#define isGRAPH_uvchr(c) _generic_invlist_uvchr(_CC_GRAPH, c)
+#define isDIGIT_uvchr(c) generic_invlist_uvchr_(CC_DIGIT_, c)
+#define isGRAPH_uvchr(c) generic_invlist_uvchr_(CC_GRAPH_, c)
#define isIDCONT_uvchr(c) \
- _generic_uvchr(_CC_WORDCHAR, _is_uni_perl_idcont, c)
+ generic_uvchr_(CC_WORDCHAR_, _is_uni_perl_idcont, c)
#define isIDFIRST_uvchr(c) \
- _generic_uvchr(_CC_IDFIRST, _is_uni_perl_idstart, c)
-#define isLOWER_uvchr(c) _generic_invlist_uvchr(_CC_LOWER, c)
-#define isPRINT_uvchr(c) _generic_invlist_uvchr(_CC_PRINT, c)
+ generic_uvchr_(CC_IDFIRST_, _is_uni_perl_idstart, c)
+#define isLOWER_uvchr(c) generic_invlist_uvchr_(CC_LOWER_, c)
+#define isPRINT_uvchr(c) generic_invlist_uvchr_(CC_PRINT_, c)
-#define isPUNCT_uvchr(c) _generic_invlist_uvchr(_CC_PUNCT, c)
-#define isSPACE_uvchr(c) _generic_uvchr(_CC_SPACE, is_XPERLSPACE_cp_high, c)
+#define isPUNCT_uvchr(c) generic_invlist_uvchr_(CC_PUNCT_, c)
+#define isSPACE_uvchr(c) generic_uvchr_(CC_SPACE_, is_XPERLSPACE_cp_high, c)
#define isPSXSPC_uvchr(c) isSPACE_uvchr(c)
-#define isUPPER_uvchr(c) _generic_invlist_uvchr(_CC_UPPER, c)
-#define isVERTWS_uvchr(c) _generic_uvchr(_CC_VERTSPACE, is_VERTWS_cp_high, c)
-#define isWORDCHAR_uvchr(c) _generic_invlist_uvchr(_CC_WORDCHAR, c)
-#define isXDIGIT_uvchr(c) _generic_uvchr(_CC_XDIGIT, is_XDIGIT_cp_high, c)
+#define isUPPER_uvchr(c) generic_invlist_uvchr_(CC_UPPER_, c)
+#define isVERTWS_uvchr(c) generic_uvchr_(CC_VERTSPACE_, is_VERTWS_cp_high, c)
+#define isWORDCHAR_uvchr(c) generic_invlist_uvchr_(CC_WORDCHAR_, c)
+#define isXDIGIT_uvchr(c) generic_uvchr_(CC_XDIGIT_, is_XDIGIT_cp_high, c)
#define toFOLD_uvchr(c,s,l) to_uni_fold(c,s,l)
#define toLOWER_uvchr(c,s,l) to_uni_lower(c,s,l)
/* For internal core Perl use only: the base macros for defining macros like
* isALPHA_LC_uvchr. These are like isALPHA_LC, but the input can be any code
- * point, not just 0-255. Like _generic_uvchr, there are two versions, one for
+ * point, not just 0-255. Like generic_uvchr_, there are two versions, one for
* simple class definitions; the other for more complex. These are like
- * _generic_uvchr, so see it for more info. */
-#define _generic_LC_uvchr(latin1, above_latin1, c) \
+ * generic_uvchr_, so see it for more info. */
+#define generic_LC_uvchr_(latin1, above_latin1, c) \
(c < 256 ? latin1(c) : above_latin1(c))
-#define _generic_LC_invlist_uvchr(latin1, classnum, c) \
+#define generic_LC_invlist_uvchr_(latin1, classnum, c) \
(c < 256 ? latin1(c) : _is_uni_FOO(classnum, c))
-#define isALPHA_LC_uvchr(c) _generic_LC_invlist_uvchr(isALPHA_LC, _CC_ALPHA, c)
-#define isALPHANUMERIC_LC_uvchr(c) _generic_LC_invlist_uvchr(isALPHANUMERIC_LC, \
- _CC_ALPHANUMERIC, c)
+#define isALPHA_LC_uvchr(c) generic_LC_invlist_uvchr_(isALPHA_LC, CC_ALPHA_, c)
+#define isALPHANUMERIC_LC_uvchr(c) generic_LC_invlist_uvchr_(isALPHANUMERIC_LC, \
+ CC_ALPHANUMERIC_, c)
#define isASCII_LC_uvchr(c) isASCII_LC(c)
-#define isBLANK_LC_uvchr(c) _generic_LC_uvchr(isBLANK_LC, \
+#define isBLANK_LC_uvchr(c) generic_LC_uvchr_(isBLANK_LC, \
is_HORIZWS_cp_high, c)
#define isCNTRL_LC_uvchr(c) (c < 256 ? isCNTRL_LC(c) : 0)
-#define isDIGIT_LC_uvchr(c) _generic_LC_invlist_uvchr(isDIGIT_LC, _CC_DIGIT, c)
-#define isGRAPH_LC_uvchr(c) _generic_LC_invlist_uvchr(isGRAPH_LC, _CC_GRAPH, c)
-#define isIDCONT_LC_uvchr(c) _generic_LC_uvchr(isIDCONT_LC, \
+#define isDIGIT_LC_uvchr(c) generic_LC_invlist_uvchr_(isDIGIT_LC, CC_DIGIT_, c)
+#define isGRAPH_LC_uvchr(c) generic_LC_invlist_uvchr_(isGRAPH_LC, CC_GRAPH_, c)
+#define isIDCONT_LC_uvchr(c) generic_LC_uvchr_(isIDCONT_LC, \
_is_uni_perl_idcont, c)
-#define isIDFIRST_LC_uvchr(c) _generic_LC_uvchr(isIDFIRST_LC, \
+#define isIDFIRST_LC_uvchr(c) generic_LC_uvchr_(isIDFIRST_LC, \
_is_uni_perl_idstart, c)
-#define isLOWER_LC_uvchr(c) _generic_LC_invlist_uvchr(isLOWER_LC, _CC_LOWER, c)
-#define isPRINT_LC_uvchr(c) _generic_LC_invlist_uvchr(isPRINT_LC, _CC_PRINT, c)
+#define isLOWER_LC_uvchr(c) generic_LC_invlist_uvchr_(isLOWER_LC, CC_LOWER_, c)
+#define isPRINT_LC_uvchr(c) generic_LC_invlist_uvchr_(isPRINT_LC, CC_PRINT_, c)
#define isPSXSPC_LC_uvchr(c) isSPACE_LC_uvchr(c)
-#define isPUNCT_LC_uvchr(c) _generic_LC_invlist_uvchr(isPUNCT_LC, _CC_PUNCT, c)
-#define isSPACE_LC_uvchr(c) _generic_LC_uvchr(isSPACE_LC, \
+#define isPUNCT_LC_uvchr(c) generic_LC_invlist_uvchr_(isPUNCT_LC, CC_PUNCT_, c)
+#define isSPACE_LC_uvchr(c) generic_LC_uvchr_(isSPACE_LC, \
is_XPERLSPACE_cp_high, c)
-#define isUPPER_LC_uvchr(c) _generic_LC_invlist_uvchr(isUPPER_LC, _CC_UPPER, c)
-#define isWORDCHAR_LC_uvchr(c) _generic_LC_invlist_uvchr(isWORDCHAR_LC, \
- _CC_WORDCHAR, c)
-#define isXDIGIT_LC_uvchr(c) _generic_LC_uvchr(isXDIGIT_LC, \
+#define isUPPER_LC_uvchr(c) generic_LC_invlist_uvchr_(isUPPER_LC, CC_UPPER_, c)
+#define isWORDCHAR_LC_uvchr(c) generic_LC_invlist_uvchr_(isWORDCHAR_LC, \
+ 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))
# define _utf8_safe_assert(p,e) ((e) > (p))
#endif
-#define _generic_utf8_safe(classnum, p, e, above_latin1) \
+#define generic_utf8_safe_(classnum, p, e, above_latin1) \
((! _utf8_safe_assert(p, e)) \
? (_force_out_malformed_utf8_message((U8 *) (p), (U8 *) (e), 0, 1), 0)\
: (UTF8_IS_INVARIANT(*(p))) \
- ? _generic_isCC(*(p), classnum) \
+ ? generic_isCC_(*(p), classnum) \
: (UTF8_IS_DOWNGRADEABLE_START(*(p)) \
? ((LIKELY((e) - (p) > 1 && UTF8_IS_CONTINUATION(*((p)+1)))) \
- ? _generic_isCC(EIGHT_BIT_UTF8_TO_NATIVE(*(p), *((p)+1 )), \
+ ? generic_isCC_(EIGHT_BIT_UTF8_TO_NATIVE(*(p), *((p)+1 )), \
classnum) \
: (_force_out_malformed_utf8_message( \
(U8 *) (p), (U8 *) (e), 0, 1), 0)) \
: above_latin1))
/* Like the above, but calls 'above_latin1(p)' to get the utf8 value.
* 'above_latin1' can be a macro */
-#define _generic_func_utf8_safe(classnum, above_latin1, p, e) \
- _generic_utf8_safe(classnum, p, e, above_latin1(p, e))
-#define _generic_non_invlist_utf8_safe(classnum, above_latin1, p, e) \
- _generic_utf8_safe(classnum, p, e, \
+#define generic_func_utf8_safe_(classnum, above_latin1, p, e) \
+ generic_utf8_safe_(classnum, p, e, above_latin1(p, e))
+#define generic_non_invlist_utf8_safe_(classnum, above_latin1, p, e) \
+ generic_utf8_safe_(classnum, p, e, \
(UNLIKELY((e) - (p) < UTF8SKIP(p)) \
? (_force_out_malformed_utf8_message( \
(U8 *) (p), (U8 *) (e), 0, 1), 0) \
: above_latin1(p)))
/* Like the above, but passes classnum to _isFOO_utf8(), instead of having an
* 'above_latin1' parameter */
-#define _generic_invlist_utf8_safe(classnum, p, e) \
- _generic_utf8_safe(classnum, p, e, _is_utf8_FOO(classnum, p, e))
+#define generic_invlist_utf8_safe_(classnum, p, e) \
+ generic_utf8_safe_(classnum, p, e, _is_utf8_FOO(classnum, p, e))
/* Like the above, but should be used only when it is known that there are no
* characters in the upper-Latin1 range (128-255 on ASCII platforms) which the
* class is TRUE for. Hence it can skip the tests for this range.
* 'above_latin1' should include its arguments */
-#define _generic_utf8_safe_no_upper_latin1(classnum, p, e, above_latin1) \
+#define generic_utf8_safe_no_upper_latin1_(classnum, p, e, above_latin1) \
(__ASSERT_(_utf8_safe_assert(p, e)) \
(isASCII(*(p))) \
- ? _generic_isCC(*(p), classnum) \
+ ? generic_isCC_(*(p), classnum) \
: (UTF8_IS_DOWNGRADEABLE_START(*(p))) \
? 0 /* Note that doesn't check validity for latin1 */ \
: above_latin1)
#define isWORDCHAR_utf8(p, e) isWORDCHAR_utf8_safe(p, e)
#define isXDIGIT_utf8(p, e) isXDIGIT_utf8_safe(p, e)
-#define isALPHA_utf8_safe(p, e) _generic_invlist_utf8_safe(_CC_ALPHA, p, e)
+#define isALPHA_utf8_safe(p, e) generic_invlist_utf8_safe_(CC_ALPHA_, p, e)
#define isALPHANUMERIC_utf8_safe(p, e) \
- _generic_invlist_utf8_safe(_CC_ALPHANUMERIC, p, e)
+ generic_invlist_utf8_safe_(CC_ALPHANUMERIC_, p, e)
#define isASCII_utf8_safe(p, e) \
/* Because ASCII is invariant under utf8, the non-utf8 macro \
* works */ \
(__ASSERT_(_utf8_safe_assert(p, e)) isASCII(*(p)))
#define isBLANK_utf8_safe(p, e) \
- _generic_non_invlist_utf8_safe(_CC_BLANK, is_HORIZWS_high, p, e)
+ generic_non_invlist_utf8_safe_(CC_BLANK_, is_HORIZWS_high, p, e)
#ifdef EBCDIC
/* Because all controls are UTF-8 invariants in EBCDIC, we can use this
# define isCNTRL_utf8_safe(p, e) \
(__ASSERT_(_utf8_safe_assert(p, e)) isCNTRL_L1(*(p)))
#else
-# define isCNTRL_utf8_safe(p, e) _generic_utf8_safe(_CC_CNTRL, p, e, 0)
+# define isCNTRL_utf8_safe(p, e) generic_utf8_safe_(CC_CNTRL_, p, e, 0)
#endif
#define isDIGIT_utf8_safe(p, e) \
- _generic_utf8_safe_no_upper_latin1(_CC_DIGIT, p, e, \
- _is_utf8_FOO(_CC_DIGIT, p, e))
-#define isGRAPH_utf8_safe(p, e) _generic_invlist_utf8_safe(_CC_GRAPH, p, e)
-#define isIDCONT_utf8_safe(p, e) _generic_func_utf8_safe(_CC_WORDCHAR, \
+ generic_utf8_safe_no_upper_latin1_(CC_DIGIT_, p, e, \
+ _is_utf8_FOO(CC_DIGIT_, p, e))
+#define isGRAPH_utf8_safe(p, e) generic_invlist_utf8_safe_(CC_GRAPH_, p, e)
+#define isIDCONT_utf8_safe(p, e) generic_func_utf8_safe_(CC_WORDCHAR_, \
_is_utf8_perl_idcont, p, e)
/* To prevent S_scan_word in toke.c from hanging, we have to make sure that
* This used to be not the XID version, but we decided to go with the more
* modern Unicode definition */
#define isIDFIRST_utf8_safe(p, e) \
- _generic_func_utf8_safe(_CC_IDFIRST, \
+ generic_func_utf8_safe_(CC_IDFIRST_, \
_is_utf8_perl_idstart, (U8 *) (p), (U8 *) (e))
-#define isLOWER_utf8_safe(p, e) _generic_invlist_utf8_safe(_CC_LOWER, p, e)
-#define isPRINT_utf8_safe(p, e) _generic_invlist_utf8_safe(_CC_PRINT, p, e)
+#define isLOWER_utf8_safe(p, e) generic_invlist_utf8_safe_(CC_LOWER_, p, e)
+#define isPRINT_utf8_safe(p, e) generic_invlist_utf8_safe_(CC_PRINT_, p, e)
#define isPSXSPC_utf8_safe(p, e) isSPACE_utf8_safe(p, e)
-#define isPUNCT_utf8_safe(p, e) _generic_invlist_utf8_safe(_CC_PUNCT, p, e)
+#define isPUNCT_utf8_safe(p, e) generic_invlist_utf8_safe_(CC_PUNCT_, p, e)
#define isSPACE_utf8_safe(p, e) \
- _generic_non_invlist_utf8_safe(_CC_SPACE, is_XPERLSPACE_high, p, e)
-#define isUPPER_utf8_safe(p, e) _generic_invlist_utf8_safe(_CC_UPPER, p, e)
+ generic_non_invlist_utf8_safe_(CC_SPACE_, is_XPERLSPACE_high, p, e)
+#define isUPPER_utf8_safe(p, e) generic_invlist_utf8_safe_(CC_UPPER_, p, e)
#define isVERTWS_utf8_safe(p, e) \
- _generic_non_invlist_utf8_safe(_CC_VERTSPACE, is_VERTWS_high, p, e)
+ generic_non_invlist_utf8_safe_(CC_VERTSPACE_, is_VERTWS_high, p, e)
#define isWORDCHAR_utf8_safe(p, e) \
- _generic_invlist_utf8_safe(_CC_WORDCHAR, p, e)
+ generic_invlist_utf8_safe_(CC_WORDCHAR_, p, e)
#define isXDIGIT_utf8_safe(p, e) \
- _generic_utf8_safe_no_upper_latin1(_CC_XDIGIT, p, e, \
+ generic_utf8_safe_no_upper_latin1_(CC_XDIGIT_, p, e, \
(UNLIKELY((e) - (p) < UTF8SKIP(p)) \
? (_force_out_malformed_utf8_message( \
(U8 *) (p), (U8 *) (e), 0, 1), 0) \
#define isXDIGIT_LC_utf8(p, e) isXDIGIT_LC_utf8_safe(p, e)
/* For internal core Perl use only: the base macros for defining macros like
- * isALPHA_LC_utf8_safe. These are like _generic_utf8, but if the first code
+ * isALPHA_LC_utf8_safe. These are like generic_utf8_, but if the first code
* point in 'p' is within the 0-255 range, it uses locale rules from the
* passed-in 'macro' parameter */
-#define _generic_LC_utf8_safe(macro, p, e, above_latin1) \
+#define generic_LC_utf8_safe_(macro, p, e, above_latin1) \
(__ASSERT_(_utf8_safe_assert(p, e)) \
(UTF8_IS_INVARIANT(*(p))) \
? macro(*(p)) \
(U8 *) (p), (U8 *) (e), 0, 1), 0)) \
: above_latin1))
-#define _generic_LC_invlist_utf8_safe(macro, classnum, p, e) \
- _generic_LC_utf8_safe(macro, p, e, \
+#define generic_LC_invlist_utf8_safe_(macro, classnum, p, e) \
+ generic_LC_utf8_safe_(macro, p, e, \
_is_utf8_FOO(classnum, p, e))
-#define _generic_LC_func_utf8_safe(macro, above_latin1, p, e) \
- _generic_LC_utf8_safe(macro, p, e, above_latin1(p, e))
+#define generic_LC_func_utf8_safe_(macro, above_latin1, p, e) \
+ generic_LC_utf8_safe_(macro, p, e, above_latin1(p, e))
-#define _generic_LC_non_invlist_utf8_safe(classnum, above_latin1, p, e) \
- _generic_LC_utf8_safe(classnum, p, e, \
+#define generic_LC_non_invlist_utf8_safe_(classnum, above_latin1, p, e) \
+ generic_LC_utf8_safe_(classnum, p, e, \
(UNLIKELY((e) - (p) < UTF8SKIP(p)) \
? (_force_out_malformed_utf8_message( \
(U8 *) (p), (U8 *) (e), 0, 1), 0) \
: above_latin1(p)))
#define isALPHANUMERIC_LC_utf8_safe(p, e) \
- _generic_LC_invlist_utf8_safe(isALPHANUMERIC_LC, \
- _CC_ALPHANUMERIC, p, e)
+ generic_LC_invlist_utf8_safe_(isALPHANUMERIC_LC, \
+ CC_ALPHANUMERIC_, p, e)
#define isALPHA_LC_utf8_safe(p, e) \
- _generic_LC_invlist_utf8_safe(isALPHA_LC, _CC_ALPHA, p, e)
+ generic_LC_invlist_utf8_safe_(isALPHA_LC, CC_ALPHA_, p, e)
#define isASCII_LC_utf8_safe(p, e) \
(__ASSERT_(_utf8_safe_assert(p, e)) isASCII_LC(*(p)))
#define isBLANK_LC_utf8_safe(p, e) \
- _generic_LC_non_invlist_utf8_safe(isBLANK_LC, is_HORIZWS_high, p, e)
+ generic_LC_non_invlist_utf8_safe_(isBLANK_LC, is_HORIZWS_high, p, e)
#define isCNTRL_LC_utf8_safe(p, e) \
- _generic_LC_utf8_safe(isCNTRL_LC, p, e, 0)
+ generic_LC_utf8_safe_(isCNTRL_LC, p, e, 0)
#define isDIGIT_LC_utf8_safe(p, e) \
- _generic_LC_invlist_utf8_safe(isDIGIT_LC, _CC_DIGIT, p, e)
+ generic_LC_invlist_utf8_safe_(isDIGIT_LC, CC_DIGIT_, p, e)
#define isGRAPH_LC_utf8_safe(p, e) \
- _generic_LC_invlist_utf8_safe(isGRAPH_LC, _CC_GRAPH, p, e)
+ generic_LC_invlist_utf8_safe_(isGRAPH_LC, CC_GRAPH_, p, e)
#define isIDCONT_LC_utf8_safe(p, e) \
- _generic_LC_func_utf8_safe(isIDCONT_LC, \
+ generic_LC_func_utf8_safe_(isIDCONT_LC, \
_is_utf8_perl_idcont, p, e)
#define isIDFIRST_LC_utf8_safe(p, e) \
- _generic_LC_func_utf8_safe(isIDFIRST_LC, \
+ generic_LC_func_utf8_safe_(isIDFIRST_LC, \
_is_utf8_perl_idstart, p, e)
#define isLOWER_LC_utf8_safe(p, e) \
- _generic_LC_invlist_utf8_safe(isLOWER_LC, _CC_LOWER, p, e)
+ generic_LC_invlist_utf8_safe_(isLOWER_LC, CC_LOWER_, p, e)
#define isPRINT_LC_utf8_safe(p, e) \
- _generic_LC_invlist_utf8_safe(isPRINT_LC, _CC_PRINT, p, e)
+ generic_LC_invlist_utf8_safe_(isPRINT_LC, CC_PRINT_, p, e)
#define isPSXSPC_LC_utf8_safe(p, e) isSPACE_LC_utf8_safe(p, e)
#define isPUNCT_LC_utf8_safe(p, e) \
- _generic_LC_invlist_utf8_safe(isPUNCT_LC, _CC_PUNCT, p, e)
+ generic_LC_invlist_utf8_safe_(isPUNCT_LC, CC_PUNCT_, p, e)
#define isSPACE_LC_utf8_safe(p, e) \
- _generic_LC_non_invlist_utf8_safe(isSPACE_LC, is_XPERLSPACE_high, p, e)
+ generic_LC_non_invlist_utf8_safe_(isSPACE_LC, is_XPERLSPACE_high, p, e)
#define isUPPER_LC_utf8_safe(p, e) \
- _generic_LC_invlist_utf8_safe(isUPPER_LC, _CC_UPPER, p, e)
+ generic_LC_invlist_utf8_safe_(isUPPER_LC, CC_UPPER_, p, e)
#define isWORDCHAR_LC_utf8_safe(p, e) \
- _generic_LC_invlist_utf8_safe(isWORDCHAR_LC, _CC_WORDCHAR, p, e)
+ generic_LC_invlist_utf8_safe_(isWORDCHAR_LC, CC_WORDCHAR_, p, e)
#define isXDIGIT_LC_utf8_safe(p, e) \
- _generic_LC_non_invlist_utf8_safe(isXDIGIT_LC, is_XDIGIT_high, p, e)
+ generic_LC_non_invlist_utf8_safe_(isXDIGIT_LC, is_XDIGIT_high, p, e)
/* Macros for backwards compatibility and for completeness when the ASCII and
* Latin1 values are identical */
Line numbers are unsigned, 32 bits.
*/
typedef U32 line_t;
+#define LINE_Tf U32uf
#define NOLINE ((line_t) 4294967295UL) /* = FFFFFFFF */
/* Helpful alias for version prescan */
MLT_DEL_SV
};
# endif
-# if defined(PERL_IN_SV_C) /* those are only used in sv.c */
-void Perl_mem_log_new_sv(const SV *sv, const char *filename, const int linenumber, const char *funcname);
-void Perl_mem_log_del_sv(const SV *sv, const char *filename, const int linenumber, const char *funcname);
-# endif
# endif
#endif
#define C_ARRAY_LENGTH(a) (sizeof(a)/sizeof((a)[0]))
#define C_ARRAY_END(a) ((a) + C_ARRAY_LENGTH(a))
+#if defined(PERL_CORE) || defined(PERL_EXT_RE_BUILD)
+/* strlen() of a literal string constant. Restricting this to core, in part
+ * because it can generate compiler warnings about comparing unlike signs */
+# define STRLENs(s) (sizeof("" s "") - 1)
+#endif
+
#ifdef NEED_VA_COPY
# ifdef va_copy
# define Perl_va_copy(s, d) va_copy(d, s)
#endif
+/* These are simple Marsaglia XOR-SHIFT RNG's for 64 and 32 bits. These
+ * RNG's are of reasonable quality, very fast, and have the interesting
+ * property that provided 'x' is non-zero they create a cycle of 2^32-1
+ * or 2^64-1 "random" like numbers, with the exception of 0. Thus they
+ * are very useful when you want an integer to "dance" in a random way,
+ * but you also never want it to become 0 and thus false.
+ *
+ * Obviously they leave x unchanged if it starts out as 0.
+ *
+ * We have two variants just because that can be helpful in certain
+ * places. There is no advantage to either, they are equally bad as each
+ * other as far RNG's go. Sufficiently random for many purposes, but
+ * insufficiently random for serious use as they fail important tests in
+ * the Test01 BigCrush RNG test suite by L’Ecuyer and Simard. (Note
+ * that Drand48 also fails BigCrush). The main point is they produce
+ * different sequences and in places where we want some randomlike
+ * behavior they are cheap and easy.
+ *
+ * Marsaglia was one of the early researchers into RNG testing and wrote
+ * the Diehard RNG test suite, which after his death become the
+ * Dieharder RNG suite, and was generally supplanted by the Test01 suite
+ * by L'Ecruyer and associates.
+ *
+ * There are dozens of shift parameters that create a pseudo random ring
+ * of integers 1..2^N-1, if you need a different sequence just read the
+ * paper and select a set of parameters. In fact, simply reversing the
+ * shift order from L/R/L to R/L/R should result in another valid
+ * example, but read the paper before you do that.
+ *
+ * PDF of the original paper:
+ * https://www.jstatsoft.org/article/download/v008i14/916
+ * Wikipedia:
+ * https://en.wikipedia.org/wiki/Xorshift
+ * Criticism:
+ * https://www.iro.umontreal.ca/~lecuyer/myftp/papers/xorshift.pdf
+ * Test01:
+ * http://simul.iro.umontreal.ca/testu01/tu01.html
+ * Diehard:
+ * https://en.wikipedia.org/wiki/Diehard_tests
+ * Dieharder:
+ * https://webhome.phy.duke.edu/~rgb/General/rand_rate/rand_rate.abs
+ *
+ */
+
+/* 32 bit version */
+#define PERL_XORSHIFT32_A(x) \
+STMT_START { \
+ (x) ^= ((x) << 13); \
+ (x) ^= ((x) >> 17); \
+ (x) ^= ((x) << 5); \
+} STMT_END
+
+/* 64 bit version */
+#define PERL_XORSHIFT64_A(x) \
+STMT_START { \
+ (x) ^= ((x) << 13); \
+ (x) ^= ((x) >> 7); \
+ (x) ^= ((x) << 17); \
+} STMT_END
+
+/* 32 bit version */
+#define PERL_XORSHIFT32_B(x) \
+STMT_START { \
+ (x) ^= ((x) << 5); \
+ (x) ^= ((x) >> 27); \
+ (x) ^= ((x) << 8); \
+} STMT_END
+
+/* 64 bit version - currently this is unused,
+ * it is provided here to complement the 32 bit _B
+ * variant which IS used. */
+#define PERL_XORSHIFT64_B(x) \
+STMT_START { \
+ (x) ^= ((x) << 15); \
+ (x) ^= ((x) >> 49); \
+ (x) ^= ((x) << 26); \
+} STMT_END
+
+
#endif /* PERL_HANDY_H_ */
/*
https://perl5.git.perl.org / perl5.git / blobdiff