diff options
| author | jan.nijtmans <nijtmans@users.sourceforge.net> | 2019-09-14 13:07:19 (GMT) |
|---|---|---|
| committer | jan.nijtmans <nijtmans@users.sourceforge.net> | 2019-09-14 13:07:19 (GMT) |
| commit | fc77f0dda4c38f5d8a31b62f9e2cb35927922dc5 (patch) | |
| tree | 6db3c9a8eecf14ab943a33f85a465aa9c37758cc | |
| parent | 4396d37f6f0fc2dc3b776872b73ff9717ea26e43 (diff) | |
| parent | 93022718af12833e135ad743bc6169bcfd443ddf (diff) | |
| download | tcl-fc77f0dda4c38f5d8a31b62f9e2cb35927922dc5.zip tcl-fc77f0dda4c38f5d8a31b62f9e2cb35927922dc5.tar.gz tcl-fc77f0dda4c38f5d8a31b62f9e2cb35927922dc5.tar.bz2 | |
Merge 8.7
| -rw-r--r-- | doc/Encoding.3 | 16 | ||||
| -rw-r--r-- | doc/Utf.3 | 40 | ||||
| -rw-r--r-- | generic/tcl.decls | 33 | ||||
| -rw-r--r-- | generic/tcl.h | 4 | ||||
| -rw-r--r-- | generic/tclDecls.h | 72 | ||||
| -rw-r--r-- | generic/tclIOSock.c | 5 | ||||
| -rw-r--r-- | generic/tclInt.h | 18 | ||||
| -rw-r--r-- | generic/tclMain.c | 6 | ||||
| -rw-r--r-- | generic/tclPlatDecls.h | 32 | ||||
| -rw-r--r-- | generic/tclStubInit.c | 47 | ||||
| -rw-r--r-- | generic/tclUtf.c | 154 | ||||
| -rw-r--r-- | generic/tclUtil.c | 2 | ||||
| -rw-r--r-- | generic/tclZipfs.c | 3 | ||||
| -rw-r--r-- | win/tclWin32Dll.c | 78 | ||||
| -rw-r--r-- | win/tclWinFCmd.c | 54 | ||||
| -rw-r--r-- | win/tclWinFile.c | 65 | ||||
| -rw-r--r-- | win/tclWinInit.c | 4 | ||||
| -rw-r--r-- | win/tclWinLoad.c | 3 | ||||
| -rw-r--r-- | win/tclWinPipe.c | 21 | ||||
| -rw-r--r-- | win/tclWinSerial.c | 3 | ||||
| -rw-r--r-- | win/tclWinSock.c | 4 |
21 files changed, 320 insertions, 344 deletions
diff --git a/doc/Encoding.3 b/doc/Encoding.3 index c3161e1..497683d 100644 --- a/doc/Encoding.3 +++ b/doc/Encoding.3 @@ -8,7 +8,7 @@ .so man.macros .BS .SH NAME -Tcl_GetEncoding, Tcl_FreeEncoding, Tcl_GetEncodingFromObj, Tcl_ExternalToUtfDString, Tcl_ExternalToUtf, Tcl_UtfToExternalDString, Tcl_UtfToExternal, Tcl_WinTCharToUtf, Tcl_WinUtfToTChar, Tcl_GetEncodingName, Tcl_SetSystemEncoding, Tcl_GetEncodingNameFromEnvironment, Tcl_GetEncodingNames, Tcl_CreateEncoding, Tcl_GetEncodingSearchPath, Tcl_SetEncodingSearchPath \- procedures for creating and using encodings +Tcl_GetEncoding, Tcl_FreeEncoding, Tcl_GetEncodingFromObj, Tcl_ExternalToUtfDString, Tcl_ExternalToUtf, Tcl_UtfToExternalDString, Tcl_UtfToExternal, Tcl_GetEncodingName, Tcl_SetSystemEncoding, Tcl_GetEncodingNameFromEnvironment, Tcl_GetEncodingNames, Tcl_CreateEncoding, Tcl_GetEncodingSearchPath, Tcl_SetEncodingSearchPath \- procedures for creating and using encodings .SH SYNOPSIS .nf \fB#include <tcl.h>\fR @@ -36,12 +36,6 @@ int \fBTcl_UtfToExternal\fR(\fIinterp, encoding, src, srcLen, flags, statePtr, dst, dstLen, srcReadPtr, dstWrotePtr, dstCharsPtr\fR) .sp -char * -\fBTcl_WinTCharToUtf\fR(\fItsrc, srcLen, dstPtr\fR) -.sp -TCHAR * -\fBTcl_WinUtfToTChar\fR(\fIsrc, srcLen, dstPtr\fR) -.sp const char * \fBTcl_GetEncodingName\fR(\fIencoding\fR) .sp @@ -79,7 +73,7 @@ Points to storage where encoding token is to be written. .AP "const char" *src in For the \fBTcl_ExternalToUtf\fR functions, an array of bytes in the specified encoding that are to be converted to UTF-8. For the -\fBTcl_UtfToExternal\fR and \fBTcl_WinUtfToTChar\fR functions, an array of +\fBTcl_UtfToExternal\fR function, an array of UTF-8 characters to be converted to the specified encoding. .AP "const TCHAR" *tsrc in An array of Windows TCHAR characters to convert to UTF-8. @@ -249,12 +243,6 @@ is filled with the corresponding number of bytes that were stored in \fIdst\fR. The return values are the same as the return values for \fBTcl_ExternalToUtf\fR. .PP -\fBTcl_WinUtfToTChar\fR and \fBTcl_WinTCharToUtf\fR are -Windows-only convenience -functions for converting between UTF-8 and Windows strings -based on the TCHAR type which is by convention -a Unicode character on Windows NT. -.PP \fBTcl_GetEncodingName\fR is roughly the inverse of \fBTcl_GetEncoding\fR. Given an \fIencoding\fR, the return value is the \fIname\fR argument that was used to create the encoding. The string returned by @@ -8,7 +8,7 @@ .so man.macros .BS .SH NAME -Tcl_UniChar, Tcl_UniCharToUtf, Tcl_UtfToUniChar, Tcl_UniCharToUtfDString, Tcl_UtfToUniCharDString, Tcl_UniCharLen, Tcl_UniCharNcmp, Tcl_UniCharNcasecmp, Tcl_UniCharCaseMatch, Tcl_UtfNcmp, Tcl_UtfNcasecmp, Tcl_UtfCharComplete, Tcl_NumUtfChars, Tcl_UtfFindFirst, Tcl_UtfFindLast, Tcl_UtfNext, Tcl_UtfPrev, Tcl_UniCharAtIndex, Tcl_UtfAtIndex, Tcl_UtfBackslash \- routines for manipulating UTF-8 strings +Tcl_UniChar, Tcl_UniCharToUtf, Tcl_UtfToUniChar, Tcl_UtfToChar16, Tcl_UtfToWChar, Tcl_UniCharToUtfDString, Tcl_UtfToUniCharDString, Tcl_WCharToUtfDString, Tcl_UtfToWCharDString, Tcl_Char16ToUtfDString, Tcl_UtfToChar16DString, Tcl_UniCharLen, Tcl_UniCharNcmp, Tcl_UniCharNcasecmp, Tcl_UniCharCaseMatch, Tcl_UtfNcmp, Tcl_UtfNcasecmp, Tcl_UtfCharComplete, Tcl_NumUtfChars, Tcl_UtfFindFirst, Tcl_UtfFindLast, Tcl_UtfNext, Tcl_UtfPrev, Tcl_UniCharAtIndex, Tcl_UtfAtIndex, Tcl_UtfBackslash \- routines for manipulating UTF-8 strings .SH SYNOPSIS .nf \fB#include <tcl.h>\fR @@ -21,12 +21,30 @@ int int \fBTcl_UtfToUniChar\fR(\fIsrc, chPtr\fR) .sp +int +\fBTcl_UtfToChar16\fR(\fIsrc, uPtr\fR) +.sp +int +\fBTcl_UtfToWChar\fR(\fIsrc, wPtr\fR) +.sp char * \fBTcl_UniCharToUtfDString\fR(\fIuniStr, uniLength, dsPtr\fR) .sp +char * +\fBTcl_Char16ToUtfDString\fR(\fIuStr, uniLength, dsPtr\fR) +.sp +char * +\fBTcl_WCharToUtfDString\fR(\fIwStr, uniLength, dsPtr\fR) +.sp Tcl_UniChar * \fBTcl_UtfToUniCharDString\fR(\fIsrc, length, dsPtr\fR) .sp +unsigned short * +\fBTcl_UtfToChar16DString\fR(\fIsrc, length, dsPtr\fR) +.sp +wchar_t * +\fBTcl_UtfToWCharDString\fR(\fIsrc, length, dsPtr\fR) +.sp int \fBTcl_UniCharLen\fR(\fIuniStr\fR) .sp @@ -80,6 +98,10 @@ Buffer in which the UTF-8 representation of the Tcl_UniChar is stored. At most The Unicode character to be converted or examined. .AP Tcl_UniChar *chPtr out Filled with the Tcl_UniChar represented by the head of the UTF-8 string. +.AP unsigned short *uPtr out +Filled with the utf-16 represented by the head of the UTF-8 string. +.AP wchar_t *wPtr out +Filled with the wchar_t represented by the head of the UTF-8 string. .AP "const char" *src in Pointer to a UTF-8 string. .AP "const char" *cs in @@ -94,9 +116,13 @@ A null-terminated Unicode string. A null-terminated Unicode string. .AP "const Tcl_UniChar" *uniPattern in A null-terminated Unicode string. +.AP "const unsigned short" *uStr in +A null-terminated UTF-16 string. +.AP "const wchar_t" *wStr in +A null-terminated wchar_t string. .AP size_t length in The length of the UTF-8 string in bytes (not UTF-8 characters). If -TCL_AUTO_LENGTH, all bytes up to the first null byte are used. +negative, all bytes up to the first null byte are used. .AP size_t uniLength in The length of the Unicode string in characters. .AP "Tcl_DString" *dsPtr in/out @@ -118,11 +144,10 @@ case-insensitive (1). .SH DESCRIPTION .PP -These routines convert between UTF-8 strings and Unicode characters. An -Unicode character represented as an unsigned, fixed-size -quantity. A UTF-8 character is a Unicode character represented as -a varying-length sequence of up to \fBTCL_UTF_MAX\fR bytes. A multibyte UTF-8 -sequence consists of a lead byte followed by some number of trail bytes. +These routines convert between UTF-8 strings and Unicode/Utf-16 characters. +A UTF-8 character is a Unicode character represented as a varying-length +sequence of up to \fBTCL_UTF_MAX\fR bytes. A multibyte UTF-8 sequence +consists of a lead byte followed by some number of trail bytes. .PP \fBTCL_UTF_MAX\fR is the maximum number of bytes that it takes to represent one Unicode character in the UTF-8 representation. @@ -150,7 +175,6 @@ byte of \fIsrc\fR in \fI*chPtr\fR as a Tcl_UniChar between 0x00A0 and .PP \fBTcl_UniCharToUtfDString\fR converts the given Unicode string to UTF-8, storing the result in a previously initialized \fBTcl_DString\fR. -You must specify \fIuniLength\fR, the length of the given Unicode string. The return value is a pointer to the UTF-8 representation of the Unicode string. Storage for the return value is appended to the end of the \fBTcl_DString\fR. diff --git a/generic/tcl.decls b/generic/tcl.decls index f72d57e..8b16a1b 100644 --- a/generic/tcl.decls +++ b/generic/tcl.decls @@ -1242,7 +1242,7 @@ declare 335 { int Tcl_UtfToTitle(char *src) } declare 336 { - int Tcl_UtfToUniChar(const char *src, Tcl_UniChar *chPtr) + int Tcl_UtfToChar16(const char *src, unsigned short *chPtr) } declare 337 { int Tcl_UtfToUpper(char *src) @@ -1299,11 +1299,11 @@ declare 353 { size_t numChars) } declare 354 { - char *Tcl_UniCharToUtfDString(const Tcl_UniChar *uniStr, + char *Tcl_Char16ToUtfDString(const unsigned short *uniStr, size_t uniLength, Tcl_DString *dsPtr) } declare 355 { - Tcl_UniChar *Tcl_UtfToUniCharDString(const char *src, + unsigned short *Tcl_UtfToChar16DString(const char *src, size_t length, Tcl_DString *dsPtr) } declare 356 { @@ -2444,6 +2444,19 @@ declare 645 { size_t endValue, size_t *indexPtr) } +# TIP #548 +declare 646 { + int Tcl_UtfToUniChar(const char *src, int *chPtr) +} +declare 647 { + char *Tcl_UniCharToUtfDString(const int *uniStr, + size_t uniLength, Tcl_DString *dsPtr) +} +declare 648 { + int *Tcl_UtfToUniCharDString(const char *src, + size_t length, Tcl_DString *dsPtr) +} + # ----- BASELINE -- FOR -- 8.7.0 ----- # ############################################################################## @@ -2460,14 +2473,14 @@ interface tclPlat ################################ # Windows specific functions -# Added in Tcl 8.1 +# Added in Tcl 8.1, Removed in Tcl 9.0 (converted to macro) -declare 0 win { - TCHAR *Tcl_WinUtfToTChar(const char *str, size_t len, Tcl_DString *dsPtr) -} -declare 1 win { - char *Tcl_WinTCharToUtf(const TCHAR *str, size_t len, Tcl_DString *dsPtr) -} +#declare 0 win { +# TCHAR *Tcl_WinUtfToTChar(const char *str, size_t len, Tcl_DString *dsPtr) +#} +#declare 1 win { +# char *Tcl_WinTCharToUtf(const TCHAR *str, size_t len, Tcl_DString *dsPtr) +#} ################################ # Mac OS X specific functions diff --git a/generic/tcl.h b/generic/tcl.h index 8017dd0..37a11ff 100644 --- a/generic/tcl.h +++ b/generic/tcl.h @@ -1984,7 +1984,7 @@ typedef struct Tcl_EncodingType { #if TCL_UTF_MAX > 4 /* - * unsigned int isn't 100% accurate as it should be a strict 4-byte value + * int isn't 100% accurate as it should be a strict 4-byte value * (perhaps wchar_t). 64-bit systems may have troubles. The size of this * value must be reflected correctly in regcustom.h and * in tclEncoding.c. @@ -1992,7 +1992,7 @@ typedef struct Tcl_EncodingType { * XXX: string rep that Tcl_UniChar represents. Changing the size * XXX: of Tcl_UniChar is /not/ supported. */ -typedef unsigned int Tcl_UniChar; +typedef int Tcl_UniChar; #else typedef unsigned short Tcl_UniChar; #endif diff --git a/generic/tclDecls.h b/generic/tclDecls.h index 35ad3fa..5d8fdd6 100644 --- a/generic/tclDecls.h +++ b/generic/tclDecls.h @@ -910,7 +910,8 @@ EXTERN int Tcl_UtfToLower(char *src); /* 335 */ EXTERN int Tcl_UtfToTitle(char *src); /* 336 */ -EXTERN int Tcl_UtfToUniChar(const char *src, Tcl_UniChar *chPtr); +EXTERN int Tcl_UtfToChar16(const char *src, + unsigned short *chPtr); /* 337 */ EXTERN int Tcl_UtfToUpper(char *src); /* 338 */ @@ -946,10 +947,10 @@ EXTERN size_t Tcl_UniCharLen(const Tcl_UniChar *uniStr); EXTERN int Tcl_UniCharNcmp(const Tcl_UniChar *ucs, const Tcl_UniChar *uct, size_t numChars); /* 354 */ -EXTERN char * Tcl_UniCharToUtfDString(const Tcl_UniChar *uniStr, +EXTERN char * Tcl_Char16ToUtfDString(const unsigned short *uniStr, size_t uniLength, Tcl_DString *dsPtr); /* 355 */ -EXTERN Tcl_UniChar * Tcl_UtfToUniCharDString(const char *src, +EXTERN unsigned short * Tcl_UtfToChar16DString(const char *src, size_t length, Tcl_DString *dsPtr); /* 356 */ EXTERN Tcl_RegExp Tcl_GetRegExpFromObj(Tcl_Interp *interp, @@ -1757,6 +1758,14 @@ EXTERN int Tcl_LinkArray(Tcl_Interp *interp, EXTERN int Tcl_GetIntForIndex(Tcl_Interp *interp, Tcl_Obj *objPtr, size_t endValue, size_t *indexPtr); +/* 646 */ +EXTERN int Tcl_UtfToUniChar(const char *src, int *chPtr); +/* 647 */ +EXTERN char * Tcl_UniCharToUtfDString(const int *uniStr, + size_t uniLength, Tcl_DString *dsPtr); +/* 648 */ +EXTERN int * Tcl_UtfToUniCharDString(const char *src, + size_t length, Tcl_DString *dsPtr); typedef struct { const struct TclPlatStubs *tclPlatStubs; @@ -2128,7 +2137,7 @@ typedef struct TclStubs { char * (*tcl_UtfToExternalDString) (Tcl_Encoding encoding, const char *src, size_t srcLen, Tcl_DString *dsPtr); /* 333 */ int (*tcl_UtfToLower) (char *src); /* 334 */ int (*tcl_UtfToTitle) (char *src); /* 335 */ - int (*tcl_UtfToUniChar) (const char *src, Tcl_UniChar *chPtr); /* 336 */ + int (*tcl_UtfToChar16) (const char *src, unsigned short *chPtr); /* 336 */ int (*tcl_UtfToUpper) (char *src); /* 337 */ size_t (*tcl_WriteChars) (Tcl_Channel chan, const char *src, size_t srcLen); /* 338 */ size_t (*tcl_WriteObj) (Tcl_Channel chan, Tcl_Obj *objPtr); /* 339 */ @@ -2146,8 +2155,8 @@ typedef struct TclStubs { int (*tcl_UniCharIsWordChar) (int ch); /* 351 */ size_t (*tcl_UniCharLen) (const Tcl_UniChar *uniStr); /* 352 */ int (*tcl_UniCharNcmp) (const Tcl_UniChar *ucs, const Tcl_UniChar *uct, size_t numChars); /* 353 */ - char * (*tcl_UniCharToUtfDString) (const Tcl_UniChar *uniStr, size_t uniLength, Tcl_DString *dsPtr); /* 354 */ - Tcl_UniChar * (*tcl_UtfToUniCharDString) (const char *src, size_t length, Tcl_DString *dsPtr); /* 355 */ + char * (*tcl_Char16ToUtfDString) (const unsigned short *uniStr, size_t uniLength, Tcl_DString *dsPtr); /* 354 */ + unsigned short * (*tcl_UtfToChar16DString) (const char *src, size_t length, Tcl_DString *dsPtr); /* 355 */ Tcl_RegExp (*tcl_GetRegExpFromObj) (Tcl_Interp *interp, Tcl_Obj *patObj, int flags); /* 356 */ void (*reserved357)(void); void (*tcl_FreeParse) (Tcl_Parse *parsePtr); /* 358 */ @@ -2438,6 +2447,9 @@ typedef struct TclStubs { int (*tcl_IsShared) (Tcl_Obj *objPtr); /* 643 */ int (*tcl_LinkArray) (Tcl_Interp *interp, const char *varName, void *addr, int type, size_t size); /* 644 */ int (*tcl_GetIntForIndex) (Tcl_Interp *interp, Tcl_Obj *objPtr, size_t endValue, size_t *indexPtr); /* 645 */ + int (*tcl_UtfToUniChar) (const char *src, int *chPtr); /* 646 */ + char * (*tcl_UniCharToUtfDString) (const int *uniStr, size_t uniLength, Tcl_DString *dsPtr); /* 647 */ + int * (*tcl_UtfToUniCharDString) (const char *src, size_t length, Tcl_DString *dsPtr); /* 648 */ } TclStubs; extern const TclStubs *tclStubsPtr; @@ -3095,8 +3107,8 @@ extern const TclStubs *tclStubsPtr; (tclStubsPtr->tcl_UtfToLower) /* 334 */ #define Tcl_UtfToTitle \ (tclStubsPtr->tcl_UtfToTitle) /* 335 */ -#define Tcl_UtfToUniChar \ - (tclStubsPtr->tcl_UtfToUniChar) /* 336 */ +#define Tcl_UtfToChar16 \ + (tclStubsPtr->tcl_UtfToChar16) /* 336 */ #define Tcl_UtfToUpper \ (tclStubsPtr->tcl_UtfToUpper) /* 337 */ #define Tcl_WriteChars \ @@ -3129,10 +3141,10 @@ extern const TclStubs *tclStubsPtr; (tclStubsPtr->tcl_UniCharLen) /* 352 */ #define Tcl_UniCharNcmp \ (tclStubsPtr->tcl_UniCharNcmp) /* 353 */ -#define Tcl_UniCharToUtfDString \ - (tclStubsPtr->tcl_UniCharToUtfDString) /* 354 */ -#define Tcl_UtfToUniCharDString \ - (tclStubsPtr->tcl_UtfToUniCharDString) /* 355 */ +#define Tcl_Char16ToUtfDString \ + (tclStubsPtr->tcl_Char16ToUtfDString) /* 354 */ +#define Tcl_UtfToChar16DString \ + (tclStubsPtr->tcl_UtfToChar16DString) /* 355 */ #define Tcl_GetRegExpFromObj \ (tclStubsPtr->tcl_GetRegExpFromObj) /* 356 */ /* Slot 357 is reserved */ @@ -3706,6 +3718,12 @@ extern const TclStubs *tclStubsPtr; (tclStubsPtr->tcl_LinkArray) /* 644 */ #define Tcl_GetIntForIndex \ (tclStubsPtr->tcl_GetIntForIndex) /* 645 */ +#define Tcl_UtfToUniChar \ + (tclStubsPtr->tcl_UtfToUniChar) /* 646 */ +#define Tcl_UniCharToUtfDString \ + (tclStubsPtr->tcl_UniCharToUtfDString) /* 647 */ +#define Tcl_UtfToUniCharDString \ + (tclStubsPtr->tcl_UtfToUniCharDString) /* 648 */ #endif /* defined(USE_TCL_STUBS) */ @@ -3855,6 +3873,36 @@ extern const TclStubs *tclStubsPtr; #define Tcl_BackgroundError(interp) Tcl_BackgroundException((interp), TCL_ERROR) #define Tcl_StringMatch(str, pattern) Tcl_StringCaseMatch((str), (pattern), 0) +#if TCL_UTF_MAX <= 4 +# undef Tcl_UniCharToUtfDString +# define Tcl_UniCharToUtfDString Tcl_Char16ToUtfDString +# undef Tcl_UtfToUniCharDString +# define Tcl_UtfToUniCharDString Tcl_UtfToChar16DString +# undef Tcl_UtfToUniChar +# define Tcl_UtfToUniChar Tcl_UtfToChar16 +#endif +#if defined(USE_TCL_STUBS) +# define Tcl_WCharToUtfDString (sizeof(wchar_t) != sizeof(short) \ + ? (char *(*)(const wchar_t *, size_t, Tcl_DString *))tclStubsPtr->tcl_UniCharToUtfDString \ + : (char *(*)(const wchar_t *, size_t, Tcl_DString *))Tcl_Char16ToUtfDString) +# define Tcl_UtfToWCharDString (sizeof(wchar_t) != sizeof(short) \ + ? (wchar_t *(*)(const char *, size_t, Tcl_DString *))tclStubsPtr->tcl_UtfToUniCharDString \ + : (wchar_t *(*)(const char *, size_t, Tcl_DString *))Tcl_UtfToChar16DString) +# define Tcl_UtfToWChar (sizeof(wchar_t) != sizeof(short) \ + ? (int (*)(const char *, wchar_t *))tclStubsPtr->tcl_UtfToChar16 \ + : (int (*)(const char *, wchar_t *))Tcl_UtfToUniChar) +#else +# define Tcl_WCharToUtfDString (sizeof(wchar_t) != sizeof(short) \ + ? (char *(*)(const wchar_t *, size_t, Tcl_DString *))Tcl_UniCharToUtfDString \ + : (char *(*)(const wchar_t *, size_t, Tcl_DString *))Tcl_Char16ToUtfDString) +# define Tcl_UtfToWCharDString (sizeof(wchar_t) != sizeof(short) \ + ? (wchar_t *(*)(const char *, size_t, Tcl_DString *))Tcl_UtfToUniCharDString \ + : (wchar_t *(*)(const char *, size_t, Tcl_DString *))Tcl_UtfToChar16DString) +# define Tcl_UtfToWChar (sizeof(wchar_t) != sizeof(short) \ + ? (int (*)(const char *, wchar_t *))Tcl_UtfToChar16 \ + : (int (*)(const char *, wchar_t *))Tcl_UtfToUniChar) +#endif + /* * Deprecated Tcl procedures: */ diff --git a/generic/tclIOSock.c b/generic/tclIOSock.c index db124fb..531c38b 100644 --- a/generic/tclIOSock.c +++ b/generic/tclIOSock.c @@ -30,11 +30,12 @@ gai_strerror( ThreadSpecificData *tsdPtr = TCL_TSD_INIT(&dataKey); if (tsdPtr->initialized) { - Tcl_DStringFree(&tsdPtr->errorMsg); + Tcl_DStringSetLength(&tsdPtr->errorMsg, 0); } else { + Tcl_DStringInit(&tsdPtr->errorMsg); tsdPtr->initialized = 1; } - Tcl_WinTCharToUtf(gai_strerrorW(code), -1, &tsdPtr->errorMsg); + Tcl_WCharToUtfDString(gai_strerrorW(code), -1, &tsdPtr->errorMsg); return Tcl_DStringValue(&tsdPtr->errorMsg); } #endif diff --git a/generic/tclInt.h b/generic/tclInt.h index 8afd63e..7bb6170 100644 --- a/generic/tclInt.h +++ b/generic/tclInt.h @@ -3172,17 +3172,6 @@ MODULE_SCOPE const char*TclGetCommandTypeName(Tcl_Command command); MODULE_SCOPE void TclRegisterCommandTypeName( Tcl_ObjCmdProc *implementationProc, const char *nameStr); -#if (TCL_UTF_MAX > 4) && (defined(__CYGWIN__) || defined(_WIN32)) -MODULE_SCOPE int TclUtfToWChar(const char *src, WCHAR *chPtr); -MODULE_SCOPE char * TclWCharToUtfDString(const WCHAR *uniStr, - size_t uniLength, Tcl_DString *dsPtr); -MODULE_SCOPE WCHAR * TclUtfToWCharDString(const char *src, - size_t length, Tcl_DString *dsPtr); -#else -# define TclUtfToWChar TclUtfToUniChar -# define TclWCharToUtfDString Tcl_UniCharToUtfDString -# define TclUtfToWCharDString Tcl_UtfToUniCharDString -#endif MODULE_SCOPE int TclUtfCmp(const char *cs, const char *ct); MODULE_SCOPE int TclUtfCasecmp(const char *cs, const char *ct); MODULE_SCOPE size_t TclUtfCount(int ch); @@ -4576,10 +4565,17 @@ MODULE_SCOPE void TclDbInitNewObj(Tcl_Obj *objPtr, const char *file, *---------------------------------------------------------------- */ +#if TCL_UTF_MAX > 4 #define TclUtfToUniChar(str, chPtr) \ ((((unsigned char) *(str)) < 0x80) ? \ ((*(chPtr) = (unsigned char) *(str)), 1) \ : Tcl_UtfToUniChar(str, chPtr)) +#else +#define TclUtfToUniChar(str, chPtr) \ + ((((unsigned char) *(str)) < 0x80) ? \ + ((*(chPtr) = (unsigned char) *(str)), 1) \ + : Tcl_UtfToChar16(str, chPtr)) +#endif /* *---------------------------------------------------------------- diff --git a/generic/tclMain.c b/generic/tclMain.c index 31e6438..e5a4408 100644 --- a/generic/tclMain.c +++ b/generic/tclMain.c @@ -70,10 +70,8 @@ NewNativeObj( Tcl_DString ds; #ifdef UNICODE - if (length != TCL_AUTO_LENGTH) { - length *= sizeof(WCHAR); - } - Tcl_WinTCharToUtf(string, length, &ds); + Tcl_DStringInit(&ds); + Tcl_WCharToUtfDString(string, length, &ds); #else Tcl_ExternalToUtfDString(NULL, (char *) string, length, &ds); #endif diff --git a/generic/tclPlatDecls.h b/generic/tclPlatDecls.h index 7218490..b1a874f 100644 --- a/generic/tclPlatDecls.h +++ b/generic/tclPlatDecls.h @@ -50,14 +50,6 @@ extern "C" { * Exported function declarations: */ -#if defined(_WIN32) || defined(__CYGWIN__) /* WIN */ -/* 0 */ -EXTERN TCHAR * Tcl_WinUtfToTChar(const char *str, size_t len, - Tcl_DString *dsPtr); -/* 1 */ -EXTERN char * Tcl_WinTCharToUtf(const TCHAR *str, size_t len, - Tcl_DString *dsPtr); -#endif /* WIN */ #ifdef MAC_OSX_TCL /* MACOSX */ /* 0 */ EXTERN int Tcl_MacOSXOpenBundleResources(Tcl_Interp *interp, @@ -75,10 +67,6 @@ typedef struct TclPlatStubs { int magic; void *hooks; -#if defined(_WIN32) || defined(__CYGWIN__) /* WIN */ - TCHAR * (*tcl_WinUtfToTChar) (const char *str, size_t len, Tcl_DString *dsPtr); /* 0 */ - char * (*tcl_WinTCharToUtf) (const TCHAR *str, size_t len, Tcl_DString *dsPtr); /* 1 */ -#endif /* WIN */ #ifdef MAC_OSX_TCL /* MACOSX */ int (*tcl_MacOSXOpenBundleResources) (Tcl_Interp *interp, const char *bundleName, int hasResourceFile, size_t maxPathLen, char *libraryPath); /* 0 */ int (*tcl_MacOSXOpenVersionedBundleResources) (Tcl_Interp *interp, const char *bundleName, const char *bundleVersion, int hasResourceFile, size_t maxPathLen, char *libraryPath); /* 1 */ @@ -97,12 +85,6 @@ extern const TclPlatStubs *tclPlatStubsPtr; * Inline function declarations: */ -#if defined(_WIN32) || defined(__CYGWIN__) /* WIN */ -#define Tcl_WinUtfToTChar \ - (tclPlatStubsPtr->tcl_WinUtfToTChar) /* 0 */ -#define Tcl_WinTCharToUtf \ - (tclPlatStubsPtr->tcl_WinTCharToUtf) /* 1 */ -#endif /* WIN */ #ifdef MAC_OSX_TCL /* MACOSX */ #define Tcl_MacOSXOpenBundleResources \ (tclPlatStubsPtr->tcl_MacOSXOpenBundleResources) /* 0 */ @@ -117,6 +99,16 @@ extern const TclPlatStubs *tclPlatStubsPtr; #undef TCL_STORAGE_CLASS #define TCL_STORAGE_CLASS DLLIMPORT -#endif /* _TCLPLATDECLS */ - +#if defined(USE_TCL_STUBS) && (defined(_WIN32) || defined(__CYGWIN__))\ + && (defined(TCL_NO_DEPRECATED) || TCL_MAJOR_VERSION > 8) +#undef Tcl_WinUtfToTChar +#undef Tcl_WinTCharToUtf +#ifdef _WIN32 +#define Tcl_WinUtfToTChar(string, len, dsPtr) (Tcl_DStringInit(dsPtr), \ + (TCHAR *)Tcl_UtfToChar16DString((string), (len), (dsPtr))) +#define Tcl_WinTCharToUtf(string, len, dsPtr) (Tcl_DStringInit(dsPtr), \ + (char *)Tcl_Char16ToUtfDString((string), ((((len) + 2) >> 1) - 1), (dsPtr))) +#endif +#endif +#endif /* _TCLPLATDECLS */ diff --git a/generic/tclStubInit.c b/generic/tclStubInit.c index 493d56f..1320b36 100644 --- a/generic/tclStubInit.c +++ b/generic/tclStubInit.c @@ -44,6 +44,9 @@ #undef TclpGetPid #undef TclStaticPackage #undef Tcl_BackgroundError +#undef Tcl_UtfToUniChar +#undef Tcl_UtfToUniCharDString +#undef Tcl_UniCharToUtfDString #define TclStaticPackage Tcl_StaticPackage #ifdef TCL_MEM_DEBUG @@ -104,37 +107,6 @@ TclpGetPid(Tcl_Pid pid) return (size_t) pid; } -char * -Tcl_WinUtfToTChar( - const char *string, - size_t len, - Tcl_DString *dsPtr) -{ - Tcl_DStringInit(dsPtr); - if (!string) { - return NULL; - } - return (char *)TclUtfToWCharDString(string, len, dsPtr); -} - -char * -Tcl_WinTCharToUtf( - const char *string, - size_t len, - Tcl_DString *dsPtr) -{ - Tcl_DStringInit(dsPtr); - if (!string) { - return NULL; - } - if (len == TCL_AUTO_LENGTH) { - len = wcslen((wchar_t *)string); - } else { - len /= 2; - } - return TclWCharToUtfDString((const WCHAR *)string, len, dsPtr); -} - #if defined(TCL_WIDE_INT_IS_LONG) /* On Cygwin64, long is 64-bit while on Win64 long is 32-bit. Therefore * we have to make sure that all stub entries on Cygwin64 follow the Win64 @@ -568,10 +540,6 @@ static const TclIntPlatStubs tclIntPlatStubs = { static const TclPlatStubs tclPlatStubs = { TCL_STUB_MAGIC, 0, -#if defined(_WIN32) || defined(__CYGWIN__) /* WIN */ - Tcl_WinUtfToTChar, /* 0 */ - Tcl_WinTCharToUtf, /* 1 */ -#endif /* WIN */ #ifdef MAC_OSX_TCL /* MACOSX */ Tcl_MacOSXOpenBundleResources, /* 0 */ Tcl_MacOSXOpenVersionedBundleResources, /* 1 */ @@ -1030,7 +998,7 @@ const TclStubs tclStubs = { Tcl_UtfToExternalDString, /* 333 */ Tcl_UtfToLower, /* 334 */ Tcl_UtfToTitle, /* 335 */ - Tcl_UtfToUniChar, /* 336 */ + Tcl_UtfToChar16, /* 336 */ Tcl_UtfToUpper, /* 337 */ Tcl_WriteChars, /* 338 */ Tcl_WriteObj, /* 339 */ @@ -1048,8 +1016,8 @@ const TclStubs tclStubs = { Tcl_UniCharIsWordChar, /* 351 */ Tcl_UniCharLen, /* 352 */ Tcl_UniCharNcmp, /* 353 */ - Tcl_UniCharToUtfDString, /* 354 */ - Tcl_UtfToUniCharDString, /* 355 */ + Tcl_Char16ToUtfDString, /* 354 */ + Tcl_UtfToChar16DString, /* 355 */ Tcl_GetRegExpFromObj, /* 356 */ 0, /* 357 */ Tcl_FreeParse, /* 358 */ @@ -1340,6 +1308,9 @@ const TclStubs tclStubs = { Tcl_IsShared, /* 643 */ Tcl_LinkArray, /* 644 */ Tcl_GetIntForIndex, /* 645 */ + Tcl_UtfToUniChar, /* 646 */ + Tcl_UniCharToUtfDString, /* 647 */ + Tcl_UtfToUniCharDString, /* 648 */ }; /* !END!: Do not edit above this line. */ diff --git a/generic/tclUtf.c b/generic/tclUtf.c index b5167a4..644939b 100644 --- a/generic/tclUtf.c +++ b/generic/tclUtf.c @@ -221,23 +221,33 @@ three: *--------------------------------------------------------------------------- */ +#undef Tcl_UniCharToUtfDString char * Tcl_UniCharToUtfDString( - const Tcl_UniChar *uniStr, /* Unicode string to convert to UTF-8. */ - size_t uniLength, /* Length of Unicode string in Tcl_UniChars - * (must be >= 0). */ + const int *uniStr, /* Unicode string to convert to UTF-8. */ + size_t uniLength, /* Length of Unicode string. */ Tcl_DString *dsPtr) /* UTF-8 representation of string is appended * to this previously initialized DString. */ { - const Tcl_UniChar *w, *wEnd; + const int *w, *wEnd; char *p, *string; size_t oldLength; - int len = 1; /* * UTF-8 string length in bytes will be <= Unicode string length * 4. */ + if (uniStr == NULL) { + return NULL; + } + if (uniLength < 0) { + uniLength = 0; + w = uniStr; + while (*w != '\0') { + uniLength++; + w++; + } + } oldLength = Tcl_DStringLength(dsPtr); Tcl_DStringSetLength(dsPtr, oldLength + (uniLength + 1) * 4); string = Tcl_DStringValue(dsPtr) + oldLength; @@ -245,46 +255,44 @@ Tcl_UniCharToUtfDString( p = string; wEnd = uniStr + uniLength; for (w = uniStr; w < wEnd; ) { - if (!len && ((*w & 0xFC00) != 0xDC00)) { - /* Special case for handling high surrogates. */ - p += Tcl_UniCharToUtf(-1, p); - } - len = Tcl_UniCharToUtf(*w, p); - p += len; - if ((*w >= 0xD800) && (len < 3)) { - len = 0; /* Indication that high surrogate was found */ - } + p += Tcl_UniCharToUtf(*w, p); w++; } - if (!len) { - /* Special case for handling high surrogates. */ - p += Tcl_UniCharToUtf(-1, p); - } Tcl_DStringSetLength(dsPtr, oldLength + (p - string)); return string; } -#if (TCL_UTF_MAX > 4) && (defined(__CYGWIN__) || defined(_WIN32)) char * -TclWCharToUtfDString( - const WCHAR *uniStr, /* WCHAR string to convert to UTF-8. */ - size_t uniLength, /* Length of WCHAR string in Tcl_UniChars - * (must be >= 0). */ +Tcl_Char16ToUtfDString( + const unsigned short *uniStr,/* Utf-16 string to convert to UTF-8. */ + size_t uniLength, /* Length of Utf-16 string. */ Tcl_DString *dsPtr) /* UTF-8 representation of string is appended * to this previously initialized DString. */ { - const WCHAR *w, *wEnd; + const unsigned short *w, *wEnd; char *p, *string; size_t oldLength; int len = 1; /* - * UTF-8 string length in bytes will be <= Unicode string length * 4. + * UTF-8 string length in bytes will be <= Utf16 string length * 3. */ + if (uniStr == NULL) { + return NULL; + } + if (uniLength < 0) { + + uniLength = 0; + w = uniStr; + while (*w != '\0') { + uniLength++; + w++; + } + } oldLength = Tcl_DStringLength(dsPtr); - Tcl_DStringSetLength(dsPtr, oldLength + (uniLength + 1) * 4); + Tcl_DStringSetLength(dsPtr, oldLength + (uniLength + 1) * 3); string = Tcl_DStringValue(dsPtr) + oldLength; p = string; @@ -309,7 +317,6 @@ TclWCharToUtfDString( return string; } -#endif /* *--------------------------------------------------------------------------- * @@ -352,13 +359,14 @@ static const unsigned short cp1252[32] = { 0x2DC, 0x2122, 0x0161, 0x203A, 0x0153, 0x9D, 0x017E, 0x0178 }; +#undef Tcl_UtfToUniChar int Tcl_UtfToUniChar( const char *src, /* The UTF-8 string. */ - Tcl_UniChar *chPtr)/* Filled with the Tcl_UniChar represented by + int *chPtr)/* Filled with the unsigned int represented by * the UTF-8 string. */ { - Tcl_UniChar byte; + int byte; /* * Unroll 1 to 4 byte UTF-8 sequences. @@ -374,20 +382,6 @@ Tcl_UtfToUniChar( * characters representing themselves. */ -#if TCL_UTF_MAX <= 4 - /* If *chPtr contains a high surrogate (produced by a previous - * Tcl_UtfToUniChar() call) and the next 3 bytes are UTF-8 continuation - * bytes, then we must produce a follow-up low surrogate. We only - * do that if the high surrogate matches the bits we encounter. - */ - if ((byte >= 0x80) - && (((((byte - 0x10) << 2) & 0xFC) | 0xD800) == (*chPtr & 0xFCFC)) - && ((src[1] & 0xF0) == (((*chPtr << 4) & 0x30) | 0x80)) - && ((src[2] & 0xC0) == 0x80)) { - *chPtr = ((src[1] & 0x0F) << 6) + (src[2] & 0x3F) + 0xDC00; - return 3; - } -#endif if ((unsigned)(byte-0x80) < (unsigned)0x20) { *chPtr = cp1252[byte-0x80]; } else { @@ -433,23 +427,11 @@ Tcl_UtfToUniChar( /* * Four-byte-character lead byte followed by three trail bytes. */ -#if TCL_UTF_MAX <= 4 - Tcl_UniChar high = (((byte & 0x07) << 8) | ((src[1] & 0x3F) << 2) - | ((src[2] & 0x3F) >> 4)) - 0x40; - if (high >= 0x400) { - /* out of range, < 0x10000 or > 0x10ffff */ - } else { - /* produce high surrogate, advance source pointer */ - *chPtr = 0xD800 + high; - return 1; - } -#else *chPtr = (((byte & 0x07) << 18) | ((src[1] & 0x3F) << 12) | ((src[2] & 0x3F) << 6) | (src[3] & 0x3F)); if ((unsigned)(*chPtr - 0x10000) <= 0xFFFFF) { return 4; } -#endif } /* @@ -462,14 +444,13 @@ Tcl_UtfToUniChar( return 1; } -#if (TCL_UTF_MAX > 4) && (defined(__CYGWIN__) || defined(_WIN32)) int -TclUtfToWChar( +Tcl_UtfToChar16( const char *src, /* The UTF-8 string. */ - WCHAR *chPtr)/* Filled with the WCHAR represented by + unsigned short *chPtr)/* Filled with the unsigned short represented by * the UTF-8 string. */ { - WCHAR byte; + unsigned short byte; /* * Unroll 1 to 4 byte UTF-8 sequences. @@ -542,7 +523,7 @@ TclUtfToWChar( /* * Four-byte-character lead byte followed by three trail bytes. */ - WCHAR high = (((byte & 0x07) << 8) | ((src[1] & 0x3F) << 2) + unsigned short high = (((byte & 0x07) << 8) | ((src[1] & 0x3F) << 2) | ((src[2] & 0x3F) >> 4)) - 0x40; if (high >= 0x400) { /* out of range, < 0x10000 or > 0x10ffff */ @@ -562,7 +543,6 @@ TclUtfToWChar( *chPtr = byte; return 1; } -#endif /* *--------------------------------------------------------------------------- @@ -582,7 +562,8 @@ TclUtfToWChar( *--------------------------------------------------------------------------- */ -Tcl_UniChar * +#undef Tcl_UtfToUniCharDString +int * Tcl_UtfToUniCharDString( const char *src, /* UTF-8 string to convert to Unicode. */ size_t length, /* Length of UTF-8 string in bytes, or -1 for @@ -591,10 +572,13 @@ Tcl_UtfToUniCharDString( * appended to this previously initialized * DString. */ { - Tcl_UniChar ch = 0, *w, *wString; + int ch = 0, *w, *wString; const char *p, *end; size_t oldLength; + if (src == NULL) { + return NULL; + } if (length == TCL_AUTO_LENGTH) { length = strlen(src); } @@ -607,20 +591,20 @@ Tcl_UtfToUniCharDString( oldLength = Tcl_DStringLength(dsPtr); Tcl_DStringSetLength(dsPtr, - oldLength + (int) ((length + 1) * sizeof(Tcl_UniChar))); - wString = (Tcl_UniChar *) (Tcl_DStringValue(dsPtr) + oldLength); + oldLength + ((length + 1) * sizeof(int))); + wString = (int *) (Tcl_DStringValue(dsPtr) + oldLength); w = wString; p = src; end = src + length - 4; while (p < end) { - p += TclUtfToUniChar(p, &ch); + p += Tcl_UtfToUniChar(p, &ch); *w++ = ch; } end += 4; while (p < end) { if (Tcl_UtfCharComplete(p, end-p)) { - p += TclUtfToUniChar(p, &ch); + p += Tcl_UtfToUniChar(p, &ch); } else { ch = UCHAR(*p++); } @@ -633,9 +617,8 @@ Tcl_UtfToUniCharDString( return wString; } -#if (TCL_UTF_MAX > 4) && (defined(__CYGWIN__) || defined(_WIN32)) -WCHAR * -TclUtfToWCharDString( +unsigned short * +Tcl_UtfToChar16DString( const char *src, /* UTF-8 string to convert to Unicode. */ size_t length, /* Length of UTF-8 string in bytes, or -1 for * strlen(). */ @@ -643,10 +626,14 @@ TclUtfToWCharDString( * appended to this previously initialized * DString. */ { - WCHAR ch = 0, *w, *wString; + unsigned short ch = 0; + unsigned short *w, *wString; const char *p, *end; size_t oldLength; + if (src == NULL) { + return NULL; + } if (length == TCL_AUTO_LENGTH) { length = strlen(src); } @@ -659,20 +646,20 @@ TclUtfToWCharDString( oldLength = Tcl_DStringLength(dsPtr); Tcl_DStringSetLength(dsPtr, - oldLength + (int) ((length + 1) * sizeof(WCHAR))); - wString = (WCHAR *) (Tcl_DStringValue(dsPtr) + oldLength); + oldLength + ((length + 1) * sizeof(unsigned short))); + wString = (unsigned short *) (Tcl_DStringValue(dsPtr) + oldLength); w = wString; p = src; end = src + length - 4; while (p < end) { - p += TclUtfToWChar(p, &ch); + p += Tcl_UtfToChar16(p, &ch); *w++ = ch; } end += 4; while (p < end) { if (Tcl_UtfCharComplete(p, end-p)) { - p += TclUtfToWChar(p, &ch); + p += Tcl_UtfToChar16(p, &ch); } else { ch = UCHAR(*p++); } @@ -684,7 +671,6 @@ TclUtfToWCharDString( return wString; } -#endif /* *--------------------------------------------------------------------------- * @@ -2161,7 +2147,7 @@ Tcl_UniCharCaseMatch( if ((p != '[') && (p != '?') && (p != '\\')) { if (nocase) { while (*uniStr && (p != *uniStr) - && (p != (Tcl_UniChar)Tcl_UniCharToLower(*uniStr))) { + && (p != Tcl_UniCharToLower(*uniStr))) { uniStr++; } } else { @@ -2201,13 +2187,13 @@ Tcl_UniCharCaseMatch( Tcl_UniChar startChar, endChar; uniPattern++; - ch1 = (nocase ? (Tcl_UniChar)Tcl_UniCharToLower(*uniStr) : *uniStr); + ch1 = (nocase ? Tcl_UniCharToLower(*uniStr) : *uniStr); uniStr++; while (1) { if ((*uniPattern == ']') || (*uniPattern == 0)) { return 0; } - startChar = (nocase ? (Tcl_UniChar)Tcl_UniCharToLower(*uniPattern) + startChar = (nocase ? Tcl_UniCharToLower(*uniPattern) : *uniPattern); uniPattern++; if (*uniPattern == '-') { @@ -2215,7 +2201,7 @@ Tcl_UniCharCaseMatch( if (*uniPattern == 0) { return 0; } - endChar = (nocase ? (Tcl_UniChar)Tcl_UniCharToLower(*uniPattern) + endChar = (nocase ? Tcl_UniCharToLower(*uniPattern) : *uniPattern); uniPattern++; if (((startChar <= ch1) && (ch1 <= endChar)) @@ -2353,7 +2339,7 @@ TclUniCharMatch( if ((p != '[') && (p != '?') && (p != '\\')) { if (nocase) { while ((string < stringEnd) && (p != *string) - && (p != (Tcl_UniChar)Tcl_UniCharToLower(*string))) { + && (p != Tcl_UniCharToLower(*string))) { string++; } } else { @@ -2394,20 +2380,20 @@ TclUniCharMatch( Tcl_UniChar ch1, startChar, endChar; pattern++; - ch1 = (nocase ? (Tcl_UniChar)Tcl_UniCharToLower(*string) : *string); + ch1 = (nocase ? Tcl_UniCharToLower(*string) : *string); string++; while (1) { if ((*pattern == ']') || (pattern == patternEnd)) { return 0; } - startChar = (nocase ? (Tcl_UniChar)Tcl_UniCharToLower(*pattern) : *pattern); + startChar = (nocase ? Tcl_UniCharToLower(*pattern) : *pattern); pattern++; if (*pattern == '-') { pattern++; if (pattern == patternEnd) { return 0; } - endChar = (nocase ? (Tcl_UniChar)Tcl_UniCharToLower(*pattern) + endChar = (nocase ? Tcl_UniCharToLower(*pattern) : *pattern); pattern++; if (((startChar <= ch1) && (ch1 <= endChar)) diff --git a/generic/tclUtil.c b/generic/tclUtil.c index 2ffd4ea..c9d7003 100644 --- a/generic/tclUtil.c +++ b/generic/tclUtil.c @@ -2211,7 +2211,7 @@ Tcl_StringCaseMatch( if (nocase) { while (*str) { charLen = TclUtfToUniChar(str, &ch1); - if (ch2==ch1 || ch2==(Tcl_UniChar)Tcl_UniCharToLower(ch1)) { + if (ch2==ch1 || ch2==Tcl_UniCharToLower(ch1)) { break; } str += charLen; diff --git a/generic/tclZipfs.c b/generic/tclZipfs.c index 914c6bb..4e7cb1f 100644 --- a/generic/tclZipfs.c +++ b/generic/tclZipfs.c @@ -4913,7 +4913,8 @@ TclZipfs_AppHook( #ifdef _WIN32 Tcl_DString ds; - archive = Tcl_WinTCharToUtf((*argvPtr)[1], -1, &ds); + Tcl_DStringInit(&ds); + archive = Tcl_WCharToUtfDString((*argvPtr)[1], -1, &ds); #else /* !_WIN32 */ archive = (*argvPtr)[1]; #endif /* _WIN32 */ diff --git a/win/tclWin32Dll.c b/win/tclWin32Dll.c index 4fc97db..a286254 100644 --- a/win/tclWin32Dll.c +++ b/win/tclWin32Dll.c @@ -421,84 +421,6 @@ TclWinDriveLetterForVolMountPoint( } /* - *--------------------------------------------------------------------------- - * - * Tcl_WinUtfToTChar, Tcl_WinTCharToUtf -- - * - * Convert between UTF-8 and Unicode when running Windows. - * - * On Mac and Unix, all strings exchanged between Tcl and the OS are - * "char" oriented. We need only one Tcl_Encoding to convert between - * UTF-8 and the system's native encoding. We use NULL to represent - * that encoding. - * - * On Windows, some strings exchanged between Tcl and the OS are "char" - * oriented, while others are in Unicode. We need two Tcl_Encoding APIs - * depending on whether we are targeting a "char" or Unicode interface. - * - * Calling Tcl_UtfToExternal() or Tcl_ExternalToUtf() with an encoding - * of NULL should always used to convert between UTF-8 and the system's - * "char" oriented encoding. The following two functions are used in - * Windows-specific code to convert between UTF-8 and Unicode strings. - * This saves you the trouble of writing the - * following type of fragment over and over: - * - * encoding <- Tcl_GetEncoding("unicode"); - * nativeBuffer <- UtfToExternal(encoding, utfBuffer); - * Tcl_FreeEncoding(encoding); - * - * By convention, in Windows a WCHAR is a Unicode character. If you plan - * on targeting a Unicode interface when running on Windows, these - * functions should be used. If you plan on targetting a "char" oriented - * function on Windows, use Tcl_UtfToExternal() with an encoding of NULL. - * - * Results: - * The result is a pointer to the string in the desired target encoding. - * Storage for the result string is allocated in dsPtr; the caller must - * call Tcl_DStringFree() when the result is no longer needed. - * - * Side effects: - * None. - * - *--------------------------------------------------------------------------- - */ - -WCHAR * -Tcl_WinUtfToTChar( - const char *string, /* Source string in UTF-8. */ - size_t len, /* Source string length in bytes, or -1 - * for strlen(). */ - Tcl_DString *dsPtr) /* Uninitialized or free DString in which the - * converted string is stored. */ -{ - Tcl_DStringInit(dsPtr); - if (!string) { - return NULL; - } - return TclUtfToWCharDString(string, len, dsPtr); -} - -char * -Tcl_WinTCharToUtf( - const WCHAR *string, /* Source string in Unicode. */ - size_t len, /* Source string length in bytes, or -1 - * for platform-specific string length. */ - Tcl_DString *dsPtr) /* Uninitialized or free DString in which the - * converted string is stored. */ -{ - Tcl_DStringInit(dsPtr); - if (!string) { - return NULL; - } - if (len == TCL_AUTO_LENGTH) { - len = wcslen((WCHAR *)string); - } else { - len /= 2; - } - return TclWCharToUtfDString((unsigned short *)string, len, dsPtr); -} - -/* *------------------------------------------------------------------------ * * TclWinCPUID -- diff --git a/win/tclWinFCmd.c b/win/tclWinFCmd.c index 6780328..f080d9f 100644 --- a/win/tclWinFCmd.c +++ b/win/tclWinFCmd.c @@ -328,8 +328,10 @@ DoRenameFile( CharLower(nativeSrcPath); CharLower(nativeDstPath); - src = Tcl_WinTCharToUtf(nativeSrcPath, -1, &srcString); - dst = Tcl_WinTCharToUtf(nativeDstPath, -1, &dstString); + Tcl_DStringInit(&srcString); + Tcl_DStringInit(&dstString); + src = Tcl_WCharToUtfDString(nativeSrcPath, -1, &srcString); + dst = Tcl_WCharToUtfDString(nativeDstPath, -1, &dstString); /* * Check whether the destination path is actually inside the @@ -911,8 +913,10 @@ TclpObjCopyDirectory( return TCL_ERROR; } - Tcl_WinUtfToTChar(TclGetString(normSrcPtr), -1, &srcString); - Tcl_WinUtfToTChar(TclGetString(normDestPtr), -1, &dstString); + Tcl_DStringInit(&srcString); + Tcl_DStringInit(&dstString); + Tcl_UtfToWCharDString(TclGetString(normSrcPtr), -1, &srcString); + Tcl_UtfToWCharDString(TclGetString(normDestPtr), -1, &dstString); ret = TraverseWinTree(TraversalCopy, &srcString, &dstString, &ds); @@ -984,7 +988,8 @@ TclpObjRemoveDirectory( if (normPtr == NULL) { return TCL_ERROR; } - Tcl_WinUtfToTChar(TclGetString(normPtr), -1, &native); + Tcl_DStringInit(&native); + Tcl_UtfToWCharDString(TclGetString(normPtr), -1, &native); ret = DoRemoveDirectory(&native, recursive, &ds); Tcl_DStringFree(&native); } else { @@ -1109,7 +1114,10 @@ DoRemoveJustDirectory( end: if (errorPtr != NULL) { - char *p = Tcl_WinTCharToUtf(nativePath, -1, errorPtr); + char *p; + + Tcl_DStringInit(errorPtr); + p = Tcl_WCharToUtfDString(nativePath, -1, errorPtr); for (; *p; ++p) { if (*p == '\\') *p = '/'; } @@ -1323,7 +1331,8 @@ TraverseWinTree( if (nativeErrfile != NULL) { TclWinConvertError(GetLastError()); if (errorPtr != NULL) { - Tcl_WinTCharToUtf(nativeErrfile, -1, errorPtr); + Tcl_DStringInit(errorPtr); + Tcl_WCharToUtfDString(nativeErrfile, -1, errorPtr); } result = TCL_ERROR; } @@ -1388,7 +1397,8 @@ TraversalCopy( */ if (errorPtr != NULL) { - Tcl_WinTCharToUtf(nativeDst, -1, errorPtr); + Tcl_DStringInit(errorPtr); + Tcl_WCharToUtfDString(nativeDst, -1, errorPtr); } return TCL_ERROR; } @@ -1443,7 +1453,8 @@ TraversalDelete( } if (errorPtr != NULL) { - Tcl_WinTCharToUtf(nativeSrc, -1, errorPtr); + Tcl_DStringInit(errorPtr); + Tcl_WCharToUtfDString(nativeSrc, -1, errorPtr); } return TCL_ERROR; } @@ -1651,7 +1662,8 @@ ConvertFileNameFormat( */ tempString = TclGetStringFromObj(tempPath, &length); - nativeName = Tcl_WinUtfToTChar(tempString, length, &ds); + Tcl_DStringInit(&ds); + nativeName = Tcl_UtfToWCharDString(tempString, length, &ds); Tcl_DecrRefCount(tempPath); handle = FindFirstFile(nativeName, &data); if (handle == INVALID_HANDLE_VALUE) { @@ -1688,7 +1700,7 @@ ConvertFileNameFormat( } /* - * Purify reports a extraneous UMR in Tcl_WinTCharToUtf() trying + * Purify reports a extraneous UMR in Tcl_WCharToUtfDString() trying * to dereference nativeName as a Unicode string. I have proven to * myself that purify is wrong by running the following example * when nativeName == data.w.cAlternateFileName and noting that @@ -1700,7 +1712,7 @@ ConvertFileNameFormat( */ Tcl_DStringInit(&dsTemp); - Tcl_WinTCharToUtf(nativeName, -1, &dsTemp); + Tcl_WCharToUtfDString(nativeName, -1, &dsTemp); Tcl_DStringFree(&ds); /* @@ -1995,9 +2007,10 @@ TclpCreateTemporaryDirectory( if (dirObj->length < 1) { goto useSystemTemp; } - Tcl_WinUtfToTChar(Tcl_GetString(dirObj), -1, &base); + Tcl_DStringInit(&base); + Tcl_UtfToWCharDString(Tcl_GetString(dirObj), -1, &base); if (dirObj->bytes[dirObj->length - 1] != '\\') { - TclUtfToWCharDString("\\", -1, &base); + Tcl_UtfToWCharDString("\\", -1, &base); } } else { useSystemTemp: @@ -2013,13 +2026,11 @@ TclpCreateTemporaryDirectory( #define SUFFIX_LENGTH 8 if (basenameObj) { - Tcl_WinUtfToTChar(Tcl_GetString(basenameObj), -1, &name); - TclDStringAppendDString(&base, &name); - Tcl_DStringFree(&name); + Tcl_UtfToWCharDString(Tcl_GetString(basenameObj), -1, &base); } else { - TclUtfToWCharDString(DEFAULT_TEMP_DIR_PREFIX, -1, &base); + Tcl_UtfToWCharDString(DEFAULT_TEMP_DIR_PREFIX, -1, &base); } - TclUtfToWCharDString("_", -1, &base); + Tcl_UtfToWCharDString("_", -1, &base); /* * Now we keep on trying random suffixes until we get one that works @@ -2046,7 +2057,7 @@ TclpCreateTemporaryDirectory( tempbuf[i] = randChars[(int) (rand() % numRandChars)]; } Tcl_DStringSetLength(&base, baseLen); - TclUtfToWCharDString(tempbuf, -1, &base); + Tcl_UtfToWCharDString(tempbuf, -1, &base); } while (!CreateDirectoryW((LPCWSTR) Tcl_DStringValue(&base), NULL) && (error = GetLastError()) == ERROR_ALREADY_EXISTS); @@ -2066,7 +2077,8 @@ TclpCreateTemporaryDirectory( * as a (clean) Tcl_Obj. */ - Tcl_WinTCharToUtf((LPCWSTR) Tcl_DStringValue(&base), -1, &name); + Tcl_DStringInit(&name); + Tcl_WCharToUtfDString((LPCWSTR) Tcl_DStringValue(&base), -1, &name); Tcl_DStringFree(&base); return TclDStringToObj(&name); } diff --git a/win/tclWinFile.c b/win/tclWinFile.c index 14c4378..4073f9a 100644 --- a/win/tclWinFile.c +++ b/win/tclWinFile.c @@ -634,10 +634,11 @@ WinReadLinkDirectory( } } - Tcl_WinTCharToUtf( + Tcl_DStringInit(&ds); + Tcl_WCharToUtfDString( reparseBuffer->MountPointReparseBuffer.PathBuffer, reparseBuffer->MountPointReparseBuffer - .SubstituteNameLength, &ds); + .SubstituteNameLength>>1, &ds); copy = Tcl_DStringValue(&ds)+offset; len = Tcl_DStringLength(&ds)-offset; @@ -1003,7 +1004,8 @@ TclpMatchInDirectory( dirName = TclDStringAppendLiteral(&dsOrig, "*.*"); } - native = Tcl_WinUtfToTChar(dirName, -1, &ds); + Tcl_DStringInit(&ds); + native = Tcl_UtfToWCharDString(dirName, -1, &ds); if ((types == NULL) || (types->type != TCL_GLOB_TYPE_DIR)) { handle = FindFirstFile(native, &data); } else { @@ -1076,7 +1078,8 @@ TclpMatchInDirectory( native = data.cFileName; attr = data.dwFileAttributes; - utfname = Tcl_WinTCharToUtf(native, -1, &ds); + Tcl_DStringInit(&ds); + utfname = Tcl_WCharToUtfDString(native, -1, &ds); if (!matchSpecialDots) { /* @@ -1451,14 +1454,14 @@ TclpGetUserHome( Tcl_DStringFree(&ds); } else { Tcl_DStringInit(&ds); - wName = TclUtfToWCharDString(domain + 1, -1, &ds); + wName = Tcl_UtfToWCharDString(domain + 1, -1, &ds); rc = NetGetDCName(NULL, wName, (LPBYTE *) &wDomain); Tcl_DStringFree(&ds); nameLen = domain - name; } if (rc == 0) { Tcl_DStringInit(&ds); - wName = TclUtfToWCharDString(name, nameLen, &ds); + wName = Tcl_UtfToWCharDString(name, nameLen, &ds); while (NetUserGetInfo(wDomain, wName, 1, (LPBYTE *) &uiPtr) != 0) { /* * User does not exist; if domain was not specified, try again @@ -1486,7 +1489,7 @@ TclpGetUserHome( wHomeDir = uiPtr->usri1_home_dir; if ((wHomeDir != NULL) && (wHomeDir[0] != '\0')) { size = lstrlenW(wHomeDir); - TclWCharToUtfDString(wHomeDir, size, bufferPtr); + Tcl_WCharToUtfDString(wHomeDir, size, bufferPtr); } else { /* * User exists but has no home dir. Return @@ -1494,7 +1497,7 @@ TclpGetUserHome( */ GetProfilesDirectoryW(buf, &size); - TclWCharToUtfDString(buf, size-1, bufferPtr); + Tcl_WCharToUtfDString(buf, size-1, bufferPtr); Tcl_DStringAppend(bufferPtr, "/", 1); Tcl_DStringAppend(bufferPtr, name, nameLen); } @@ -1864,11 +1867,11 @@ NativeIsExec( } path += len-3; - if ((wcsicmp(path, L"exe") == 0) - || (wcsicmp(path, L"com") == 0) - || (wcsicmp(path, L"cmd") == 0) - || (wcsicmp(path, L"cmd") == 0) - || (wcsicmp(path, L"bat") == 0)) { + if ((_wcsicmp(path, L"exe") == 0) + || (_wcsicmp(path, L"com") == 0) + || (_wcsicmp(path, L"cmd") == 0) + || (_wcsicmp(path, L"cmd") == 0) + || (_wcsicmp(path, L"bat") == 0)) { return 1; } return 0; @@ -1962,7 +1965,8 @@ TclpGetCwd( && (native[2] == '\\') && (native[3] == '\\')) { native += 2; } - Tcl_WinTCharToUtf(native, -1, bufferPtr); + Tcl_DStringInit(bufferPtr); + Tcl_WCharToUtfDString(native, -1, bufferPtr); /* * Convert to forward slashes for easier use in scripts. @@ -2170,7 +2174,8 @@ NativeDev( const char *fullPath; GetFullPathName(nativePath, MAX_PATH, nativeFullPath, &nativePart); - fullPath = Tcl_WinTCharToUtf(nativeFullPath, -1, &ds); + Tcl_DStringInit(&ds); + fullPath = Tcl_WCharToUtfDString(nativeFullPath, -1, &ds); if ((fullPath[0] == '\\') && (fullPath[1] == '\\')) { const char *p; @@ -2191,7 +2196,8 @@ NativeDev( } else { p++; } - nativeVol = Tcl_WinUtfToTChar(fullPath, p - fullPath, &volString); + Tcl_DStringInit(&volString); + nativeVol = Tcl_UtfToWCharDString(fullPath, p - fullPath, &volString); dw = (DWORD) -1; GetVolumeInformation(nativeVol, NULL, 0, &dw, NULL, NULL, NULL, 0); @@ -2471,7 +2477,8 @@ TclpFilesystemPathType( } else { Tcl_DString ds; - Tcl_WinTCharToUtf(volType, -1, &ds); + Tcl_DStringInit(&ds); + Tcl_WCharToUtfDString(volType, -1, &ds); return TclDStringToObj(&ds); } #undef VOL_BUF_SIZE @@ -2541,7 +2548,10 @@ TclpObjNormalizePath( */ WIN32_FILE_ATTRIBUTE_DATA data; - const WCHAR *nativePath = Tcl_WinUtfToTChar(path, + const WCHAR *nativePath; + + Tcl_DStringInit(&ds); + nativePath = Tcl_UtfToWCharDString(path, currentPathEndPosition - path, &ds); if (GetFileAttributesEx(nativePath, @@ -2743,11 +2753,14 @@ TclpObjNormalizePath( if (1) { WCHAR wpath[MAX_PATH]; - const WCHAR *nativePath = - Tcl_WinUtfToTChar(path, lastValidPathEnd - path, &ds); - DWORD wpathlen = GetLongPathNameProc(nativePath, - (WCHAR *) wpath, MAX_PATH); + const WCHAR *nativePath; + DWORD wpathlen; + Tcl_DStringInit(&ds); + nativePath = + Tcl_UtfToWCharDString(path, lastValidPathEnd - path, &ds); + wpathlen = GetLongPathNameProc(nativePath, + (WCHAR *) wpath, MAX_PATH); /* * We have to make the drive letter uppercase. */ @@ -2774,8 +2787,9 @@ TclpObjNormalizePath( * native encoding, so we have to convert it to Utf. */ - Tcl_WinTCharToUtf((const WCHAR *) Tcl_DStringValue(&dsNorm), - Tcl_DStringLength(&dsNorm), &ds); + Tcl_DStringInit(&ds); + Tcl_WCharToUtfDString((const WCHAR *) Tcl_DStringValue(&dsNorm), + Tcl_DStringLength(&dsNorm)>>1, &ds); nextCheckpoint = Tcl_DStringLength(&ds); if (*lastValidPathEnd != 0) { /* @@ -2950,7 +2964,8 @@ TclpNativeToNormalized( size_t len; char *copy, *p; - Tcl_WinTCharToUtf((const WCHAR *) clientData, -1, &ds); + Tcl_DStringInit(&ds); + Tcl_WCharToUtfDString((const WCHAR *) clientData, -1, &ds); copy = Tcl_DStringValue(&ds); len = Tcl_DStringLength(&ds); diff --git a/win/tclWinInit.c b/win/tclWinInit.c index 1bd1548..c6d5c1d 100644 --- a/win/tclWinInit.c +++ b/win/tclWinInit.c @@ -476,8 +476,8 @@ TclpGetUserName( return NULL; } cchUserNameLen--; - cchUserNameLen *= sizeof(WCHAR); - Tcl_WinTCharToUtf(szUserName, cchUserNameLen, bufferPtr); + Tcl_DStringInit(bufferPtr); + Tcl_WCharToUtfDString(szUserName, cchUserNameLen, bufferPtr); } return Tcl_DStringValue(bufferPtr); } diff --git a/win/tclWinLoad.c b/win/tclWinLoad.c index be6299c..254f59a 100644 --- a/win/tclWinLoad.c +++ b/win/tclWinLoad.c @@ -95,7 +95,8 @@ TclpDlopen( firstError = (nativeName == NULL) ? ERROR_MOD_NOT_FOUND : GetLastError(); - nativeName = Tcl_WinUtfToTChar(TclGetString(pathPtr), -1, &ds); + Tcl_DStringInit(&ds); + nativeName = Tcl_UtfToWCharDString(TclGetString(pathPtr), -1, &ds); hInstance = LoadLibraryEx(nativeName, NULL, LOAD_WITH_ALTERED_SEARCH_PATH); Tcl_DStringFree(&ds); diff --git a/win/tclWinPipe.c b/win/tclWinPipe.c index 1aa28bc..fae5ed3 100644 --- a/win/tclWinPipe.c +++ b/win/tclWinPipe.c @@ -577,7 +577,8 @@ TclpOpenFile( break; } - nativePath = Tcl_WinUtfToTChar(path, -1, &ds); + Tcl_DStringInit(&ds); + nativePath = Tcl_UtfToWCharDString(path, -1, &ds); /* * If the file is not being created, use the existing file attributes. @@ -1290,7 +1291,8 @@ ApplicationType( for (i = 0; i < (int) (sizeof(extensions) / sizeof(extensions[0])); i++) { Tcl_DStringSetLength(&nameBuf, nameLen); Tcl_DStringAppend(&nameBuf, extensions[i], -1); - nativeName = Tcl_WinUtfToTChar(Tcl_DStringValue(&nameBuf), + Tcl_DStringInit(&ds); + nativeName = Tcl_UtfToWCharDString(Tcl_DStringValue(&nameBuf), Tcl_DStringLength(&nameBuf), &ds); found = SearchPath(NULL, nativeName, NULL, MAX_PATH, nativeFullPath, &rest); @@ -1308,7 +1310,8 @@ ApplicationType( if ((attr == 0xffffffff) || (attr & FILE_ATTRIBUTE_DIRECTORY)) { continue; } - strcpy(fullName, Tcl_WinTCharToUtf(nativeFullPath, -1, &ds)); + Tcl_DStringInit(&ds); + strcpy(fullName, Tcl_WCharToUtfDString(nativeFullPath, -1, &ds)); Tcl_DStringFree(&ds); ext = strrchr(fullName, '.'); @@ -1399,7 +1402,8 @@ ApplicationType( */ GetShortPathName(nativeFullPath, nativeFullPath, MAX_PATH); - strcpy(fullName, Tcl_WinTCharToUtf(nativeFullPath, -1, &ds)); + Tcl_DStringInit(&ds); + strcpy(fullName, Tcl_WCharToUtfDString(nativeFullPath, -1, &ds)); Tcl_DStringFree(&ds); } return applType; @@ -1727,7 +1731,8 @@ BuildCommandLine( } } Tcl_DStringFree(linePtr); - Tcl_WinUtfToTChar(Tcl_DStringValue(&ds), Tcl_DStringLength(&ds), linePtr); + Tcl_DStringInit(linePtr); + Tcl_UtfToWCharDString(Tcl_DStringValue(&ds), Tcl_DStringLength(&ds), linePtr); Tcl_DStringFree(&ds); } @@ -3210,7 +3215,8 @@ TclpOpenTemporaryFile( if (basenameObj) { const char *string = TclGetStringFromObj(basenameObj, &length); - Tcl_WinUtfToTChar(string, length, &buf); + Tcl_DStringInit(&buf); + Tcl_UtfToWCharDString(string, length, &buf); memcpy(namePtr, Tcl_DStringValue(&buf), Tcl_DStringLength(&buf)); namePtr += Tcl_DStringLength(&buf); Tcl_DStringFree(&buf); @@ -3230,7 +3236,8 @@ TclpOpenTemporaryFile( sprintf(number, "%d.TMP", counter); counter = (unsigned short) (counter + 1); - Tcl_WinUtfToTChar(number, strlen(number), &buf); + Tcl_DStringInit(&buf); + Tcl_UtfToWCharDString(number, strlen(number), &buf); Tcl_DStringSetLength(&buf, Tcl_DStringLength(&buf) + 1); memcpy(namePtr, Tcl_DStringValue(&buf), Tcl_DStringLength(&buf) + 1); Tcl_DStringFree(&buf); diff --git a/win/tclWinSerial.c b/win/tclWinSerial.c index 9f559e6..fed1c5e 100644 --- a/win/tclWinSerial.c +++ b/win/tclWinSerial.c @@ -1669,7 +1669,8 @@ SerialSetOptionProc( if (!GetCommState(infoPtr->handle, &dcb)) { goto getStateFailed; } - native = Tcl_WinUtfToTChar(value, -1, &ds); + Tcl_DStringInit(&ds); + native = Tcl_UtfToWCharDString(value, -1, &ds); result = BuildCommDCB(native, &dcb); Tcl_DStringFree(&ds); diff --git a/win/tclWinSock.c b/win/tclWinSock.c index 5cb19e9..1679bbe 100644 --- a/win/tclWinSock.c +++ b/win/tclWinSock.c @@ -367,15 +367,15 @@ InitializeHostName( DWORD length = MAX_COMPUTERNAME_LENGTH + 1; Tcl_DString ds; + Tcl_DStringInit(&ds); if (GetComputerName(tbuf, &length) != 0) { /* * Convert string from native to UTF then change to lowercase. */ - Tcl_UtfToLower(Tcl_WinTCharToUtf(tbuf, -1, &ds)); + Tcl_UtfToLower(Tcl_WCharToUtfDString(tbuf, -1, &ds)); } else { - Tcl_DStringInit(&ds); if (TclpHasSockets(NULL) == TCL_OK) { /* * The buffer size of 256 is recommended by the MSDN page that |
