diff options
Diffstat (limited to 'generic/tclParse.c')
| -rw-r--r-- | generic/tclParse.c | 401 |
1 files changed, 209 insertions, 192 deletions
diff --git a/generic/tclParse.c b/generic/tclParse.c index 2b0dab4..b40b636 100644 --- a/generic/tclParse.c +++ b/generic/tclParse.c @@ -12,7 +12,7 @@ * See the file "license.terms" for information on usage and redistribution of * this file, and for a DISCLAIMER OF ALL WARRANTIES. */ - + #include "tclInt.h" /* @@ -182,13 +182,13 @@ static int ParseWhiteSpace(const char *src, int numBytes, * * TclParseInit -- * - * Initialize the fields of a Tcl_Parse struct. + * Initialize the fields of a Tcl_Parse struct. * * Results: - * None. + * None. * * Side effects: - * The Tcl_Parse struct pointed to by parsePtr gets initialized. + * The Tcl_Parse struct pointed to by parsePtr gets initialized. * *---------------------------------------------------------------------- */ @@ -251,7 +251,7 @@ Tcl_ParseCommand( * command terminator. If zero, then close * bracket has no special meaning. */ register Tcl_Parse *parsePtr) - /* Structure to fill in with information about + /* Structure to fill in with information about * the parsed command; any previous * information in the structure is ignored. */ { @@ -496,10 +496,9 @@ Tcl_ParseCommand( * tokens representing the expanded list. */ - const char *listStart; + CONST char *listStart; int growthNeeded = wordIndex + 2*elemCount - parsePtr->numTokens; - parsePtr->numWords += elemCount - 1; if (growthNeeded > 0) { TclGrowParseTokenArray(parsePtr, growthNeeded); @@ -519,7 +518,7 @@ Tcl_ParseCommand( listStart = nextElem = tokenPtr[1].start; while (nextElem < listEnd) { int quoted; - + tokenPtr->type = TCL_TOKEN_SIMPLE_WORD; tokenPtr->numComponents = 1; @@ -629,6 +628,47 @@ TclIsSpaceProc( /* *---------------------------------------------------------------------- * + * TclIsBareword-- + * + * Report whether byte is one that can be part of a "bareword". + * This concept is named in expression parsing, where it determines + * what can be a legal function name, but is the same definition used + * in determining what variable names can be parsed as variable + * substitutions without the benefit of enclosing braces. The set of + * ASCII chars that are accepted are the numeric chars ('0'-'9'), + * the alphabetic chars ('a'-'z', 'A'-'Z') and underscore ('_'). + * + * Results: + * Returns 1, if byte is in the accepted set of chars, 0 otherwise. + * + * Side effects: + * None. + * + *---------------------------------------------------------------------- + */ + +int +TclIsBareword( + char byte) +{ + if (byte < '0' || byte > 'z') { + return 0; + } + if (byte <= '9' || byte >= 'a') { + return 1; + } + if (byte == '_') { + return 1; + } + if (byte < 'A' || byte > 'Z') { + return 0; + } + return 1; +} + +/* + *---------------------------------------------------------------------- + * * ParseWhiteSpace -- * * Scans up to numBytes bytes starting at src, consuming white space @@ -673,7 +713,7 @@ ParseWhiteSpace( if (p[1] != '\n') { break; } - p += 2; + p+=2; if (--numBytes == 0) { *incompletePtr = 1; break; @@ -744,11 +784,11 @@ int TclParseHex( const char *src, /* First character to parse. */ int numBytes, /* Max number of byes to scan */ - int *resultPtr) /* Points to storage provided by caller where - * the character resulting from the + Tcl_UniChar *resultPtr) /* Points to storage provided by caller where + * the Tcl_UniChar resulting from the * conversion is to be written. */ { - int result = 0; + Tcl_UniChar result = 0; register const char *p = src; while (numBytes--) { @@ -758,7 +798,7 @@ TclParseHex( break; } - p++; + ++p; result <<= 4; if (digit >= 'a') { @@ -783,14 +823,14 @@ TclParseHex( * sequence as defined by Tcl's parsing rules. * * Results: - * Records at readPtr the number of bytes making up the backslash - * sequence. Records at dst the UTF-8 encoded equivalent of that - * backslash sequence. Returns the number of bytes written to dst, at - * most TCL_UTF_MAX. Either readPtr or dst may be NULL, if the results - * are not needed, but the return value is the same either way. + * Records at readPtr the number of bytes making up the backslash + * sequence. Records at dst the UTF-8 encoded equivalent of that + * backslash sequence. Returns the number of bytes written to dst, at + * most TCL_UTF_MAX. Either readPtr or dst may be NULL, if the results + * are not needed, but the return value is the same either way. * * Side effects: - * None. + * None. * *---------------------------------------------------------------------- */ @@ -808,8 +848,7 @@ TclParseBackslash( * written there. */ { register const char *p = src+1; - Tcl_UniChar unichar; - int result; + Tcl_UniChar result; int count; char buf[TCL_UTF_MAX]; @@ -907,21 +946,21 @@ TclParseBackslash( */ if (isdigit(UCHAR(*p)) && (UCHAR(*p) < '8')) { /* INTL: digit */ - result = *p - '0'; + result = (unsigned char)(*p - '0'); p++; if ((numBytes == 2) || !isdigit(UCHAR(*p)) /* INTL: digit */ || (UCHAR(*p) >= '8')) { break; } count = 3; - result = (result << 3) + (*p - '0'); + result = (unsigned char)((result << 3) + (*p - '0')); p++; if ((numBytes == 3) || !isdigit(UCHAR(*p)) /* INTL: digit */ || (UCHAR(*p) >= '8')) { break; } count = 4; - result = UCHAR((result << 3) + (*p - '0')); + result = (unsigned char)((result << 3) + (*p - '0')); break; } @@ -933,15 +972,14 @@ TclParseBackslash( */ if (Tcl_UtfCharComplete(p, numBytes - 1)) { - count = Tcl_UtfToUniChar(p, &unichar) + 1; /* +1 for '\' */ + count = Tcl_UtfToUniChar(p, &result) + 1; /* +1 for '\' */ } else { char utfBytes[TCL_UTF_MAX]; memcpy(utfBytes, p, (size_t) (numBytes - 1)); utfBytes[numBytes - 1] = '\0'; - count = Tcl_UtfToUniChar(utfBytes, &unichar) + 1; + count = Tcl_UtfToUniChar(utfBytes, &result) + 1; } - result = unichar; break; } @@ -949,7 +987,7 @@ TclParseBackslash( if (readPtr != NULL) { *readPtr = count; } - return Tcl_UniCharToUtf(result, dst); + return Tcl_UniCharToUtf((int) result, dst); } /* @@ -961,11 +999,11 @@ TclParseBackslash( * defined by Tcl's parsing rules. * * Results: - * Records in parsePtr information about the parse. Returns the number of - * bytes consumed. + * Records in parsePtr information about the parse. Returns the number of + * bytes consumed. * * Side effects: - * None. + * None. * *---------------------------------------------------------------------- */ @@ -1118,7 +1156,7 @@ ParseTokens( } /* - * This is a variable reference. Call Tcl_ParseVarName to do all + * This is a variable reference. Call Tcl_ParseVarName to do all * the dirty work of parsing the name. */ @@ -1142,14 +1180,15 @@ ParseTokens( } /* - * Command substitution. Call Tcl_ParseCommand recursively (and + * Command substitution. Call Tcl_ParseCommand recursively (and * repeatedly) to parse the nested command(s), then throw away the * parse information. */ src++; numBytes--; - nestedPtr = TclStackAlloc(parsePtr->interp, sizeof(Tcl_Parse)); + nestedPtr = (Tcl_Parse *) + TclStackAlloc(parsePtr->interp, sizeof(Tcl_Parse)); while (1) { if (Tcl_ParseCommand(parsePtr->interp, src, numBytes, 1, nestedPtr) != TCL_OK) { @@ -1295,7 +1334,7 @@ Tcl_FreeParse( * call to Tcl_ParseCommand. */ { if (parsePtr->tokenPtr != parsePtr->staticTokens) { - ckfree(parsePtr->tokenPtr); + ckfree((char *) parsePtr->tokenPtr); parsePtr->tokenPtr = parsePtr->staticTokens; } } @@ -1345,9 +1384,7 @@ Tcl_ParseVarName( { Tcl_Token *tokenPtr; register const char *src; - unsigned char c; - int varIndex, offset; - Tcl_UniChar ch; + int varIndex; unsigned array; if ((numBytes == 0) || (start == NULL)) { @@ -1430,22 +1467,12 @@ Tcl_ParseVarName( tokenPtr->numComponents = 0; while (numBytes) { - if (Tcl_UtfCharComplete(src, numBytes)) { - offset = Tcl_UtfToUniChar(src, &ch); - } else { - char utfBytes[TCL_UTF_MAX]; - - memcpy(utfBytes, src, (size_t) numBytes); - utfBytes[numBytes] = '\0'; - offset = Tcl_UtfToUniChar(utfBytes, &ch); - } - c = UCHAR(ch); - if (isalnum(c) || (c == '_')) { /* INTL: ISO only, UCHAR. */ - src += offset; - numBytes -= offset; + if (TclIsBareword(*src)) { + src += 1; + numBytes -= 1; continue; } - if ((c == ':') && (numBytes != 1) && (src[1] == ':')) { + if ((src[0] == ':') && (numBytes != 1) && (src[1] == ':')) { src += 2; numBytes -= 2; while (numBytes && (*src == ':')) { @@ -1546,7 +1573,8 @@ Tcl_ParseVar( { register Tcl_Obj *objPtr; int code; - Tcl_Parse *parsePtr = TclStackAlloc(interp, sizeof(Tcl_Parse)); + Tcl_Parse *parsePtr = (Tcl_Parse *) + TclStackAlloc(interp, sizeof(Tcl_Parse)); if (Tcl_ParseVarName(interp, start, -1, parsePtr, 0) != TCL_OK) { TclStackFree(interp, parsePtr); @@ -1567,6 +1595,7 @@ Tcl_ParseVar( code = TclSubstTokens(interp, parsePtr->tokenPtr, parsePtr->numTokens, NULL, 1, NULL, NULL); + Tcl_FreeParse(parsePtr); TclStackFree(interp, parsePtr); if (code != TCL_OK) { return NULL; @@ -1629,7 +1658,7 @@ Tcl_ParseBraces( * the string consists of all bytes up to the * first null character. */ register Tcl_Parse *parsePtr, - /* Structure to fill in with information about + /* Structure to fill in with information about * the string. */ int append, /* Non-zero means append tokens to existing * information in parsePtr; zero means ignore @@ -1830,7 +1859,7 @@ Tcl_ParseQuotedString( * the string consists of all bytes up to the * first null character. */ register Tcl_Parse *parsePtr, - /* Structure to fill in with information about + /* Structure to fill in with information about * the string. */ int append, /* Non-zero means append tokens to existing * information in parsePtr; zero means ignore @@ -1878,42 +1907,33 @@ Tcl_ParseQuotedString( /* *---------------------------------------------------------------------- * - * TclSubstParse -- + * Tcl_SubstObj -- + * + * This function performs the substitutions specified on the given string + * as described in the user documentation for the "subst" Tcl command. * - * Token parser used by the [subst] command. Parses the string made up of - * 'numBytes' bytes starting at 'bytes'. Parsing is controlled by the - * flags argument to provide support for the -nobackslashes, -nocommands, - * and -novariables options, as represented by the flag values - * TCL_SUBST_BACKSLASHES, TCL_SUBST_COMMANDS, TCL_SUBST_VARIABLES. - * * Results: - * None. + * A Tcl_Obj* containing the substituted string, or NULL to indicate that + * an error occurred. * * Side effects: - * The Tcl_Parse struct '*parsePtr' is filled with parse results. - * The caller is expected to eventually call Tcl_FreeParse() to properly - * cleanup the value written there. - * - * If a parse error occurs, the Tcl_InterpState value '*statePtr' is - * filled with the state created by that error. When *statePtr is written - * to, the caller is expected to make the required calls to either - * Tcl_RestoreInterpState() or Tcl_DiscardInterpState() to dispose of the - * value written there. + * See the user documentation. * *---------------------------------------------------------------------- */ -void -TclSubstParse( - Tcl_Interp *interp, - const char *bytes, - int numBytes, - int flags, - Tcl_Parse *parsePtr, - Tcl_InterpState *statePtr) +Tcl_Obj * +Tcl_SubstObj( + Tcl_Interp *interp, /* Interpreter in which substitution occurs */ + Tcl_Obj *objPtr, /* The value to be substituted. */ + int flags) /* What substitutions to do. */ { - int length = numBytes; - const char *p = bytes; + int length, tokensLeft, code; + Tcl_Token *endTokenPtr; + Tcl_Obj *result, *errMsg = NULL; + const char *p = TclGetStringFromObj(objPtr, &length); + Tcl_Parse *parsePtr = (Tcl_Parse *) + TclStackAlloc(interp, sizeof(Tcl_Parse)); TclParseInit(interp, p, length, parsePtr); @@ -1925,11 +1945,12 @@ TclSubstParse( if (TCL_OK != ParseTokens(p, length, /* mask */ 0, flags, parsePtr)) { /* - * There was a parse error. Save the interpreter state for possible - * error reporting later. + * There was a parse error. Save the error message for possible + * reporting later. */ - *statePtr = Tcl_SaveInterpState(interp, TCL_ERROR); + errMsg = Tcl_GetObjResult(interp); + Tcl_IncrRefCount(errMsg); /* * We need to re-parse to get the portion of the string we can [subst] @@ -1995,10 +2016,10 @@ TclSubstParse( parsePtr->tokenPtr + parsePtr->numTokens - 2; if (varTokenPtr->type != TCL_TOKEN_VARIABLE) { - Tcl_Panic("TclSubstParse: programming error"); + Tcl_Panic("Tcl_SubstObj: programming error"); } if (varTokenPtr[1].type != TCL_TOKEN_TEXT) { - Tcl_Panic("TclSubstParse: programming error"); + Tcl_Panic("Tcl_SubstObj: programming error"); } parsePtr->numTokens -= 2; } @@ -2027,7 +2048,7 @@ TclSubstParse( Tcl_Token *tokenPtr; const char *lastTerm = parsePtr->term; - Tcl_Parse *nestedPtr = + Tcl_Parse *nestedPtr = (Tcl_Parse *) TclStackAlloc(interp, sizeof(Tcl_Parse)); while (TCL_OK == @@ -2072,8 +2093,63 @@ TclSubstParse( break; default: - Tcl_Panic("bad parse in TclSubstParse: %c", p[length]); + Tcl_Panic("bad parse in Tcl_SubstObj: %c", p[length]); + } + } + + /* + * Next, substitute the parsed tokens just as in normal Tcl evaluation. + */ + + endTokenPtr = parsePtr->tokenPtr + parsePtr->numTokens; + tokensLeft = parsePtr->numTokens; + code = TclSubstTokens(interp, endTokenPtr - tokensLeft, tokensLeft, + &tokensLeft, 1, NULL, NULL); + if (code == TCL_OK) { + Tcl_FreeParse(parsePtr); + TclStackFree(interp, parsePtr); + if (errMsg != NULL) { + Tcl_SetObjResult(interp, errMsg); + Tcl_DecrRefCount(errMsg); + return NULL; + } + return Tcl_GetObjResult(interp); + } + + result = Tcl_NewObj(); + while (1) { + switch (code) { + case TCL_ERROR: + Tcl_FreeParse(parsePtr); + TclStackFree(interp, parsePtr); + Tcl_DecrRefCount(result); + if (errMsg != NULL) { + Tcl_DecrRefCount(errMsg); + } + return NULL; + case TCL_BREAK: + tokensLeft = 0; /* Halt substitution */ + default: + Tcl_AppendObjToObj(result, Tcl_GetObjResult(interp)); } + + if (tokensLeft == 0) { + Tcl_FreeParse(parsePtr); + TclStackFree(interp, parsePtr); + if (errMsg != NULL) { + if (code != TCL_BREAK) { + Tcl_DecrRefCount(result); + Tcl_SetObjResult(interp, errMsg); + Tcl_DecrRefCount(errMsg); + return NULL; + } + Tcl_DecrRefCount(errMsg); + } + return result; + } + + code = TclSubstTokens(interp, endTokenPtr - tokensLeft, tokensLeft, + &tokensLeft, 1, NULL, NULL); } } @@ -2088,13 +2164,13 @@ TclSubstParse( * non-TCL_OK completion code arises. * * Results: - * The return value is a standard Tcl completion code. The result in - * interp is the substituted value, or an error message if TCL_ERROR is - * returned. If tokensLeftPtr is not NULL, then it points to an int where - * the number of tokens remaining to be processed is written. + * The return value is a standard Tcl completion code. The result in + * interp is the substituted value, or an error message if TCL_ERROR is + * returned. If tokensLeftPtr is not NULL, then it points to an int where + * the number of tokens remaining to be processed is written. * * Side effects: - * Can be anything, depending on the types of substitution done. + * Can be anything, depending on the types of substitution done. * *---------------------------------------------------------------------- */ @@ -2112,30 +2188,29 @@ TclSubstTokens( * integer representing the number of tokens * left to be substituted will be written */ int line, /* The line the script starts on. */ - int *clNextOuter, /* Information about an outer context for */ - const char *outerScript) /* continuation line data. This is set by - * EvalEx() to properly handle [...]-nested - * commands. The 'outerScript' refers to the - * most-outer script containing the embedded - * command, which is refered to by 'script'. - * The 'clNextOuter' refers to the current - * entry in the table of continuation lines in - * this "master script", and the character - * offsets are relative to the 'outerScript' - * as well. - * - * If outerScript == script, then this call is - * for words in the outer-most script or - * command. See Tcl_EvalEx and TclEvalObjEx - * for the places generating arguments for - * which this is true. */ + int* clNextOuter, /* Information about an outer context for */ + CONST char* outerScript) /* continuation line data. This is set by + * EvalEx() to properly handle [...]-nested + * commands. The 'outerScript' refers to the + * most-outer script containing the embedded + * command, which is refered to by 'script'. The + * 'clNextOuter' refers to the current entry in + * the table of continuation lines in this + * "master script", and the character offsets are + * relative to the 'outerScript' as well. + * + * If outerScript == script, then this call is for + * words in the outer-most script/command. See + * Tcl_EvalEx() and TclEvalObjEx() for the places + * generating arguments for which this is true. + */ { Tcl_Obj *result; int code = TCL_OK; #define NUM_STATIC_POS 20 int isLiteral, maxNumCL, numCL, i, adjust; - int *clPosition = NULL; - Interp *iPtr = (Interp *) interp; + int* clPosition = NULL; + Interp* iPtr = (Interp*) interp; int inFile = iPtr->evalFlags & TCL_EVAL_FILE; /* @@ -2152,24 +2227,24 @@ TclSubstTokens( * For the handling of continuation lines in literals we first check if * this is actually a literal. For if not we can forego the additional * processing. Otherwise we pre-allocate a small table to store the - * locations of all continuation lines we find in this literal, if any. - * The table is extended if needed. + * locations of all continuation lines we find in this literal, if + * any. The table is extended if needed. */ - numCL = 0; - maxNumCL = 0; + numCL = 0; + maxNumCL = 0; isLiteral = 1; for (i=0 ; i < count; i++) { - if ((tokenPtr[i].type != TCL_TOKEN_TEXT) - && (tokenPtr[i].type != TCL_TOKEN_BS)) { + if ((tokenPtr[i].type != TCL_TOKEN_TEXT) && + (tokenPtr[i].type != TCL_TOKEN_BS)) { isLiteral = 0; break; } } if (isLiteral) { - maxNumCL = NUM_STATIC_POS; - clPosition = ckalloc(maxNumCL * sizeof(int)); + maxNumCL = NUM_STATIC_POS; + clPosition = (int*) ckalloc (maxNumCL*sizeof(int)); } adjust = 0; @@ -2190,7 +2265,6 @@ TclSubstTokens( appendByteLength = TclParseBackslash(tokenPtr->start, tokenPtr->size, NULL, utfCharBytes); append = utfCharBytes; - /* * If the backslash sequence we found is in a literal, and * represented a continuation line, we compute and store its @@ -2206,11 +2280,10 @@ TclSubstTokens( * correction. */ - if ((appendByteLength == 1) && (utfCharBytes[0] == ' ') - && (tokenPtr->start[1] == '\n')) { + if ((appendByteLength == 1) && (utfCharBytes[0] == ' ') && + (tokenPtr->start[1] == '\n')) { if (isLiteral) { int clPos; - if (result == 0) { clPos = 0; } else { @@ -2219,18 +2292,19 @@ TclSubstTokens( if (numCL >= maxNumCL) { maxNumCL *= 2; - clPosition = ckrealloc(clPosition, - maxNumCL * sizeof(int)); + clPosition = (int*) ckrealloc ((char*)clPosition, + maxNumCL*sizeof(int)); } clPosition[numCL] = clPos; - numCL++; + numCL ++; } - adjust++; + adjust ++; } break; case TCL_TOKEN_COMMAND: { - /* TIP #280: Transfer line information to nested command */ + Interp *iPtr = (Interp *) interp; + iPtr->numLevels++; code = TclInterpReady(interp); if (code == TCL_OK) { @@ -2239,27 +2313,21 @@ TclSubstTokens( */ int theline; - - TclAdvanceContinuations(&line, &clNextOuter, - tokenPtr->start - outerScript); + TclAdvanceContinuations (&line, &clNextOuter, + tokenPtr->start - outerScript); theline = line + adjust; + /* TIP #280: Transfer line information to nested command */ code = TclEvalEx(interp, tokenPtr->start+1, tokenPtr->size-2, 0, theline, clNextOuter, outerScript); - - TclAdvanceLines(&line, tokenPtr->start+1, - tokenPtr->start + tokenPtr->size - 1); - /* * Restore flag reset by nested eval for future bracketed * commands and their cmdframe setup */ - - if (inFile) { + if (inFile) { iPtr->evalFlags |= TCL_EVAL_FILE; } } iPtr->numLevels--; - TclResetCancellation(interp, 0); appendObj = Tcl_GetObjResult(interp); break; } @@ -2358,7 +2426,6 @@ TclSubstTokens( if (code != TCL_ERROR) { /* Keep error message in result! */ if (result != NULL) { Tcl_SetObjResult(interp, result); - /* * If the code found continuation lines (which implies that this * word is a literal), then we store the accumulated table of @@ -2377,7 +2444,7 @@ TclSubstTokens( */ if (maxNumCL) { - ckfree(clPosition); + ckfree ((char*) clPosition); } } else { Tcl_ResetResult(interp); @@ -2496,56 +2563,6 @@ TclObjCommandComplete( } /* - *---------------------------------------------------------------------- - * - * TclIsLocalScalar -- - * - * Check to see if a given string is a legal scalar variable name with no - * namespace qualifiers or substitutions. - * - * Results: - * Returns 1 if the variable is a local scalar. - * - * Side effects: - * None. - * - *---------------------------------------------------------------------- - */ - -int -TclIsLocalScalar( - const char *src, - int len) -{ - const char *p; - const char *lastChar = src + (len - 1); - - for (p=src ; p<=lastChar ; p++) { - if ((CHAR_TYPE(*p) != TYPE_NORMAL) - && (CHAR_TYPE(*p) != TYPE_COMMAND_END)) { - /* - * TCL_COMMAND_END is returned for the last character of the - * string. By this point we know it isn't an array or namespace - * reference. - */ - - return 0; - } - if (*p == '(') { - if (*lastChar == ')') { /* We have an array element */ - return 0; - } - } else if (*p == ':') { - if ((p != lastChar) && *(p+1) == ':') { /* qualified name */ - return 0; - } - } - } - - return 1; -} - -/* * Local Variables: * mode: c * c-basic-offset: 4 |
