Tcl Source Code

Check-in [2aa14f69f0]
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:Fix 2 failing test-cases, broken by some earlier commit
Downloads: Tarball | ZIP archive | SQL archive
Timelines: family | ancestors | descendants | both | no-wideint
Files: files | file ages | folders
SHA3-256: 2aa14f69f0e391dbb35d6bbc53789e06e1206cbd888aff53fdd713a13c1658a6
User & Date: jan.nijtmans 2017-10-31 11:37:29
Context
2017-10-31
12:39
more simplifications check-in: fc07c2bbab user: jan.nijtmans tags: no-wideint
11:37
Fix 2 failing test-cases, broken by some earlier commit check-in: 2aa14f69f0 user: jan.nijtmans tags: no-wideint
2017-10-30
16:46
Fix some pointer arthemeric (only visible on big-endian systems) check-in: d3adf52e7a user: jan.nijtmans tags: no-wideint
Changes
Hide Diffs Unified Diffs Ignore Whitespace Patch

Changes to generic/tclExecute.c.

5889
5890
5891
5892
5893
5894
5895

5896
5897
5898
5899
5900
5901
5902

5903
5904
5905
5906
5907
5908
5909
5910
5911

5912

5913
5914


5915
5916
5917
5918
5919
5920
5921
5922
....
5947
5948
5949
5950
5951
5952
5953
5954
5955
5956
5957
5958
5959
5960
5961
5962
5963
5964
....
6276
6277
6278
6279
6280
6281
6282
6283
6284
6285
6286


6287
6288
6289
6290
6291
6292
6293
6294
6295
6296
6297
6298
6299
6300
6301
6302
6303
6304
6305
6306
6307
6308
6309
6310
6311
....
6324
6325
6326
6327
6328
6329
6330
6331
6332
6333
6334
6335
6336
6337
6338
6339
6340
6341
6342
6343
6344
6345
6346
6347
6348
6349
6350
6351
6352
6353
6354
6355
6356
6357
6358
6359
6360
6361
6362
6363
6364
6365
6366
6367
6368
6369
6370
6371
....
8223
8224
8225
8226
8227
8228
8229
8230
8231
8232
8233
8234
8235
8236
8237
8238
....
8246
8247
8248
8249
8250
8251
8252
8253
8254
8255
8256
8257
8258
8259
8260
....
8304
8305
8306
8307
8308
8309
8310
8311
8312
8313
8314
8315
8316
8317
8318
8319
....
8977
8978
8979
8980
8981
8982
8983
8984
8985
8986
8987
8988
8989
8990
8991
     *	   Start of numeric operator instructions.
     */

    {
	ClientData ptr1, ptr2;
	int type1, type2;
	long l1, l2, lResult;


    case INST_NUM_TYPE:
	if (GetNumberFromObj(NULL, OBJ_AT_TOS, &ptr1, &type1) != TCL_OK) {
	    type1 = 0;
	} else if (type1 == TCL_NUMBER_LONG || type1 == TCL_NUMBER_WIDE) {
	    /* value is between LONG_MIN and LONG_MAX */
	    /* [string is integer] is -UINT_MAX to UINT_MAX range */

	    int i;

	    if (Tcl_GetIntFromObj(NULL, OBJ_AT_TOS, &i) != TCL_OK) {
		type1 = TCL_NUMBER_WIDE;
	    } else {
		type1 = TCL_NUMBER_LONG;
	    }
	} else if (type1 == TCL_NUMBER_BIG) {
	    /* value is an integer outside the WIDE_MIN to WIDE_MAX range */

	    /* [string is wideinteger] is -UWIDE_MAX to UWIDE_MAX range */

	    Tcl_WideInt w;



	    if (Tcl_GetWideIntFromObj(NULL, OBJ_AT_TOS, &w) == TCL_OK) {
		type1 = TCL_NUMBER_WIDE;
	    }
	}
	TclNewLongObj(objResultPtr, type1);
	TRACE(("\"%.20s\" => %d\n", O2S(OBJ_AT_TOS), type1));
	NEXT_INST_F(1, 1, 1);

................................................................................
	    iResult = (*pc == INST_NEQ);
	    goto foundResult;
	}
	if (valuePtr == value2Ptr) {
	    compare = MP_EQ;
	    goto convertComparison;
	}
	if ((type1 == TCL_NUMBER_LONG) && (type2 == TCL_NUMBER_LONG)) {
	    l1 = *((const Tcl_WideInt *)ptr1);
	    l2 = *((const Tcl_WideInt *)ptr2);
	    compare = (l1 < l2) ? MP_LT : ((l1 > l2) ? MP_GT : MP_EQ);
	} else {
	    compare = TclCompareTwoNumbers(valuePtr, value2Ptr);
	}

	/*
	 * Turn comparison outcome into appropriate result for opcode.
	 */
................................................................................

	/*
	 * Handle (long,long) arithmetic as best we can without going out to
	 * an external function.
	 */

	if ((type1 == TCL_NUMBER_LONG) && (type2 == TCL_NUMBER_LONG)) {
	    Tcl_WideInt w1, w2, wResult;

	    l1 = *((const Tcl_WideInt *)ptr1);
	    l2 = *((const Tcl_WideInt *)ptr2);



	    switch (*pc) {
	    case INST_ADD:
		w1 = (Tcl_WideInt) l1;
		w2 = (Tcl_WideInt) l2;
		wResult = w1 + w2;
		/*
		 * Check for overflow.
		 */

		if (Overflowing(w1, w2, wResult)) {
		    goto overflow;
		}
		goto wideResultOfArithmetic;

	    case INST_SUB:
		w1 = (Tcl_WideInt) l1;
		w2 = (Tcl_WideInt) l2;
		wResult = w1 - w2;
		/*
		 * Must check for overflow. The macro tests for overflows in
		 * sums by looking at the sign bits. As we have a subtraction
		 * here, we are adding -w2. As -w2 could in turn overflow, we
		 * test with ~w2 instead: it has the opposite sign bit to w2
		 * so it does the job. Note that the only "bad" case (w2==0)
................................................................................
		    NEXT_INST_F(1, 2, 1);
		}
		Tcl_SetWideIntObj(valuePtr, wResult);
		TRACE(("%s\n", O2S(valuePtr)));
		NEXT_INST_F(1, 1, 0);

	    case INST_DIV:
		if (l2 == 0) {
		    TRACE(("%s %s => DIVIDE BY ZERO\n",
			    O2S(valuePtr), O2S(value2Ptr)));
		    goto divideByZero;
		} else if ((l1 == LONG_MIN) && (l2 == -1)) {
		    /*
		     * Can't represent (-LONG_MIN) as a long.
		     */

		    goto overflow;
		}
		lResult = l1 / l2;

		/*
		 * Force Tcl's integer division rules.
		 * TODO: examine for logic simplification
		 */

		if (((lResult < 0) || ((lResult == 0) &&
			((l1 < 0 && l2 > 0) || (l1 > 0 && l2 < 0)))) &&
			((lResult * l2) != l1)) {
		    lResult -= 1;
		}
		goto longResultOfArithmetic;

	    case INST_MULT:
		if (((sizeof(long) >= 2*sizeof(int))
			&& (l1 <= INT_MAX) && (l1 >= INT_MIN)
			&& (l2 <= INT_MAX) && (l2 >= INT_MIN))
			|| ((sizeof(long) >= 2*sizeof(short))
			&& (l1 <= SHRT_MAX) && (l1 >= SHRT_MIN)
			&& (l2 <= SHRT_MAX) && (l2 >= SHRT_MIN))) {
		    lResult = l1 * l2;
		    goto longResultOfArithmetic;
		}
	    }

	    /*
	     * Fall through with INST_EXPON, INST_DIV and large multiplies.
	     */
	}
................................................................................
    case INST_LSHIFT:
    case INST_RSHIFT: {
	/*
	 * Reject negative shift argument.
	 */

	switch (type2) {
	case TCL_NUMBER_WIDE:
	case TCL_NUMBER_LONG:
	    invalid = (*((const Tcl_WideInt *)ptr2) < (Tcl_WideInt)0);
	    break;
	case TCL_NUMBER_BIG:
	    Tcl_TakeBignumFromObj(NULL, value2Ptr, &big2);
	    invalid = (mp_cmp_d(&big2, 0) == MP_LT);
	    mp_clear(&big2);
	    break;
................................................................................
	    return GENERAL_ARITHMETIC_ERROR;
	}

	/*
	 * Zero shifted any number of bits is still zero.
	 */

	if ((type1==TCL_NUMBER_LONG) && (*((const Tcl_WideInt *)ptr1) == (long)0)) {
	    return constants[0];
	}

	if (opcode == INST_LSHIFT) {
	    /*
	     * Large left shifts create integer overflow.
	     *
................................................................................
		 * not take us to the result of 0 or -1, but since we're using
		 * mp_div_2d to do the work, and it takes only an int
		 * argument, we draw the line there.
		 */

		switch (type1) {
		case TCL_NUMBER_LONG:
		    zero = (*(const Tcl_WideInt *)ptr1 > 0L);
		    break;
		case TCL_NUMBER_WIDE:
		    zero = (*(const Tcl_WideInt *)ptr1 > (Tcl_WideInt)0);
		    break;
		case TCL_NUMBER_BIG:
		    Tcl_TakeBignumFromObj(NULL, valuePtr, &big1);
		    zero = (mp_cmp_d(&big1, 0) == MP_GT);
		    mp_clear(&big1);
................................................................................
			goto overflowBasic;
		    }
		}
		break;

	    case INST_SUB:
		wResult = w1 - w2;
		if ((type1 == TCL_NUMBER_WIDE) || (type2 == TCL_NUMBER_WIDE))
		{
		    /*
		     * Must check for overflow. The macro tests for overflows
		     * in sums by looking at the sign bits. As we have a
		     * subtraction here, we are adding -w2. As -w2 could in
		     * turn overflow, we test with ~w2 instead: it has the
		     * opposite sign bit to w2 so it does the job. Note that






>





|

>









>

>


>
>
|







 







|
|
|
|







 







<
<
|
|
>
>



<
<











<
<







 







|



|

|




|






|
|
|
|

|



|
|

|
|
|
|







 







|
|







 







|







 







<
<







 







|







5889
5890
5891
5892
5893
5894
5895
5896
5897
5898
5899
5900
5901
5902
5903
5904
5905
5906
5907
5908
5909
5910
5911
5912
5913
5914
5915
5916
5917
5918
5919
5920
5921
5922
5923
5924
5925
5926
5927
5928
....
5953
5954
5955
5956
5957
5958
5959
5960
5961
5962
5963
5964
5965
5966
5967
5968
5969
5970
....
6282
6283
6284
6285
6286
6287
6288


6289
6290
6291
6292
6293
6294
6295


6296
6297
6298
6299
6300
6301
6302
6303
6304
6305
6306


6307
6308
6309
6310
6311
6312
6313
....
6326
6327
6328
6329
6330
6331
6332
6333
6334
6335
6336
6337
6338
6339
6340
6341
6342
6343
6344
6345
6346
6347
6348
6349
6350
6351
6352
6353
6354
6355
6356
6357
6358
6359
6360
6361
6362
6363
6364
6365
6366
6367
6368
6369
6370
6371
6372
6373
....
8225
8226
8227
8228
8229
8230
8231
8232
8233
8234
8235
8236
8237
8238
8239
8240
....
8248
8249
8250
8251
8252
8253
8254
8255
8256
8257
8258
8259
8260
8261
8262
....
8306
8307
8308
8309
8310
8311
8312


8313
8314
8315
8316
8317
8318
8319
....
8977
8978
8979
8980
8981
8982
8983
8984
8985
8986
8987
8988
8989
8990
8991
     *	   Start of numeric operator instructions.
     */

    {
	ClientData ptr1, ptr2;
	int type1, type2;
	long l1, l2, lResult;
    Tcl_WideInt w1, w2, wResult;

    case INST_NUM_TYPE:
	if (GetNumberFromObj(NULL, OBJ_AT_TOS, &ptr1, &type1) != TCL_OK) {
	    type1 = 0;
	} else if (type1 == TCL_NUMBER_LONG || type1 == TCL_NUMBER_WIDE) {
	    /* value is between WIDE_MIN and WIDE_MAX */
	    /* [string is integer] is -UINT_MAX to UINT_MAX range */
	    /* [string is wideinteger] is -UWIDE_MAX to UWIDE_MAX range */
	    int i;

	    if (Tcl_GetIntFromObj(NULL, OBJ_AT_TOS, &i) != TCL_OK) {
		type1 = TCL_NUMBER_WIDE;
	    } else {
		type1 = TCL_NUMBER_LONG;
	    }
	} else if (type1 == TCL_NUMBER_BIG) {
	    /* value is an integer outside the WIDE_MIN to WIDE_MAX range */
	    /* [string is integer] is -UINT_MAX to UINT_MAX range */
	    /* [string is wideinteger] is -UWIDE_MAX to UWIDE_MAX range */
	    int i;
	    Tcl_WideInt w;

	    if (Tcl_GetIntFromObj(NULL, OBJ_AT_TOS, &i) == TCL_OK) {
		type1 = TCL_NUMBER_LONG;
	    } else if (Tcl_GetWideIntFromObj(NULL, OBJ_AT_TOS, &w) == TCL_OK) {
		type1 = TCL_NUMBER_WIDE;
	    }
	}
	TclNewLongObj(objResultPtr, type1);
	TRACE(("\"%.20s\" => %d\n", O2S(OBJ_AT_TOS), type1));
	NEXT_INST_F(1, 1, 1);

................................................................................
	    iResult = (*pc == INST_NEQ);
	    goto foundResult;
	}
	if (valuePtr == value2Ptr) {
	    compare = MP_EQ;
	    goto convertComparison;
	}
	if ((type1 == TCL_NUMBER_LONG || type1 == TCL_NUMBER_WIDE) && (type2 == TCL_NUMBER_LONG || type2 == TCL_NUMBER_WIDE)) {
	    w1 = *((const Tcl_WideInt *)ptr1);
	    w2 = *((const Tcl_WideInt *)ptr2);
	    compare = (w1 < w2) ? MP_LT : ((w1 > w2) ? MP_GT : MP_EQ);
	} else {
	    compare = TclCompareTwoNumbers(valuePtr, value2Ptr);
	}

	/*
	 * Turn comparison outcome into appropriate result for opcode.
	 */
................................................................................

	/*
	 * Handle (long,long) arithmetic as best we can without going out to
	 * an external function.
	 */

	if ((type1 == TCL_NUMBER_LONG) && (type2 == TCL_NUMBER_LONG)) {


	    w1 = *((const Tcl_WideInt *)ptr1);
	    w2 = *((const Tcl_WideInt *)ptr2);
	    l1 = w1;
	    l2 = w2;

	    switch (*pc) {
	    case INST_ADD:


		wResult = w1 + w2;
		/*
		 * Check for overflow.
		 */

		if (Overflowing(w1, w2, wResult)) {
		    goto overflow;
		}
		goto wideResultOfArithmetic;

	    case INST_SUB:


		wResult = w1 - w2;
		/*
		 * Must check for overflow. The macro tests for overflows in
		 * sums by looking at the sign bits. As we have a subtraction
		 * here, we are adding -w2. As -w2 could in turn overflow, we
		 * test with ~w2 instead: it has the opposite sign bit to w2
		 * so it does the job. Note that the only "bad" case (w2==0)
................................................................................
		    NEXT_INST_F(1, 2, 1);
		}
		Tcl_SetWideIntObj(valuePtr, wResult);
		TRACE(("%s\n", O2S(valuePtr)));
		NEXT_INST_F(1, 1, 0);

	    case INST_DIV:
		if (w2 == 0) {
		    TRACE(("%s %s => DIVIDE BY ZERO\n",
			    O2S(valuePtr), O2S(value2Ptr)));
		    goto divideByZero;
		} else if ((w1 == LLONG_MIN) && (w2 == -1)) {
		    /*
		     * Can't represent (-LLONG_MIN) as a long.
		     */

		    goto overflow;
		}
		wResult = w1 / w2;

		/*
		 * Force Tcl's integer division rules.
		 * TODO: examine for logic simplification
		 */

		if (((wResult < 0) || ((wResult == 0) &&
			((w1 < 0 && w2 > 0) || (w1 > 0 && w2 < 0)))) &&
			((wResult * w2) != w1)) {
		    wResult -= 1;
		}
		goto wideResultOfArithmetic;

	    case INST_MULT:
		if (((sizeof(long) >= 2*sizeof(int))
			&& (w1 <= INT_MAX) && (w1 >= INT_MIN)
			&& (w2 <= INT_MAX) && (w2 >= INT_MIN))
			|| ((sizeof(long) >= 2*sizeof(short))
			&& (w1 <= SHRT_MAX) && (w1 >= SHRT_MIN)
			&& (w2 <= SHRT_MAX) && (w2 >= SHRT_MIN))) {
		    wResult = w1 * w2;
		    goto wideResultOfArithmetic;
		}
	    }

	    /*
	     * Fall through with INST_EXPON, INST_DIV and large multiplies.
	     */
	}
................................................................................
    case INST_LSHIFT:
    case INST_RSHIFT: {
	/*
	 * Reject negative shift argument.
	 */

	switch (type2) {
	case TCL_NUMBER_LONG:
	case TCL_NUMBER_WIDE:
	    invalid = (*((const Tcl_WideInt *)ptr2) < (Tcl_WideInt)0);
	    break;
	case TCL_NUMBER_BIG:
	    Tcl_TakeBignumFromObj(NULL, value2Ptr, &big2);
	    invalid = (mp_cmp_d(&big2, 0) == MP_LT);
	    mp_clear(&big2);
	    break;
................................................................................
	    return GENERAL_ARITHMETIC_ERROR;
	}

	/*
	 * Zero shifted any number of bits is still zero.
	 */

	if ((type1==TCL_NUMBER_LONG || type1==TCL_NUMBER_WIDE) && (*((const Tcl_WideInt *)ptr1) == (Tcl_WideInt)0)) {
	    return constants[0];
	}

	if (opcode == INST_LSHIFT) {
	    /*
	     * Large left shifts create integer overflow.
	     *
................................................................................
		 * not take us to the result of 0 or -1, but since we're using
		 * mp_div_2d to do the work, and it takes only an int
		 * argument, we draw the line there.
		 */

		switch (type1) {
		case TCL_NUMBER_LONG:


		case TCL_NUMBER_WIDE:
		    zero = (*(const Tcl_WideInt *)ptr1 > (Tcl_WideInt)0);
		    break;
		case TCL_NUMBER_BIG:
		    Tcl_TakeBignumFromObj(NULL, valuePtr, &big1);
		    zero = (mp_cmp_d(&big1, 0) == MP_GT);
		    mp_clear(&big1);
................................................................................
			goto overflowBasic;
		    }
		}
		break;

	    case INST_SUB:
		wResult = w1 - w2;
		if ((type1 == TCL_NUMBER_LONG || type1 == TCL_NUMBER_WIDE) || (type2 == TCL_NUMBER_LONG || type2 == TCL_NUMBER_WIDE))
		{
		    /*
		     * Must check for overflow. The macro tests for overflows
		     * in sums by looking at the sign bits. As we have a
		     * subtraction here, we are adding -w2. As -w2 could in
		     * turn overflow, we test with ~w2 instead: it has the
		     * opposite sign bit to w2 so it does the job. Note that