Tcl Source Code

Check-in [e6e6fee0e1]
Login
Bounty program for improvements to Tcl and certain Tcl packages.
Tcl 2019 Conference, Houston/TX, US, Nov 4-8
Send your abstracts to [email protected]
or submit via the online form by Sep 9.

Many hyperlinks are disabled.
Use anonymous login to enable hyperlinks.

Overview
Comment:Route all [string repeat] operations through a common implementation. Code that to preserve bytearrays, eliminate unnecessary string rep generation, increase efficiency, and encapsulate access to internal rep details.
Downloads: Tarball | ZIP archive | SQL archive
Timelines: family | ancestors | descendants | both | trunk
Files: files | file ages | folders
SHA1: e6e6fee0e1aa40924675ecb370dc09503c98f36a
User & Date: dgp 2016-11-30 03:08:24
Context
2016-12-01
12:01
merge trunk check-in: 4162fbf706 user: limeboy tags: tip-456
11:44
compatability -> compatibility check-in: f94307dc05 user: jan.nijtmans tags: trunk
2016-11-30
14:27
Merge trunk check-in: 39918e8e32 user: jan.nijtmans tags: novem
03:16
merge trunk check-in: c2b9c28871 user: dgp tags: tip-445
03:08
Route all [string repeat] operations through a common implementation. Code that to preserve bytearra... check-in: e6e6fee0e1 user: dgp tags: trunk
2016-11-29
18:46
Where feasible, convert concatenation panics into errors. check-in: 3e5cafe2a5 user: dgp tags: trunk
Changes
Hide Diffs Unified Diffs Ignore Whitespace Patch

Changes to generic/tclCmdMZ.c.

2114
2115
2116
2117
2118
2119
2120
2121
2122
2123
2124
2125
2126
2127
2128
2129
2130
....
2134
2135
2136
2137
2138
2139
2140
2141
2142
2143
2144
2145
2146
2147
2148
2149
2150
2151
2152
2153
2154
2155
2156
2157
2158
2159
2160
2161
2162
2163
2164
2165
2166
2167
2168
2169
2170
2171
2172
2173
2174
2175
2176
2177
2178
2179
2180
2181
2182
2183
2184
2185

2186
2187
2188
2189
2190
2191
2192
2193
2194
2195
2196
2197
2198
2199
2200
2201
2202
2203
2204
2205
2206
2207
2208
2209
2210
2211
static int
StringReptCmd(
    ClientData dummy,		/* Not used. */
    Tcl_Interp *interp,		/* Current interpreter. */
    int objc,			/* Number of arguments. */
    Tcl_Obj *const objv[])	/* Argument objects. */
{
    const char *string1;
    char *string2;
    int count, index, length1, length2;
    Tcl_Obj *resultPtr;

    if (objc != 3) {
	Tcl_WrongNumArgs(interp, 1, objv, "string count");
	return TCL_ERROR;
    }

................................................................................

    /*
     * Check for cases that allow us to skip copying stuff.
     */

    if (count == 1) {
	Tcl_SetObjResult(interp, objv[1]);
	goto done;
    } else if (count < 1) {
	goto done;
    }
    string1 = TclGetStringFromObj(objv[1], &length1);
    if (length1 <= 0) {
	goto done;
    }

    /*
     * Only build up a string that has data. Instead of building it up with
     * repeated appends, we just allocate the necessary space once and copy
     * the string value in.
     *
     * We have to worry about overflow [Bugs 714106, 2561746].
     * At this point we know 1 <= length1 <= INT_MAX and 2 <= count <= INT_MAX.
     * We need to keep 2 <= length2 <= INT_MAX.
     */

    if (count > INT_MAX/length1) {
	Tcl_SetObjResult(interp, Tcl_ObjPrintf(
		"result exceeds max size for a Tcl value (%d bytes)",
		INT_MAX));
	Tcl_SetErrorCode(interp, "TCL", "MEMORY", NULL);
	return TCL_ERROR;
    }
    length2 = length1 * count;

    /*
     * Include space for the NUL.
     */

    string2 = attemptckalloc((unsigned) length2 + 1);
    if (string2 == NULL) {
	/*
	 * Alloc failed. Note that in this case we try to do an error message
	 * since this is a case that's most likely when the alloc is large and
	 * that's easy to do with this API. Note that if we fail allocating a
	 * short string, this will likely keel over too (and fatally).
	 */

	Tcl_SetObjResult(interp, Tcl_ObjPrintf(
		"string size overflow, out of memory allocating %u bytes",
		length2 + 1));
	Tcl_SetErrorCode(interp, "TCL", "MEMORY", NULL);

	return TCL_ERROR;
    }
    for (index = 0; index < count; index++) {
	memcpy(string2 + (length1 * index), string1, (size_t) length1);
    }
    string2[length2] = '\0';

    /*
     * We have to directly assign this instead of using Tcl_SetStringObj (and
     * indirectly TclInitStringRep) because that makes another copy of the
     * data.
     */

    TclNewObj(resultPtr);
    resultPtr->bytes = string2;
    resultPtr->length = length2;
    Tcl_SetObjResult(interp, resultPtr);

  done:
    return TCL_OK;
}
 
/*
 *----------------------------------------------------------------------
 *
 * StringRplcCmd --






<
<
|







 







|

<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
|

<

<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
>


<
<
<
<
<
<
<
<
<
<
<
<
<
<

<
<







2114
2115
2116
2117
2118
2119
2120


2121
2122
2123
2124
2125
2126
2127
2128
....
2132
2133
2134
2135
2136
2137
2138
2139
2140






















2141
2142

2143

















2144
2145
2146














2147


2148
2149
2150
2151
2152
2153
2154
static int
StringReptCmd(
    ClientData dummy,		/* Not used. */
    Tcl_Interp *interp,		/* Current interpreter. */
    int objc,			/* Number of arguments. */
    Tcl_Obj *const objv[])	/* Argument objects. */
{


    int count;
    Tcl_Obj *resultPtr;

    if (objc != 3) {
	Tcl_WrongNumArgs(interp, 1, objv, "string count");
	return TCL_ERROR;
    }

................................................................................

    /*
     * Check for cases that allow us to skip copying stuff.
     */

    if (count == 1) {
	Tcl_SetObjResult(interp, objv[1]);
	return TCL_OK;
    } else if (count < 1) {






















	return TCL_OK;
    }



















    if (TCL_OK != TclStringRepeat(interp, objv[1], count, &resultPtr)) {
	return TCL_ERROR;
    }














    Tcl_SetObjResult(interp, resultPtr);


    return TCL_OK;
}
 
/*
 *----------------------------------------------------------------------
 *
 * StringRplcCmd --

Changes to generic/tclInt.h.

3143
3144
3145
3146
3147
3148
3149


3150
3151
3152
3153
3154
3155
3156
MODULE_SCOPE int	TclStringLast(Tcl_Obj *needle, Tcl_Obj *haystack,
			    int last);
MODULE_SCOPE int	TclStringMatch(const char *str, int strLen,
			    const char *pattern, int ptnLen, int flags);
MODULE_SCOPE int	TclStringMatchObj(Tcl_Obj *stringObj,
			    Tcl_Obj *patternObj, int flags);
MODULE_SCOPE Tcl_Obj *	TclStringObjReverse(Tcl_Obj *objPtr);


MODULE_SCOPE void	TclSubstCompile(Tcl_Interp *interp, const char *bytes,
			    int numBytes, int flags, int line,
			    struct CompileEnv *envPtr);
MODULE_SCOPE int	TclSubstOptions(Tcl_Interp *interp, int numOpts,
			    Tcl_Obj *const opts[], int *flagPtr);
MODULE_SCOPE void	TclSubstParse(Tcl_Interp *interp, const char *bytes,
			    int numBytes, int flags, Tcl_Parse *parsePtr,






>
>







3143
3144
3145
3146
3147
3148
3149
3150
3151
3152
3153
3154
3155
3156
3157
3158
MODULE_SCOPE int	TclStringLast(Tcl_Obj *needle, Tcl_Obj *haystack,
			    int last);
MODULE_SCOPE int	TclStringMatch(const char *str, int strLen,
			    const char *pattern, int ptnLen, int flags);
MODULE_SCOPE int	TclStringMatchObj(Tcl_Obj *stringObj,
			    Tcl_Obj *patternObj, int flags);
MODULE_SCOPE Tcl_Obj *	TclStringObjReverse(Tcl_Obj *objPtr);
MODULE_SCOPE int	TclStringRepeat(Tcl_Interp *interp, Tcl_Obj *objPtr,
			    int count, Tcl_Obj **objPtrPtr);
MODULE_SCOPE void	TclSubstCompile(Tcl_Interp *interp, const char *bytes,
			    int numBytes, int flags, int line,
			    struct CompileEnv *envPtr);
MODULE_SCOPE int	TclSubstOptions(Tcl_Interp *interp, int numOpts,
			    Tcl_Obj *const opts[], int *flagPtr);
MODULE_SCOPE void	TclSubstParse(Tcl_Interp *interp, const char *bytes,
			    int numBytes, int flags, Tcl_Parse *parsePtr,

Changes to generic/tclStringObj.c.

2605
2606
2607
2608
2609
2610
2611












































































































































2612
2613
2614
2615
2616
2617
2618
	return TclGetStringFromObj(objPtr, (int *)sizePtr);
    }

    stringPtr = GET_STRING(objPtr);
    *sizePtr = stringPtr->allocated;
    return objPtr->bytes;
}












































































































































 
/*
 *---------------------------------------------------------------------------
 *
 * TclStringCatObjv --
 *
 *	Performs the [string cat] function.






>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>







2605
2606
2607
2608
2609
2610
2611
2612
2613
2614
2615
2616
2617
2618
2619
2620
2621
2622
2623
2624
2625
2626
2627
2628
2629
2630
2631
2632
2633
2634
2635
2636
2637
2638
2639
2640
2641
2642
2643
2644
2645
2646
2647
2648
2649
2650
2651
2652
2653
2654
2655
2656
2657
2658
2659
2660
2661
2662
2663
2664
2665
2666
2667
2668
2669
2670
2671
2672
2673
2674
2675
2676
2677
2678
2679
2680
2681
2682
2683
2684
2685
2686
2687
2688
2689
2690
2691
2692
2693
2694
2695
2696
2697
2698
2699
2700
2701
2702
2703
2704
2705
2706
2707
2708
2709
2710
2711
2712
2713
2714
2715
2716
2717
2718
2719
2720
2721
2722
2723
2724
2725
2726
2727
2728
2729
2730
2731
2732
2733
2734
2735
2736
2737
2738
2739
2740
2741
2742
2743
2744
2745
2746
2747
2748
2749
2750
2751
2752
2753
2754
2755
2756
2757
2758
	return TclGetStringFromObj(objPtr, (int *)sizePtr);
    }

    stringPtr = GET_STRING(objPtr);
    *sizePtr = stringPtr->allocated;
    return objPtr->bytes;
}
 
/*
 *---------------------------------------------------------------------------
 *
 * TclStringRepeat --
 *
 *	Performs the [string repeat] function.
 *
 * Results:
 * 	A standard Tcl result.
 *
 * Side effects:
 * 	Writes to *objPtrPtr the address of Tcl_Obj that is concatenation
 * 	of count copies of the value in objPtr.
 *
 *---------------------------------------------------------------------------
 */

int
TclStringRepeat(
    Tcl_Interp *interp,
    Tcl_Obj *objPtr,
    int count,
    Tcl_Obj **objPtrPtr)
{
    Tcl_Obj *objResultPtr;
    int length = 0, unichar = 0, done = 1;
    int binary = TclIsPureByteArray(objPtr);

    /* assert (count >= 2) */

    /*
     * Analyze to determine what representation result should be.
     * GOALS:	Avoid shimmering & string rep generation.
     * 		Produce pure bytearray when possible.
     * 		Error on overflow.
     */

    if (!binary) {
	if (objPtr->typePtr == &tclStringType) {
	    String *stringPtr = GET_STRING(objPtr);
	    if (stringPtr->hasUnicode) {
		unichar = 1;
	    }
	}
    }

    if (binary) {
	/* Result will be pure byte array. Pre-size it */
	Tcl_GetByteArrayFromObj(objPtr, &length);
    } else if (unichar) {
	/* Result will be pure Tcl_UniChar array. Pre-size it. */
	Tcl_GetUnicodeFromObj(objPtr, &length);
    } else {
	/* Result will be concat of string reps. Pre-size it. */
	Tcl_GetStringFromObj(objPtr, &length);
    }

    if (length == 0) {
	/* Any repeats of empty is empty. */
	*objPtrPtr = objPtr;
	return TCL_OK;
    }

    if (count > INT_MAX/length) {
	if (interp) {
	    Tcl_SetObjResult(interp, Tcl_ObjPrintf(
		    "max size for a Tcl value (%d bytes) exceeded", INT_MAX));
	    Tcl_SetErrorCode(interp, "TCL", "MEMORY", NULL);
	}
	return TCL_ERROR;
    }

    if (binary) {
	/* Efficiently produce a pure byte array result */
	objResultPtr = Tcl_IsShared(objPtr) ? Tcl_DuplicateObj(objPtr)
		: objPtr;

	Tcl_SetByteArrayLength(objResultPtr, count*length); /* PANIC? */
	Tcl_SetByteArrayLength(objResultPtr, length);
	while (count - done > done) {
	    Tcl_AppendObjToObj(objResultPtr, objResultPtr);
	    done *= 2;
	}
	TclAppendBytesToByteArray(objResultPtr,
		Tcl_GetByteArrayFromObj(objResultPtr, NULL),
		(count - done) * length);
    } else if (unichar) {
	/* Efficiently produce a pure Tcl_UniChar array result */
	if (Tcl_IsShared(objPtr)) {
	    objResultPtr = Tcl_NewUnicodeObj(Tcl_GetUnicode(objPtr), length);
	} else {
	    TclInvalidateStringRep(objPtr);
	    objResultPtr = objPtr;
	}

        if (0 == Tcl_AttemptSetObjLength(objResultPtr, count*length)) {
	    if (interp) {
		Tcl_SetObjResult(interp, Tcl_ObjPrintf(
			"string size overflow: unable to alloc %lu bytes",
			STRING_SIZE(count*length)));
		Tcl_SetErrorCode(interp, "TCL", "MEMORY", NULL);
	    }
	    return TCL_ERROR;
	}
	Tcl_SetObjLength(objResultPtr, length);
	while (count - done > done) {
	    Tcl_AppendObjToObj(objResultPtr, objResultPtr);
	    done *= 2;
	}
	Tcl_AppendUnicodeToObj(objResultPtr, Tcl_GetUnicode(objResultPtr),
		(count - done) * length);
    } else {
	/* Efficiently concatenate string reps */
	if (Tcl_IsShared(objPtr)) {
	    objResultPtr = Tcl_NewStringObj(Tcl_GetString(objPtr), length);
	} else {
	    TclFreeIntRep(objPtr);
	    objResultPtr = objPtr;
	}
        if (0 == Tcl_AttemptSetObjLength(objResultPtr, count*length)) {
	    if (interp) {
		Tcl_SetObjResult(interp, Tcl_ObjPrintf(
			"string size overflow: unable to alloc %u bytes",
			count*length));
		Tcl_SetErrorCode(interp, "TCL", "MEMORY", NULL);
	    }
	    return TCL_ERROR;
	}
	Tcl_SetObjLength(objResultPtr, length);
	while (count - done > done) {
	    Tcl_AppendObjToObj(objResultPtr, objResultPtr);
	    done *= 2;
	}
	Tcl_AppendToObj(objResultPtr, Tcl_GetString(objResultPtr),
		(count - done) * length);
    }
    *objPtrPtr = objResultPtr;
    return TCL_OK;
}
 
/*
 *---------------------------------------------------------------------------
 *
 * TclStringCatObjv --
 *
 *	Performs the [string cat] function.