Add TODO comments. Further small review comments

src/pcre2_auto_possess.c

@@ -482,6 +482,8 @@ switch(c)
   case OP_CLASS:
   case OP_XCLASS:
   case OP_ECLASS:
+  /* TODO: [EC] https://github.com/PCRE2Project/pcre2/issues/537
+  Add back the "ifdef SUPPORT_WIDE_CHARS" once we stop emitting ECLASS for this case. */
   if (c == OP_XCLASS || c == OP_ECLASS)
     end = code + GET(code, 0) - 1;
   else
@@ -1117,6 +1119,8 @@ for(;;)
       break;
 #endif
 
+      /* TODO: [EC] https://github.com/PCRE2Project/pcre2/issues/537
+      Enclose in "ifdef SUPPORT_WIDE_CHARS" once we stop emitting ECLASS for this case. */
       case OP_ECLASS:
       if (PRIV(eclass)(chr,
           (list_ptr == list ? code : base_end) - list_ptr[2] + LINK_SIZE,
@@ -1226,6 +1230,8 @@ for (;;)
     }
   else if (c == OP_CLASS || c == OP_NCLASS || c == OP_XCLASS || c == OP_ECLASS)
     {
+    /* TODO: [EC] https://github.com/PCRE2Project/pcre2/issues/537
+    Add back the "ifdef SUPPORT_WIDE_CHARS" once we stop emitting ECLASS for this case. */
     if (c == OP_XCLASS || c == OP_ECLASS)
       repeat_opcode = code + GET(code, 1);
     else
@@ -1301,6 +1307,8 @@ for (;;)
     code += GET(code, 1 + 2*LINK_SIZE);
     break;
 
+    /* TODO: [EC] https://github.com/PCRE2Project/pcre2/issues/537
+    Add back the "ifdef SUPPORT_WIDE_CHARS" once we stop emitting ECLASS for this case. */
     case OP_ECLASS:
     case OP_XCLASS:
     code += GET(code, 1);

src/pcre2_compile.c

@@ -3572,6 +3572,13 @@ while (ptr < ptrend)
     case CHAR_LEFT_SQUARE_BRACKET:
     okquantifier = TRUE;
 
+    /* TODO: [EC] https://github.com/PCRE2Project/pcre2/issues/536
+    We shall support Perl's (?[...]) syntax. We need a variable class_perlext = true
+    and a goto jumping here if we see "(?[...". We need to check for closing "])" and
+    also implement the completely idiosyncratic nesting and operator rules in this
+    mode. We can hopefully emit exactly the same META codes as for the UTS#18
+    syntax, so that only parser changes are required for the Perl syntax. */
+
     /* In another (POSIX) regex library, the ugly syntax [[:<:]] and [[:>:]] is
     used for "start of word" and "end of word". As these are otherwise illegal
     sequences, we don't break anything by recognizing them. They are replaced
@@ -3721,7 +3728,8 @@ while (ptr < ptrend)
               ptr[1] == CHAR_MINUS) &&
             (ptr + 1 == ptrend || ptr[1] != CHAR_RIGHT_SQUARE_BRACKET))
           {
-          // TODO: We need a better lookahead here to match Perl; should skip \Q\E, and whitespace in /xx mode.
+          /* TODO: [EC] https://github.com/PCRE2Project/pcre2/issues/538
+          We need a better lookahead here to match Perl; should skip \Q\E, and whitespace in /xx mode. */
           errorcode = ERR50;
           goto FAILED;
           }
@@ -4094,7 +4102,8 @@ while (ptr < ptrend)
               ptr[1] == CHAR_MINUS) &&
             (ptr + 1 == ptrend || ptr[1] != CHAR_RIGHT_SQUARE_BRACKET))
           {
-          // TODO: We need a better lookahead here to match Perl; should skip \Q\E, and whitespace in /xx mode.
+          /* TODO: [EC] https://github.com/PCRE2Project/pcre2/issues/538
+          We need a better lookahead here to match Perl; should skip \Q\E, and whitespace in /xx mode. */
           errorcode = ERR50;
           goto FAILED;
           }
@@ -6781,6 +6790,8 @@ for (;; pptr++)
 #ifdef SUPPORT_WIDE_CHARS
       case OP_XCLASS:
 #endif
+      /* TODO: [EC] https://github.com/PCRE2Project/pcre2/issues/537
+      Enclose in the "ifdef SUPPORT_WIDE_CHARS" once we stop emitting ECLASS for this case. */
       case OP_ECLASS:
       case OP_CLASS:
       case OP_NCLASS:
@@ -6815,7 +6826,8 @@ for (;; pptr++)
       (?!), disallow a quantifier at parse time. We ought to be able to ignore this. */
 
       case OP_FAIL:
-      /* TODO: Should this be removed, now that '[]' has been changed so it doesn't produce OP_FAIL? */
+      /* TODO: [EC] https://github.com/PCRE2Project/pcre2/issues/541
+      Should this be removed, now that '[]' has been changed so it doesn't produce OP_FAIL? */
       PCRE2_UNREACHABLE();
       goto END_REPEAT;
 
@@ -7430,6 +7442,8 @@ for (;; pptr++)
         tempcode += 1 + 32/sizeof(PCRE2_UCHAR);
         break;
 
+        /* TODO: [EC] https://github.com/PCRE2Project/pcre2/issues/537
+        Add back the "ifdef SUPPORT_WIDE_CHARS" once we stop emitting ECLASS for this case. */
         case OP_XCLASS:
         case OP_ECLASS:
         tempcode += GET(tempcode, 1);

src/pcre2_compile.h

@@ -231,8 +231,9 @@ extern const int PRIV(posix_class_maps)[];
 
 /* Merge intersecting ranges of classes. */
 
-class_ranges *PRIV(optimize_class)(uint32_t *start_ptr, uint32_t *end_ptr,
-  uint32_t options, uint32_t xoptions, compile_block* cb);
+class_ranges *PRIV(optimize_class)(uint32_t *start_ptr,
+  const uint32_t *end_ptr, uint32_t options, uint32_t xoptions,
+  compile_block *cb);
 
 /* Set bits in classbits according to the property type */
 
@@ -258,7 +259,7 @@ BOOL PRIV(compile_class_nested)(uint32_t options, uint32_t xoptions,
 OP_CLASS, OP_NCLASS, OP_XCLASS, or OP_ALLANY into pcode. */
 
 BOOL PRIV(compile_class_not_nested)(uint32_t options, uint32_t xoptions,
-  uint32_t *start_ptr, uint32_t *end_ptr, PCRE2_UCHAR **pcode,
+  uint32_t *start_ptr, const uint32_t *end_ptr, PCRE2_UCHAR **pcode,
   BOOL negate_class, int *errorcodeptr, compile_block *cb,
   PCRE2_SIZE *lengthptr);
 

src/pcre2_compile_class.c

@@ -319,7 +319,8 @@ append_non_ascii_range(uint32_t options, uint32_t *buffer)
 }
 
 static size_t
-parse_class(uint32_t *ptr, uint32_t *end_ptr, uint32_t options, uint32_t *buffer)
+parse_class(uint32_t *ptr, const uint32_t *end_ptr, uint32_t options,
+  uint32_t *buffer)
 {
 size_t total_size = 0;
 size_t size;
@@ -457,8 +458,9 @@ static const uint32_t char_list_starts[] = {
   XCL_CHAR_LIST_LOW_16_START,
 };
 
-class_ranges *PRIV(optimize_class)(uint32_t *start_ptr, uint32_t *end_ptr,
-  uint32_t options, uint32_t xoptions, compile_block* cb)
+class_ranges *PRIV(optimize_class)(uint32_t *start_ptr,
+  const uint32_t *end_ptr, uint32_t options, uint32_t xoptions,
+  compile_block *cb)
 {
 class_ranges* cranges;
 uint32_t *ptr;
@@ -933,11 +935,9 @@ while (p[0] < 256)
   p += n + 1;
   }
 }
-#endif
 
 
 
-#if PCRE2_CODE_UNIT_WIDTH == 8
 /*************************************************
 *    Add characters not in a list to a class     *
 *************************************************/
@@ -966,7 +966,7 @@ while (p[0] < 256)
   p++;
   }
 }
-#endif
+#endif /* PCRE2_CODE_UNIT_WIDTH == 8 */
 
 
 
@@ -1042,8 +1042,9 @@ become a single OP_CLASS (or OP_NCLASS, OP_XCLASS, or OP_ALLANY). */
 
 BOOL
 PRIV(compile_class_not_nested)(uint32_t options, uint32_t xoptions,
-  uint32_t *start_ptr, uint32_t *end_ptr, PCRE2_UCHAR **pcode, BOOL negate_class,
-  int *errorcodeptr, compile_block *cb, PCRE2_SIZE *lengthptr)
+  uint32_t *start_ptr, const uint32_t *end_ptr, PCRE2_UCHAR **pcode,
+  BOOL negate_class, int *errorcodeptr, compile_block *cb,
+  PCRE2_SIZE *lengthptr)
 {
 uint32_t *pptr = start_ptr;
 PCRE2_UCHAR *code = *pcode;
@@ -1942,6 +1943,12 @@ PRIV(compile_class_nested)(uint32_t options, uint32_t xoptions,
   uint32_t **pptr, PCRE2_UCHAR **pcode, int *errorcodeptr,
   compile_block *cb, PCRE2_SIZE *lengthptr)
 {
+/* TODO: [EC] https://github.com/PCRE2Project/pcre2/issues/537
+We shall convert this recursive descent into a stack-based precedence parser.
+We shall optimise it, so that OP_CLASS/NCLASS are constant-folded.
+We shall potentially fold all the bitsets, so that there's only one bitset
+held by the OP_ECLASS. */
+
 uint32_t *ptr = *pptr;
 PCRE2_UCHAR *code = *pcode;
 BOOL negated;

src/pcre2_jit_compile.c

@@ -1102,6 +1102,8 @@ switch(*cc)
   case OP_CALLOUT_STR:
   return cc + GET(cc, 1 + 2*LINK_SIZE);
 
+  /* TODO: [EC] https://github.com/PCRE2Project/pcre2/issues/537
+  Add back the "if defined SUPPORT_UNICODE || PCRE2_CODE_UNIT_WIDTH != 8" once we stop emitting ECLASS for this case. */
   case OP_ECLASS:
   case OP_XCLASS:
   return cc + GET(cc, 1);

src/pcre2_match.c

@@ -2330,6 +2330,8 @@ fprintf(stderr, "++ %2ld op=%3d %s\n", Fecode - mb->start_code, *Fecode,
 #define Lmin         F->temp_32[0]
 #define Lmax         F->temp_32[1]
 
+    /* TODO: [EC] https://github.com/PCRE2Project/pcre2/issues/537
+    Enclose in "ifdef SUPPORT_WIDE_CHARS" once we stop emitting ECLASS for this case. */
     case OP_ECLASS:
       {
       Leclass_data = Fecode + 1 + LINK_SIZE;  /* Save for matching */

src/pcre2_study.c

@@ -415,6 +415,9 @@ for (;;)
 
     /* Check a class for variable quantification */
 
+    /* TODO: [EC] https://github.com/PCRE2Project/pcre2/issues/537
+    Add back the "ifdef SUPPORT_WIDE_CHARS" once we stop emitting ECLASS for this case. */
+
     case OP_CLASS:
     case OP_NCLASS:
     case OP_XCLASS:
@@ -1714,6 +1717,9 @@ do
 
       /* Set-based ECLASS: treat it the same as a "complex" XCLASS; give up. */
 
+      /* TODO: [EC] https://github.com/PCRE2Project/pcre2/issues/537
+      Enclose in "ifdef SUPPORT_WIDE_CHARS" once we stop emitting ECLASS for this case. */
+
       case OP_ECLASS:
       return SSB_FAIL;
 

src/pcre2_xclass.c

@@ -530,7 +530,7 @@ while (ptr < data_end)
   }
 
 /* The final bit left on the stack now holds the match result. */
-return (stack & 1u);
+return (stack & 1u) != 0;
 }
 
 /* End of pcre2_xclass.c */