Tcl Source Code

2013-05-22  Donal K. Fellows  <[email protected]>

	* generic/tclUtf.c (TclUtfCasecmp): [Bug 3613609]: Replace problematic
	uses of strcasecmp with a proper UTF-8-aware version. Affects both
	[lsearch -nocase] and [lsort -nocase].

2013-05-22  Donal K. Fellows  <[email protected]>

	* doc/file.n: [Bug 3613671]: Added note to portability section on the
	fact that [file owned] does not produce useful results on Windows.

2013-05-20  Donal K. Fellows  <[email protected]>

	case LSEARCH_INLINE:		/* -inline */
	    inlineReturn = 1;
	    break;
	case LSEARCH_INTEGER:		/* -integer */
	    dataType = INTEGER;
	    break;
	case LSEARCH_NOCASE:		/* -nocase */
	    strCmpFn = TclUtfCasecmp;
	    noCase = 1;
	    break;
	case LSEARCH_NOT:		/* -not */
	    negatedMatch = 1;
	    break;
	case LSEARCH_REAL:		/* -real */
	    dataType = REAL;

		    if (length == elemLen) {
			/*
			 * This split allows for more optimal compilation of
			 * memcmp/strcasecmp.
			 */

			if (noCase) {
			    match = (TclUtfCasecmp(bytes, patternBytes) == 0);
			} else {
			    match = (memcmp(bytes, patternBytes,
				    (size_t) length) == 0);
			}
		    }
		    break;

	    long a;

	    if (TclGetLongFromObj(sortInfo.interp, indexPtr, &a) != TCL_OK) {
		sortInfo.resultCode = TCL_ERROR;
		goto done1;
	    }
	    elementArray[i].collationKey.intValue = a;
	} else if (sortMode == SORTMODE_REAL) {
	    double a;

	    if (Tcl_GetDoubleFromObj(sortInfo.interp, indexPtr,
		    &a) != TCL_OK) {
		sortInfo.resultCode = TCL_ERROR;
		goto done1;
	    }

	Tcl_SetObjResult(interp, resultPtr);
    }

  done1:
    TclStackFree(interp, elementArray);

  done:
    if (sortMode == SORTMODE_COMMAND) {
	TclDecrRefCount(sortInfo.compareCmdPtr);
	TclDecrRefCount(listObj);
	sortInfo.compareCmdPtr = NULL;
    }
  done2:
    if (allocatedIndexVector) {
	TclStackFree(interp, sortInfo.indexv);

{
    int order = 0;

    if (infoPtr->sortMode == SORTMODE_ASCII) {
	order = strcmp(elemPtr1->collationKey.strValuePtr,
		elemPtr2->collationKey.strValuePtr);
    } else if (infoPtr->sortMode == SORTMODE_ASCII_NC) {
	order = TclUtfCasecmp(elemPtr1->collationKey.strValuePtr,
		elemPtr2->collationKey.strValuePtr);
    } else if (infoPtr->sortMode == SORTMODE_DICTIONARY) {
	order = DictionaryCompare(elemPtr1->collationKey.strValuePtr,
		elemPtr2->collationKey.strValuePtr);
    } else if (infoPtr->sortMode == SORTMODE_INTEGER) {
	long a, b;

	     * General options.
	     */

	case OPT_LAST:
	    i++;
	    goto finishedOptions;
	case OPT_NOCASE:
	    strCmpFn = TclUtfCasecmp;
	    noCase = 1;
	    break;

	    /*
	     * Handle the different switch mode options.
	     */

MODULE_SCOPE int	TclSubstTokens(Tcl_Interp *interp, Tcl_Token *tokenPtr,
			    int count, int *tokensLeftPtr, int line,
			    int *clNextOuter, const char *outerScript);
MODULE_SCOPE int	TclTrimLeft(const char *bytes, int numBytes,
			    const char *trim, int numTrim);
MODULE_SCOPE int	TclTrimRight(const char *bytes, int numBytes,
			    const char *trim, int numTrim);
MODULE_SCOPE int	TclUtfCasecmp(const char *cs, const char *ct);
MODULE_SCOPE Tcl_Obj *	TclpNativeToNormalized(ClientData clientData);
MODULE_SCOPE Tcl_Obj *	TclpFilesystemPathType(Tcl_Obj *pathPtr);
MODULE_SCOPE int	TclpDlopen(Tcl_Interp *interp, Tcl_Obj *pathPtr,
			    Tcl_LoadHandle *loadHandle,
			    Tcl_FSUnloadFileProc **unloadProcPtr, int flags);
MODULE_SCOPE int	TclpUtime(Tcl_Obj *pathPtr, struct utimbuf *tval);
#ifdef TCL_LOAD_FROM_MEMORY

	    if (ch1 != ch2) {
		return (ch1 - ch2);
	    }
	}
    }
    return 0;
}

/*
 *----------------------------------------------------------------------
 *
 * Tcl_UtfNcasecmp --
 *
 *	Compare UTF chars of string cs to string ct case insensitively.
 *	Replacement for strcasecmp in Tcl core, in places where UTF-8 should
 *	be handled.
 *
 * Results:
 *	Return <0 if cs < ct, 0 if cs == ct, or >0 if cs > ct.
 *
 * Side effects:
 *	None.
 *
 *----------------------------------------------------------------------
 */

int
TclUtfCasecmp(
    const char *cs,		/* UTF string to compare to ct. */
    const char *ct)		/* UTF string cs is compared to. */
{
    while (*cs && *ct) {
	Tcl_UniChar ch1, ch2;

	cs += TclUtfToUniChar(cs, &ch1);
	ct += TclUtfToUniChar(ct, &ch2);
	if (ch1 != ch2) {
	    ch1 = Tcl_UniCharToLower(ch1);
	    ch2 = Tcl_UniCharToLower(ch2);
	    if (ch1 != ch2) {
		return ch1 - ch2;
	    }
	}
    }
    return UCHAR(*cs) - UCHAR(*ct);
}


/*
 *----------------------------------------------------------------------
 *
 * Tcl_UniCharToUpper --
 *
 *	Compute the uppercase equivalent of the given Unicode character.

} [list ! ` AA c CC]
test cmdIL-4.34 {SortCompare procedure, -ascii option with -nocase option} {
    lsort -ascii -nocase {d e c b a d35 d300 100 20}
} {100 20 a b c d d300 d35 e}
test cmdIL-4.35 {SortCompare procedure, -ascii option with -nocase option} {
    lsort -ascii -nocase {d E c B a D35 d300 100 20}
} {100 20 a B c d d300 D35 E}
test cmdIL-4.36 {SortCompare procedure, UTF-8 with -nocase option} {
    scan [lsort -ascii -nocase [list \u101 \u100]] %c%c%c
} {257 32 256}
test cmdIL-4.37 {SortCompare procedure, UTF-8 with -nocase option} {
    scan [lsort -ascii -nocase [list a\u0000a a]] %c%c%c%c%c
} {97 32 97 0 97}
test cmdIL-4.38 {SortCompare procedure, UTF-8 with -nocase option} {
    scan [lsort -ascii -nocase [list a a\u0000a]] %c%c%c%c%c
} {97 32 97 0 97}

test cmdIL-5.1 {lsort with list style index} {
    lsort -ascii -decreasing -index {0 1} {
	{{Jim Alpha} 20000410}
	{{Joe Bravo} 19990320}
	{{Jacky Charlie} 19390911}
    }