handy.h: Make some macros more time efficient
authorKarl Williamson <public@khwilliamson.com>
Mon, 3 Dec 2012 03:09:46 +0000 (20:09 -0700)
committerKarl Williamson <public@khwilliamson.com>
Sun, 9 Dec 2012 17:30:00 +0000 (10:30 -0700)
These macros check if a UTF-8 encoded character is of particular types
for use with locales.  Prior to this patch, they called a function to
convert the character to a code point value.  This was used as input to
another macro that handles code points.  For values above the Latin1
range, that macro calls a function, which converts back to UTF-8 and
calls another function.

This commit changes that to call the UTF-8 function directly for
above-Latin1 code points.  No conversion need be done.  For Latin1 code
points, it converts, if necessary, to the code point and calls a macro
that handles these directly.

Some of these macros now use a macro instead of a function call for
above-Latin1 code points, as is done in various other places in this
file.

handy.h

index 5636bcc..6819164 100644 (file)
--- a/handy.h
+++ b/handy.h
@@ -1028,19 +1028,35 @@ EXTCONST U32 PL_charclass[];
  * Latin1 */
 #define isPSXSPC_utf8(p)        _generic_utf8(_CC_PSXSPC, is_XPERLSPACE_high, p)
 
-#define isALNUM_LC_utf8(p)     isALNUM_LC_uvchr(valid_utf8_to_uvchr(p,  0))
-#define isIDFIRST_LC_utf8(p)   isIDFIRST_LC_uvchr(valid_utf8_to_uvchr(p,  0))
-#define isALPHA_LC_utf8(p)     isALPHA_LC_uvchr(valid_utf8_to_uvchr(p,  0))
-#define isBLANK_LC_utf8(p)     isBLANK_LC_uvchr(valid_utf8_to_uvchr(p,  0))
-#define isSPACE_LC_utf8(p)     isSPACE_LC_uvchr(valid_utf8_to_uvchr(p,  0))
-#define isDIGIT_LC_utf8(p)     isDIGIT_LC_uvchr(valid_utf8_to_uvchr(p,  0))
-#define isUPPER_LC_utf8(p)     isUPPER_LC_uvchr(valid_utf8_to_uvchr(p,  0))
-#define isLOWER_LC_utf8(p)     isLOWER_LC_uvchr(valid_utf8_to_uvchr(p,  0))
-#define isALNUMC_LC_utf8(p)    isALNUMC_LC_uvchr(valid_utf8_to_uvchr(p,  0))
-#define isCNTRL_LC_utf8(p)     isCNTRL_LC_uvchr(valid_utf8_to_uvchr(p,  0))
-#define isGRAPH_LC_utf8(p)     isGRAPH_LC_uvchr(valid_utf8_to_uvchr(p,  0))
-#define isPRINT_LC_utf8(p)     isPRINT_LC_uvchr(valid_utf8_to_uvchr(p,  0))
-#define isPUNCT_LC_utf8(p)     isPUNCT_LC_uvchr(valid_utf8_to_uvchr(p,  0))
+/* For internal core Perl use only.  If the input is in the Latin1 range, use
+ * the macro 'macro' on 'p' which is a pointer to a UTF-8 string.  Otherwise
+ * use the value given by the 'utf8' parameter.  This relies on the fact that
+ * ASCII characters have the same representation whether utf8 or not.  Note
+ * that it assumes that the utf8 has been validated, and ignores 'use bytes' */
+#define _generic_LC_utf8_utf8(macro, p, utf8) \
+                         (UTF8_IS_INVARIANT(*(p))                          \
+                         ? macro(*(p))                                     \
+                         : (UTF8_IS_DOWNGRADEABLE_START(*(p)))             \
+                           ? macro(TWO_BYTE_UTF8_TO_UNI(*(p), *((p)+1)))   \
+                           : utf8)
+
+#define _generic_LC_utf8(macro, utf8_func, p)                              \
+                            _generic_LC_utf8_utf8(macro, p, utf8_func(p))
+
+#define isALNUM_LC_utf8(p)   _generic_LC_utf8(isALNUM_LC, is_utf8_alnum, p)
+#define isIDFIRST_LC_utf8(p) _generic_LC_utf8(isIDFIRST_LC,                \
+                                                    _is_utf8_perl_idstart, p)
+#define isALPHA_LC_utf8(p)   _generic_LC_utf8(isALPHA_LC, is_utf8_alpha, p)
+#define isBLANK_LC_utf8(p)   _generic_LC_utf8(isBLANK_LC, is_HORIZWS_high, p)
+#define isSPACE_LC_utf8(p)   _generic_LC_utf8(isSPACE_LC, is_XPERLSPACE_high, p)
+#define isDIGIT_LC_utf8(p)   _generic_LC_utf8(isDIGIT_LC, is_utf8_digit, p)
+#define isUPPER_LC_utf8(p)   _generic_LC_utf8(isUPPER_LC, is_utf8_upper, p)
+#define isLOWER_LC_utf8(p)   _generic_LC_utf8(isLOWER_LC, is_utf8_lower, p)
+#define isALNUMC_LC_utf8(p)  _generic_LC_utf8(isALNUMC_LC, is_utf8_alnumc, p)
+#define isCNTRL_LC_utf8(p)   _generic_LC_utf8_utf8(isCNTRL_LC, p, 0)
+#define isGRAPH_LC_utf8(p)   _generic_LC_utf8(isGRAPH_LC, is_utf8_graph, p)
+#define isPRINT_LC_utf8(p)   _generic_LC_utf8(isPRINT_LC, is_utf8_print, p)
+#define isPUNCT_LC_utf8(p)   _generic_LC_utf8(isPUNCT_LC, is_utf8_punct, p)
 
 #define isPSXSPC_LC_utf8(c)    (isSPACE_LC_utf8(c) ||(c) == '\f')