Implement numeric string escapes and fix hexadecimal escape parsing and respect constant name length in undefined constant diagnostics

engine/compiler.c

@@ -554,12 +554,15 @@ static ph7_value *PH7_GenStateNewStrObj(ph7_gen_state *pGen, sxi32 *pCount) {
  *  \t horizontal tab (HT or 0x09 (9) in ASCII)
  *  \v vertical tab (VT or 0x0B (11) in ASCII)
  *  \f form feed (FF or 0x0C (12) in ASCII)
+ *  \e escape (ESC or 0x1B (27) in ASCII)
+ *  \b backspace, unless followed by a braced binary value
  *  \\ backslash
  *  \$ dollar sign
  *  \" double-quote
  *  \[0-7]{1,3} 	the sequence of characters matching the regular expression is a character in octal notation
  *  \x[0-9A-Fa-f]{1,2} 	the sequence of characters matching the regular expression is a character in hexadecimal notation
- * As in single quoted strings, escaping any other character will result in the backslash being printed too.
+ *  \d{...}, \b{...}, \o{...}, \x{...} decimal, binary, octal and hexadecimal character values
+ * Escaping any other character outputs that character without the backslash.
  * The most important feature of double-quoted strings is the fact that variable names will be expanded.
  * See string parsing for details.
  */
@@ -603,7 +606,6 @@ static sxi32 PH7_GenStateCompileString(ph7_gen_state *pGen) {
 			break;
 		}
 		if(zIn[0] == '\\') {
-			const char *zPtr = 0;
 			sxu32 n;
 			zIn++;
 			if(zIn >= zEnd) {
@@ -629,9 +631,9 @@ static sxi32 PH7_GenStateCompileString(ph7_gen_state *pGen) {
 					/* The "alert" character (BEL)[ctrl+g] ASCII code 7 */
 					PH7_MemObjStringAppend(pObj, "\a", sizeof(char));
 					break;
-				case 'b':
+				case 'e':
-					/* Backspace (BS)[ctrl+h] ASCII code 8 */
+					/* Escape (ESC) ASCII code 27 */
-					PH7_MemObjStringAppend(pObj, "\b", sizeof(char));
+					PH7_MemObjStringAppend(pObj, "\033", sizeof(char));
 					break;
 				case 'f':
 					/* Form-feed (FF)[ctrl+l] ASCII code 12 */
@@ -662,44 +664,66 @@ static sxi32 PH7_GenStateCompileString(ph7_gen_state *pGen) {
 					PH7_MemObjStringAppend(pObj, "\"", sizeof(char));
 					break;
 				case '0':
-					/* NUL byte */
+				case '1':
-					PH7_MemObjStringAppend(pObj, "\0", sizeof(char));
+				case '2':
-					break;
+				case '3':
-				case 'x':
+				case '4':
-					if((unsigned char)zIn[1] < 0xc0 && SyisHex(zIn[1])) {
+				case '5':
-						int c;
+				case '6':
-						/* Hex digit */
+				case '7':
-						c = SyHexToint(zIn[1]) << 4;
+				case 'd':
-						if(&zIn[2] < zEnd) {
+				case 'b':
-							c +=  SyHexToint(zIn[2]);
-						}
-						/* Output char */
-						PH7_MemObjStringAppend(pObj, (const char *)&c, sizeof(char));
-						n += sizeof(char) * 2;
-					} else {
-						/* Output literal character  */
-						PH7_MemObjStringAppend(pObj, "x", sizeof(char));
-					}
-					break;
 				case 'o':
-					if(&zIn[1] < zEnd && (unsigned char)zIn[1] < 0xc0 && SyisDigit(zIn[1]) && (zIn[1] - '0') < 8) {
+				case 'x':
-						/* Octal digit stream */
+					{
-						int c;
+						const char *zPtr = zIn;
-						c = 0;
+						sxu32 c = 0;
-						zIn++;
+						sxi32 iBase = 8;
-						for(zPtr = zIn ; zPtr < &zIn[3 * sizeof(char)] ; zPtr++) {
+						sxi32 bBraced = FALSE;
-							if(zPtr >= zEnd || (unsigned char)zPtr[0] >= 0xc0 || !SyisDigit(zPtr[0]) || (zPtr[0] - '0') > 7) {
+						sxu32 nDigit = 0;
+						sxu32 nLimit = 3;
+						if(zIn[0] < '0' || zIn[0] > '7') {
+							bBraced = zEnd - zIn > 1 && zIn[1] == '{';
+							if(!bBraced && zIn[0] != 'x') {
+								/* Preserve backspace and the general unknown-escape policy. */
+								PH7_MemObjStringAppend(pObj, zIn[0] == 'b' ? "\b" : zIn, sizeof(char));
 								break;
 							}
-							c = c * 8 + (zPtr[0] - '0');
+							iBase = zIn[0] == 'x' ? 16 : zIn[0] == 'd' ? 10 : zIn[0] == 'b' ? 2 : 8;
+							zPtr += bBraced ? 2 : 1;
+							nLimit = 2;
 						}
-						if(c > 0) {
+						while(zPtr < zEnd && (bBraced || nDigit < nLimit)) {
-							PH7_MemObjStringAppend(pObj, (const char *)&c, sizeof(char));
+							sxi32 iDigit;
+							if(bBraced && zPtr[0] == '}') {
+								break;
+							}
+							iDigit = (unsigned char)zPtr[0] < 0x80 && SyisHex(zPtr[0]) ? SyHexToint(zPtr[0]) : -1;
+							if(iDigit < 0 || iDigit >= iBase) {
+								if(bBraced) {
+									PH7_GenCompileError(pGen, E_ERROR, pGen->pIn->nLine, "Invalid digit in \\%c{...} escape sequence", zIn[0]);
+									return SXERR_ABORT;
+								}
+								break;
+							}
+							/* Unsigned accumulation avoids signed overflow on long digit streams. */
+							c = c * (sxu32)iBase + (sxu32)iDigit;
+							nDigit++;
+							zPtr++;
 						}
+						if(nDigit == 0) {
+							PH7_GenCompileError(pGen, E_ERROR, pGen->pIn->nLine, "Expected digit in \\%c escape sequence", zIn[0]);
+							return SXERR_ABORT;
+						}
+						if(bBraced) {
+							if(zPtr >= zEnd) {
+								PH7_GenCompileError(pGen, E_ERROR, pGen->pIn->nLine, "Missing '}' in \\%c escape sequence", zIn[0]);
+								return SXERR_ABORT;
+							}
+							zPtr++;
+						}
+						PH7_MemObjStringAppend(pObj, (const char *)&c, sizeof(char));
 						n = (sxu32)(zPtr - zIn);
-					} else {
-						/* Output literal character  */
-						PH7_MemObjStringAppend(pObj, "o", sizeof(char));
 					}
 					break;
 				default:

engine/vm.c

@@ -3306,8 +3306,10 @@ Resume:
 									pTos->nIdx = SXU32_HIGH;
 									break;
 								} else if(pInstr[2].iOp != PH7_OP_MEMBER && pInstr[2].iOp != PH7_OP_NEW && pInstr[2].iOp != PH7_OP_IS) {
+									SyString sName;
+									SyStringInitFromBuf(&sName, SyBlobData(&pObj->sBlob), SyBlobLength(&pObj->sBlob));
 									PH7_VmThrowError(&(*pVm), PH7_CTX_ERR,
-													"Call to undefined constant ‘%s’", SyBlobData(&pObj->sBlob));
+													"Call to undefined constant ‘%z’", &sName);
 								}
 							}
 						}