From eaefc3bdf38a0256fca08f8d0b9a2a137cf8706e Mon Sep 17 00:00:00 2001 From: "jan.nijtmans" Date: Thu, 12 Sep 2019 11:01:18 +0000 Subject: Code cleanup: Add some initialization to "Tcl_UniChar ch" declaration, making the chance higher that 4-byte UTF-8 sequences are handled more reasonable internally (see: [https://core.tcl-lang.org/tk/tktview?name=a179564826|a179564826]). Use more TclGetString() in stead of Tcl_GetString(), which is slightly more efficient. --- generic/tclCompile.c | 4 ++-- generic/tclEncoding.c | 43 ++++++++++++++++++++----------------------- generic/tclUtil.c | 16 ++++++++-------- 3 files changed, 30 insertions(+), 33 deletions(-) diff --git a/generic/tclCompile.c b/generic/tclCompile.c index 680ab66..41c81af 100644 --- a/generic/tclCompile.c +++ b/generic/tclCompile.c @@ -2126,7 +2126,7 @@ TclCompileScript( if (envPtr->iPtr == NULL) { Tcl_Panic("TclCompileScript() called on uninitialized CompileEnv"); } - /* + /* * Check depth to avoid overflow of the C execution stack by too many * nested calls of TclCompileScript (considering interp recursionlimit). * Factor 5/4 (1.25) is used to avoid too mistaken limit recognition @@ -2218,7 +2218,7 @@ TclCompileScript( continue; } - /* + /* * Avoid stack exhaustion by too many nested calls of TclCompileScript * (considering interp recursionlimit). */ diff --git a/generic/tclEncoding.c b/generic/tclEncoding.c index 144954b..002c765 100644 --- a/generic/tclEncoding.c +++ b/generic/tclEncoding.c @@ -305,7 +305,7 @@ Tcl_GetEncodingFromObj( Tcl_Obj *objPtr, Tcl_Encoding *encodingPtr) { - const char *name = Tcl_GetString(objPtr); + const char *name = TclGetString(objPtr); if (objPtr->typePtr != &encodingType) { Tcl_Encoding encoding = Tcl_GetEncoding(interp, name); @@ -704,7 +704,7 @@ Tcl_GetDefaultEncodingDir(void) } Tcl_ListObjIndex(NULL, searchPath, 0, &first); - return Tcl_GetString(first); + return TclGetString(first); } /* @@ -1260,7 +1260,7 @@ Tcl_ExternalToUtf( if (*dstCharsPtr <= maxChars) { break; } - dstLen = Tcl_UtfAtIndex(dst, maxChars) - 1 - dst + TCL_UTF_MAX; + dstLen = Tcl_UtfAtIndex(dst, maxChars) - dst + (TCL_UTF_MAX - 1); flags = savedFlags; *statePtr = savedState; } while (1); @@ -1518,10 +1518,10 @@ OpenEncodingFileChannel( } } if (!verified) { - const char *dirString = Tcl_GetString(directory); + const char *dirString = TclGetString(directory); for (i=0; itoUnicode[hi] = pageMemPtr; p += 2; @@ -2054,13 +2054,13 @@ LoadEscapeEncoding( + Tcl_DStringLength(&escapeData); dataPtr = ckalloc(size); dataPtr->initLen = strlen(init); - memcpy(dataPtr->init, init, (unsigned) dataPtr->initLen + 1); + memcpy(dataPtr->init, init, dataPtr->initLen + 1); dataPtr->finalLen = strlen(final); - memcpy(dataPtr->final, final, (unsigned) dataPtr->finalLen + 1); + memcpy(dataPtr->final, final, dataPtr->finalLen + 1); dataPtr->numSubTables = Tcl_DStringLength(&escapeData) / sizeof(EscapeSubTable); memcpy(dataPtr->subTables, Tcl_DStringValue(&escapeData), - (size_t) Tcl_DStringLength(&escapeData)); + Tcl_DStringLength(&escapeData)); Tcl_DStringFree(&escapeData); memset(dataPtr->prefixBytes, 0, sizeof(dataPtr->prefixBytes)); @@ -2148,7 +2148,7 @@ BinaryProc( *srcReadPtr = srcLen; *dstWrotePtr = srcLen; *dstCharsPtr = srcLen; - memcpy(dst, src, (size_t) srcLen); + memcpy(dst, src, srcLen); return result; } @@ -2425,11 +2425,8 @@ UnicodeToUtfProc( const char *srcStart, *srcEnd; const char *dstEnd, *dstStart; int result, numChars, charLimit = INT_MAX; - Tcl_UniChar *chPtr = (Tcl_UniChar *) statePtr; + Tcl_UniChar ch; - if (flags & TCL_ENCODING_START) { - *statePtr = 0; - } if (flags & TCL_ENCODING_CHAR_LIMIT) { charLimit = *dstCharsPtr; } @@ -2457,11 +2454,11 @@ UnicodeToUtfProc( * Tcl_UniChar-size data. */ - *chPtr = *(Tcl_UniChar *)src; - if (*chPtr && *chPtr < 0x80) { - *dst++ = (*chPtr & 0xFF); + ch = *(Tcl_UniChar *)src; + if (ch && ch < 0x80) { + *dst++ = (ch & 0xFF); } else { - dst += Tcl_UniCharToUtf(*chPtr, dst); + dst += Tcl_UniCharToUtf(ch, dst); } src += sizeof(Tcl_UniChar); } @@ -2953,6 +2950,7 @@ Iso88591FromUtfProc( const char *srcStart, *srcEnd, *srcClose; const char *dstStart, *dstEnd; int result, numChars; + Tcl_UniChar ch = 0; result = TCL_OK; @@ -2967,7 +2965,6 @@ Iso88591FromUtfProc( dstEnd = dst + dstLen - 1; for (numChars = 0; src < srcEnd; numChars++) { - Tcl_UniChar ch = 0; int len; if ((src > srcClose) && (!Tcl_UtfCharComplete(src, srcEnd - src))) { @@ -3321,6 +3318,7 @@ EscapeFromUtfProc( const TableEncodingData *tableDataPtr; const char *tablePrefixBytes; const unsigned short *const *tableFromUnicode; + Tcl_UniChar ch = 0; result = TCL_OK; @@ -3346,7 +3344,7 @@ EscapeFromUtfProc( *dstWrotePtr = 0; return TCL_CONVERT_NOSPACE; } - memcpy(dst, dataPtr->init, (size_t)dataPtr->initLen); + memcpy(dst, dataPtr->init, dataPtr->initLen); dst += dataPtr->initLen; } else { state = PTR2INT(*statePtr); @@ -3361,7 +3359,6 @@ EscapeFromUtfProc( for (numChars = 0; src < srcEnd; numChars++) { unsigned len; int word; - Tcl_UniChar ch = 0; if ((src > srcClose) && (!Tcl_UtfCharComplete(src, srcEnd - src))) { /* @@ -3468,7 +3465,7 @@ EscapeFromUtfProc( memcpy(dst, dataPtr->subTables[0].sequence, len); dst += len; } - memcpy(dst, dataPtr->final, (size_t) dataPtr->finalLen); + memcpy(dst, dataPtr->final, dataPtr->finalLen); dst += dataPtr->finalLen; state &= ~TCL_ENCODING_END; } diff --git a/generic/tclUtil.c b/generic/tclUtil.c index fc5a2ac..941a71d 100644 --- a/generic/tclUtil.c +++ b/generic/tclUtil.c @@ -1725,7 +1725,7 @@ TrimRight( */ do { - Tcl_UniChar ch1; + Tcl_UniChar ch1 = 0; const char *q = trim; int bytesLeft = numTrim; @@ -1737,7 +1737,7 @@ TrimRight( */ do { - Tcl_UniChar ch2; + Tcl_UniChar ch2 = 0; int qInc = TclUtfToUniChar(q, &ch2); if (ch1 == ch2) { @@ -1824,7 +1824,7 @@ TrimLeft( */ do { - Tcl_UniChar ch1; + Tcl_UniChar ch1 = 0; int pInc = TclUtfToUniChar(p, &ch1); const char *q = trim; int bytesLeft = numTrim; @@ -1834,7 +1834,7 @@ TrimLeft( */ do { - Tcl_UniChar ch2; + Tcl_UniChar ch2 = 0; int qInc = TclUtfToUniChar(q, &ch2); if (ch1 == ch2) { @@ -2237,7 +2237,7 @@ Tcl_StringCaseMatch( { int p, charLen; const char *pstart = pattern; - Tcl_UniChar ch1, ch2; + Tcl_UniChar ch1 = 0, ch2 = 0; while (1) { p = *pattern; @@ -2347,7 +2347,7 @@ Tcl_StringCaseMatch( */ if (p == '[') { - Tcl_UniChar startChar, endChar; + Tcl_UniChar startChar = 0, endChar = 0; pattern++; if (UCHAR(*str) < 0x80) { @@ -3069,7 +3069,7 @@ Tcl_DStringGetResult( dsPtr->length = 0; dsPtr->spaceAvl = TCL_DSTRING_STATIC_SIZE; } else { - dsPtr->string = Tcl_GetString(iPtr->objResultPtr); + dsPtr->string = TclGetString(iPtr->objResultPtr); dsPtr->length = iPtr->objResultPtr->length; dsPtr->spaceAvl = dsPtr->length + 1; TclFreeIntRep(iPtr->objResultPtr); @@ -3754,7 +3754,7 @@ TclGetIntForIndex( parseError: if (interp != NULL) { - bytes = Tcl_GetString(objPtr); + bytes = TclGetString(objPtr); Tcl_SetObjResult(interp, Tcl_ObjPrintf( "bad index \"%s\": must be integer?[+-]integer? or" " end?[+-]integer?", bytes)); -- cgit v0.12