diff options
-rw-r--r-- | generic/tcl.decls | 2 | ||||
-rw-r--r-- | generic/tclCmdMZ.c | 4 | ||||
-rw-r--r-- | generic/tclDecls.h | 6 | ||||
-rw-r--r-- | generic/tclInt.h | 28 | ||||
-rw-r--r-- | generic/tclStringObj.c | 2 | ||||
-rw-r--r-- | generic/tclStubInit.c | 15 | ||||
-rw-r--r-- | generic/tclUtf.c | 34 |
7 files changed, 46 insertions, 45 deletions
diff --git a/generic/tcl.decls b/generic/tcl.decls index 8e21b1d..4a637ad 100644 --- a/generic/tcl.decls +++ b/generic/tcl.decls @@ -1248,7 +1248,7 @@ declare 352 { int Tcl_Char16Len(const unsigned short *uniStr) } declare 353 {deprecated {Use Tcl_UtfNcmp}} { - int Tcl_UniCharNcmp(const Tcl_UniChar *ucs, const Tcl_UniChar *uct, + int Tcl_UniCharNcmp(const unsigned short *ucs, const unsigned short *uct, unsigned long numChars) } declare 354 { diff --git a/generic/tclCmdMZ.c b/generic/tclCmdMZ.c index f394035..5e95217 100644 --- a/generic/tclCmdMZ.c +++ b/generic/tclCmdMZ.c @@ -604,7 +604,7 @@ Tcl_RegsubObjCmd( numMatches = 0; nocase = (cflags & TCL_REG_NOCASE); - strCmpFn = nocase ? Tcl_UniCharNcasecmp : Tcl_UniCharNcmp; + strCmpFn = nocase ? Tcl_UniCharNcasecmp : TclUniCharNcmp; wsrc = Tcl_GetUnicodeFromObj(objv[0], &slen); wstring = Tcl_GetUnicodeFromObj(objv[1], &wlen); @@ -2070,7 +2070,7 @@ StringMapCmd( } end = ustring1 + length1; - strCmpFn = (nocase ? Tcl_UniCharNcasecmp : Tcl_UniCharNcmp); + strCmpFn = (nocase ? Tcl_UniCharNcasecmp : TclUniCharNcmp); /* * Force result to be Unicode diff --git a/generic/tclDecls.h b/generic/tclDecls.h index e84a7e8..49ce21d 100644 --- a/generic/tclDecls.h +++ b/generic/tclDecls.h @@ -1066,8 +1066,8 @@ EXTERN int Tcl_UniCharIsWordChar(int ch); EXTERN int Tcl_Char16Len(const unsigned short *uniStr); /* 353 */ TCL_DEPRECATED("Use Tcl_UtfNcmp") -int Tcl_UniCharNcmp(const Tcl_UniChar *ucs, - const Tcl_UniChar *uct, +int Tcl_UniCharNcmp(const unsigned short *ucs, + const unsigned short *uct, unsigned long numChars); /* 354 */ EXTERN char * Tcl_Char16ToUtfDString(const unsigned short *uniStr, @@ -2345,7 +2345,7 @@ typedef struct TclStubs { int (*tcl_UniCharIsUpper) (int ch); /* 350 */ int (*tcl_UniCharIsWordChar) (int ch); /* 351 */ int (*tcl_Char16Len) (const unsigned short *uniStr); /* 352 */ - TCL_DEPRECATED_API("Use Tcl_UtfNcmp") int (*tcl_UniCharNcmp) (const Tcl_UniChar *ucs, const Tcl_UniChar *uct, unsigned long numChars); /* 353 */ + TCL_DEPRECATED_API("Use Tcl_UtfNcmp") int (*tcl_UniCharNcmp) (const unsigned short *ucs, const unsigned short *uct, unsigned long numChars); /* 353 */ char * (*tcl_Char16ToUtfDString) (const unsigned short *uniStr, int uniLength, Tcl_DString *dsPtr); /* 354 */ unsigned short * (*tcl_UtfToChar16DString) (const char *src, int length, Tcl_DString *dsPtr); /* 355 */ Tcl_RegExp (*tcl_GetRegExpFromObj) (Tcl_Interp *interp, Tcl_Obj *patObj, int flags); /* 356 */ diff --git a/generic/tclInt.h b/generic/tclInt.h index 2873ad3..a0f9622 100644 --- a/generic/tclInt.h +++ b/generic/tclInt.h @@ -3319,6 +3319,16 @@ MODULE_SCOPE void TclErrorStackResetIf(Tcl_Interp *interp, MODULE_SCOPE int TclZipfs_Init(Tcl_Interp *interp); +MODULE_SCOPE int *TclGetUnicode(Tcl_Obj *); +MODULE_SCOPE int *TclGetUnicodeFromObj_(Tcl_Obj *, int *); +MODULE_SCOPE Tcl_Obj *TclNewUnicodeObj(const int *, int); +MODULE_SCOPE void TclSetUnicodeObj(Tcl_Obj *, const int *, int); +MODULE_SCOPE void TclAppendUnicodeToObj(Tcl_Obj *, const int *, int); +MODULE_SCOPE int TclUniCharNcasecmp(const int *, const int *, unsigned long); +MODULE_SCOPE int TclUniCharCaseMatch(const int *, const int *, int); +MODULE_SCOPE int TclUniCharNcmp(const int *, const int *, unsigned long); + + /* * Many parsing tasks need a common definition of whitespace. * Use this routine and macro to achieve that and place @@ -4777,24 +4787,6 @@ MODULE_SCOPE int TclIsPureByteArray(Tcl_Obj *objPtr); /* *---------------------------------------------------------------- - * Macro used by the Tcl core to compare Unicode strings. On big-endian - * systems we can use the more efficient memcmp, but this would not be - * lexically correct on little-endian systems. The ANSI C "prototype" for - * this macro is: - * - * MODULE_SCOPE int TclUniCharNcmp(const Tcl_UniChar *cs, - * const Tcl_UniChar *ct, unsigned long n); - *---------------------------------------------------------------- - */ - -#if defined(WORDS_BIGENDIAN) && (TCL_UTF_MAX > 3) -# define TclUniCharNcmp(cs,ct,n) memcmp((cs),(ct),(n)*sizeof(Tcl_UniChar)) -#else /* !WORDS_BIGENDIAN */ -# define TclUniCharNcmp Tcl_UniCharNcmp -#endif /* WORDS_BIGENDIAN */ - -/* - *---------------------------------------------------------------- * Macro used by the Tcl core to increment a namespace's export epoch * counter. The ANSI C "prototype" for this macro is: * diff --git a/generic/tclStringObj.c b/generic/tclStringObj.c index 7d4aef3..521d13b 100644 --- a/generic/tclStringObj.c +++ b/generic/tclStringObj.c @@ -3492,7 +3492,7 @@ TclStringCmp( s1len *= sizeof(Tcl_UniChar); s2len *= sizeof(Tcl_UniChar); } else { - memCmpFn = (memCmpFn_t)(void *)Tcl_UniCharNcmp; + memCmpFn = (memCmpFn_t)(void *)TclUniCharNcmp; } } } diff --git a/generic/tclStubInit.c b/generic/tclStubInit.c index 1aec652..f9430cb 100644 --- a/generic/tclStubInit.c +++ b/generic/tclStubInit.c @@ -81,21 +81,6 @@ #define TclWinConvertError (void (*)(DWORD))(void *)Tcl_WinConvertError #endif - -#if TCL_UTF_MAX > 3 -static void uniCodePanic(void) { - Tcl_Panic("This extension uses a deprecated function, not available now: Tcl is compiled with -DTCL_UTF_MAX==%d", TCL_UTF_MAX); -} -# define Tcl_GetUnicode (int *(*)(Tcl_Obj *))(void *)uniCodePanic -# define Tcl_GetUnicodeFromObj (int *(*)(Tcl_Obj *, Tcl_UniChar *))(void *)uniCodePanic -# define Tcl_NewUnicodeObj (Tcl_Obj *(*)(const int *, Tcl_UniChar))(void *)uniCodePanic -# define Tcl_SetUnicodeObj (void(*)(Tcl_Obj *, const Tcl_UniChar *, int))(void *)uniCodePanic -# define Tcl_AppendUnicodeToObj (void(*)(Tcl_Obj *, const Tcl_UniChar *, int))(void *)uniCodePanic -# define Tcl_UniCharNcasecmp (int(*)(const Tcl_UniChar *, const Tcl_UniChar *, unsigned long))(void *)uniCodePanic -# define Tcl_UniCharCaseMatch (int(*)(const Tcl_UniChar *, const Tcl_UniChar *, int))(void *)uniCodePanic -# define Tcl_UniCharNcmp (int(*)(const Tcl_UniChar *, const Tcl_UniChar *, unsigned long))(void *)uniCodePanic -#endif - #define TclUtfCharComplete UtfCharComplete #define TclUtfNext UtfNext #define TclUtfPrev UtfPrev diff --git a/generic/tclUtf.c b/generic/tclUtf.c index 169f240..e024b65 100644 --- a/generic/tclUtf.c +++ b/generic/tclUtf.c @@ -1849,9 +1849,36 @@ Tcl_UniCharLen( */ int +TclUniCharNcmp( + const int *ucs, /* Unicode string to compare to uct. */ + const int *uct, /* Unicode string ucs is compared to. */ + unsigned long numChars) /* Number of unichars to compare. */ +{ +#if defined(WORDS_BIGENDIAN) && (TCL_UTF_MAX > 3) + /* + * We are definitely on a big-endian machine; memcmp() is safe + */ + + return memcmp(ucs, uct, numChars*sizeof(Tcl_UniChar)); + +#else /* !WORDS_BIGENDIAN */ + /* + * We can't simply call memcmp() because that is not lexically correct. + */ + + for ( ; numChars != 0; ucs++, uct++, numChars--) { + if (*ucs != *uct) { + return (*ucs - *uct); + } + } + return 0; +#endif /* WORDS_BIGENDIAN */ +} + +int Tcl_UniCharNcmp( - const Tcl_UniChar *ucs, /* Unicode string to compare to uct. */ - const Tcl_UniChar *uct, /* Unicode string ucs is compared to. */ + const unsigned short *ucs, /* Unicode string to compare to uct. */ + const unsigned short *uct, /* Unicode string ucs is compared to. */ unsigned long numChars) /* Number of unichars to compare. */ { #if defined(WORDS_BIGENDIAN) && (TCL_UTF_MAX > 3) @@ -1868,21 +1895,18 @@ Tcl_UniCharNcmp( for ( ; numChars != 0; ucs++, uct++, numChars--) { if (*ucs != *uct) { -#if TCL_UTF_MAX < 4 /* special case for handling upper surrogates */ if (((*ucs & 0xFC00) == 0xD800) && ((*uct & 0xFC00) != 0xD800)) { return 1; } else if (((*uct & 0xFC00) == 0xD800)) { return -1; } -#endif return (*ucs - *uct); } } return 0; #endif /* WORDS_BIGENDIAN */ } - /* *---------------------------------------------------------------------- * |