Tcl Source Code

int
Tcl_RegexpObjCmd(
    ClientData dummy,		/* Not used. */
    Tcl_Interp *interp,		/* Current interpreter. */
    int objc,			/* Number of arguments. */
    Tcl_Obj *const objv[])	/* Argument objects. */
{
    size_t offset, stringLength, matchLength, cflags, eflags;
    int i, indices, match, about, all, doinline, numMatchesSaved;

    Tcl_RegExp regExpr;
    Tcl_Obj *objPtr, *startIndex = NULL, *resultPtr = NULL;
    Tcl_RegExpInfo info;
    static const char *const options[] = {
	"-all",		"-about",	"-indices",	"-inline",
	"-expanded",	"-line",	"-linestop",	"-lineanchor",
	"-nocase",	"-start",	"--",		NULL

		resultPtr = Tcl_NewObj();
	    }
	}
	for (i = 0; i < objc; i++) {
	    Tcl_Obj *newPtr;

	    if (indices) {
		size_t start, end;
		Tcl_Obj *objs[2];

		/*
		 * Only adjust the match area if there was a match for that
		 * area. (Scriptics Bug 4391/SF Bug #219232)
		 */

		if (i <= (int)info.nsubs && info.matches[i].start != TCL_INDEX_NONE) {
		    start = offset + info.matches[i].start;
		    end = offset + info.matches[i].end;

		    /*
		     * Adjust index so it refers to the last character in the
		     * match instead of the first character after the match.
		     */

		    if (end + 1 >= offset + 1) {
			end--;
		    }
		} else {
		    start = TCL_INDEX_NONE;
		    end = TCL_INDEX_NONE;
		}

		objs[0] = TclNewWideIntObjFromSize(start);
		objs[1] = TclNewWideIntObjFromSize(end);

		newPtr = Tcl_NewListObj(2, objs);
	    } else {
		if (i <= (int)info.nsubs) {
		    newPtr = Tcl_GetRange(objPtr,
			    offset + info.matches[i].start,
			    offset + info.matches[i].end - 1);

int
Tcl_RegsubObjCmd(
    ClientData dummy,		/* Not used. */
    Tcl_Interp *interp,		/* Current interpreter. */
    int objc,			/* Number of arguments. */
    Tcl_Obj *const objv[])	/* Argument objects. */
{
    int result, cflags, all, match, command, numParts;
    size_t idx, wlen, wsublen = 0, offset, numMatches;
    size_t start, end, subStart, subEnd;
    Tcl_RegExp regExpr;
    Tcl_RegExpInfo info;
    Tcl_Obj *resultPtr, *subPtr, *objPtr, *startIndex = NULL;
    Tcl_UniChar ch, *wsrc, *wfirstChar, *wstring, *wsubspec = 0, *wend;

    static const char *const options[] = {
	"-all",		"-command",	"-expanded",	"-line",

    cflags = TCL_REG_ADVANCED;
    all = 0;
    offset = TCL_INDEX_START;
    command = 0;
    resultPtr = NULL;

    for (idx = 1; idx < (size_t)objc; idx++) {
	const char *name;
	int index;

	name = TclGetString(objv[idx]);
	if (name[0] != '-') {
	    break;
	}

	    cflags |= TCL_REG_NLSTOP;
	    break;
	case REGSUB_LINEANCHOR:
	    cflags |= TCL_REG_NLANCH;
	    break;
	case REGSUB_START: {
	    size_t temp;
	    if (++idx >= (size_t)objc) {
		goto endOfForLoop;
	    }
	    if (TclGetIntForIndexM(interp, objv[idx], TCL_INDEX_START, &temp) != TCL_OK) {
		goto optionError;
	    }
	    if (startIndex) {
		Tcl_DecrRefCount(startIndex);
	    }
	    startIndex = objv[idx];
	    Tcl_IncrRefCount(startIndex);
	    break;
	}
	case REGSUB_LAST:
	    idx++;
	    goto endOfForLoop;
	}
    }

  endOfForLoop:
    if ((size_t)objc < idx + 3 || (size_t)objc > idx + 4) {
	Tcl_WrongNumArgs(interp, 1, objv,
		"?-option ...? exp string subSpec ?varName?");
    optionError:
	if (startIndex) {
	    Tcl_DecrRefCount(startIndex);
	}
	return TCL_ERROR;

	    int numArgs;

	    Tcl_ListObjGetElements(interp, subPtr, &numParts, &parts);
	    numArgs = numParts + info.nsubs + 1;
	    args = Tcl_Alloc(sizeof(Tcl_Obj*) * numArgs);
	    memcpy(args, parts, sizeof(Tcl_Obj*) * numParts);

	    for (idx = 0 ; idx <= info.nsubs ; idx++) {
		subStart = info.matches[idx].start;
		subEnd = info.matches[idx].end;
		if ((subStart >= 0) && (subEnd >= 0)) {
		    args[idx + numParts] = Tcl_NewUnicodeObj(
			    wstring + offset + subStart, subEnd - subStart);
		} else {
		    args[idx + numParts] = Tcl_NewObj();

	     * because Tcl_EvalObjv is "hairy monster" in terms of refcount
	     * handling, being able to optionally add references to any of its
	     * argument words. We'll drop the local refs immediately
	     * afterwards; subPtr is handled in the main exit stanza.
	     */

	    result = Tcl_EvalObjv(interp, numArgs, args, 0);
	    for (idx = 0 ; idx <= info.nsubs ; idx++) {
		TclDecrRefCount(args[idx + numParts]);
	    }
	    Tcl_Free(args);
	    if (result != TCL_OK) {
		if (result == TCL_ERROR) {
		    Tcl_AppendObjToErrorInfo(interp, Tcl_ObjPrintf(
			    "\n    (%s substitution computation script)",

	    }

	    if (wfirstChar != wsrc) {
		Tcl_AppendUnicodeToObj(resultPtr, wfirstChar,
			wsrc - wfirstChar);
	    }

	    if (idx <= info.nsubs) {
		subStart = info.matches[idx].start;
		subEnd = info.matches[idx].end;
		if ((subStart >= 0) && (subEnd >= 0)) {
		    Tcl_AppendUnicodeToObj(resultPtr,
			    wstring + offset + subStart, subEnd - subStart);
		}
	    }

	    Tcl_ListObjAppendElement(NULL, listPtr, objPtr);
	    stringPtr = p + 1;
	}
	TclNewStringObj(objPtr, stringPtr, end - stringPtr);
	Tcl_ListObjAppendElement(NULL, listPtr, objPtr);
    } else {
	const char *element, *p, *splitEnd;
	size_t splitLen;
	Tcl_UniChar splitChar = 0;

	/*
	 * Normal case: split on any of a given set of characters. Discard
	 * instances of the split characters.
	 */

static int
StringMapCmd(
    ClientData dummy,		/* Not used. */
    Tcl_Interp *interp,		/* Current interpreter. */
    int objc,			/* Number of arguments. */
    Tcl_Obj *const objv[])	/* Argument objects. */
{
    size_t length1, length2,  mapElemc, index;

    int nocase = 0, mapWithDict = 0, copySource = 0;
    Tcl_Obj **mapElemv, *sourceObj, *resultPtr;
    Tcl_UniChar *ustring1, *ustring2, *p, *end;
    int (*strCmpFn)(const Tcl_UniChar*, const Tcl_UniChar*, size_t);

    if (objc < 3 || objc > 4) {
	Tcl_WrongNumArgs(interp, 1, objv, "?-nocase? charMap string");

	Tcl_DictSearch search;

	/*
	 * We know the type exactly, so all dict operations will succeed for
	 * sure. This shortens this code quite a bit.
	 */

	Tcl_DictObjSize(interp, objv[objc-2], &i);
	if (i == 0) {
	    /*
	     * Empty charMap, just return whatever string was given.
	     */

	    Tcl_SetObjResult(interp, objv[objc-1]);
	    return TCL_OK;
	}

	mapElemc = 2 * i;
	mapWithDict = 1;

	/*
	 * Copy the dictionary out into an array; that's the easiest way to
	 * adapt this code...
	 */

	mapElemv = TclStackAlloc(interp, sizeof(Tcl_Obj *) * mapElemc);
	Tcl_DictObjFirst(interp, objv[objc-2], &search, mapElemv+0,
		mapElemv+1, &done);
	for (index=2 ; index<mapElemc ; index+=2) {
	    Tcl_DictObjNext(&search, mapElemv+index, mapElemv+index+1, &done);
	}
	Tcl_DictObjDone(&search);
    } else {
	int i;
	if (TclListObjGetElements(interp, objv[objc-2], &i,
		&mapElemv) != TCL_OK) {
	    return TCL_ERROR;
	}
	mapElemc = i;
	if (mapElemc == 0) {
	    /*
	     * empty charMap, just return whatever string was given.
	     */

	    Tcl_SetObjResult(interp, objv[objc-1]);
	    return TCL_OK;

int
TclNRSwitchObjCmd(
    ClientData dummy,		/* Not used. */
    Tcl_Interp *interp,		/* Current interpreter. */
    int objc,			/* Number of arguments. */
    Tcl_Obj *const objv[])	/* Argument objects. */
{
    int i, index, mode, foundmode, splitObjs, numMatchesSaved;
    int noCase;
    size_t patternLength, j;
    const char *pattern;
    Tcl_Obj *stringObj, *indexVarObj, *matchVarObj;
    Tcl_Obj *const *savedObjv = objv;
    Tcl_RegExp regExpr = NULL;
    Interp *iPtr = (Interp *) interp;
    int pc = 0;
    int bidx = 0;		/* Index of body argument. */

	} else {
	    matchesObj = NULL;
	}
	if (indexVarObj != NULL) {
	    TclNewObj(indicesObj);
	}

	for (j=0 ; j<=info.nsubs ; j++) {
	    if (indexVarObj != NULL) {
		Tcl_Obj *rangeObjAry[2];

		if (info.matches[j].end + 1 > 1) {
		    rangeObjAry[0] = TclNewWideIntObjFromSize(info.matches[j].start);
		    rangeObjAry[1] = TclNewWideIntObjFromSize(info.matches[j].end-1);
		} else {

	    for (k=0; k < objc; k++) {
		ctxPtr->line[k] = -1;
	    }
	}
    }

    for (j = i + 1; ; j += 2) {
	if (j >= (size_t)objc) {
	    /*
	     * This shouldn't happen since we've checked that the last body is
	     * not a continuation...
	     */

	    Tcl_Panic("fall-out when searching for body to match pattern");
	}

void
TclRegExpRangeUniChar(
    Tcl_RegExp re,		/* Compiled regular expression that has been
				 * passed to Tcl_RegExpExec. */
    size_t index,			/* 0 means give the range of the entire match,
				 * > 0 means give the range of a matching
				 * subrange, TCL_INDEX_NONE means the range of the
				 * rm_extend field. */
    size_t *startPtr,		/* Store address of first character in
				 * (sub-)range here. */
    size_t *endPtr)		/* Store address of character just after last
				 * in (sub-)range here. */
{
    TclRegexp *regexpPtr = (TclRegexp *) re;

    if ((regexpPtr->flags&REG_EXPECT) && (index == TCL_INDEX_NONE)) {
	*startPtr = regexpPtr->details.rm_extend.rm_so;
	*endPtr = regexpPtr->details.rm_extend.rm_eo;
    } else if (index + 1 > regexpPtr->re.re_nsub + 1) {
	*startPtr = TCL_INDEX_NONE;
	*endPtr = TCL_INDEX_NONE;
    } else {
	*startPtr = regexpPtr->matches[index].rm_so;
	*endPtr = regexpPtr->matches[index].rm_eo;
    }
}

/*