Update UCP handling of \b and \B to match recent changes to \w

Author: PhilipHazel

ChangeLog

@@ -134,6 +134,9 @@ matches the "fullwidth" versions of the hex digits. Just like it is done for
 
 34. GitHub PR305 fixes a potential integer overflow in pcre2_dfa_match().
 
+35. Updated handling of \b and \B in UCP mode to match the changes to \w in 32 
+above because \b and \B are defined in terms of \w.
+
 
 Version 10.42 11-December-2022
 ------------------------------

src/pcre2_dfa_match.c

@@ -1119,11 +1119,10 @@ for (;;)
           if (codevalue == OP_UCP_WORD_BOUNDARY ||
               codevalue == OP_NOT_UCP_WORD_BOUNDARY)
             {
-            if (d == '_') left_word = TRUE; else
-              {
-              uint32_t cat = UCD_CATEGORY(d);
-              left_word = (cat == ucp_L || cat == ucp_N);
-              }
+            int chartype = UCD_CHARTYPE(d);
+            int category = PRIV(ucp_gentype)[chartype];
+            left_word = (category == ucp_L || category == ucp_N ||
+              chartype == ucp_Mn || chartype == ucp_Pc);
             }
           else
 #endif
@@ -1145,11 +1144,10 @@ for (;;)
           if (codevalue == OP_UCP_WORD_BOUNDARY ||
               codevalue == OP_NOT_UCP_WORD_BOUNDARY)
             {
-            if (c == '_') right_word = TRUE; else
-              {
-              uint32_t cat = UCD_CATEGORY(c);
-              right_word = (cat == ucp_L || cat == ucp_N);
-              }
+            int chartype = UCD_CHARTYPE(c);
+            int category = PRIV(ucp_gentype)[chartype];
+            right_word = (category == ucp_L || category == ucp_N ||
+              chartype == ucp_Mn || chartype == ucp_Pc);
             }
           else
 #endif

src/pcre2_match.c

@@ -6180,11 +6180,10 @@ fprintf(stderr, "++ op=%d\n", *Fecode);
 #ifdef SUPPORT_UNICODE
       if (Fop == OP_UCP_WORD_BOUNDARY || Fop == OP_NOT_UCP_WORD_BOUNDARY)
         {
-        if (fc == '_') prev_is_word = TRUE; else
-          {
-          int cat = UCD_CATEGORY(fc);
-          prev_is_word = (cat == ucp_L || cat == ucp_N);
-          }
+        int chartype = UCD_CHARTYPE(fc);
+        int category = PRIV(ucp_gentype)[chartype];
+        prev_is_word = (category == ucp_L || category == ucp_N ||
+          chartype == ucp_Mn || chartype == ucp_Pc);
         }
       else
 #endif  /* SUPPORT_UNICODE */
@@ -6214,11 +6213,10 @@ fprintf(stderr, "++ op=%d\n", *Fecode);
 #ifdef SUPPORT_UNICODE
       if (Fop == OP_UCP_WORD_BOUNDARY || Fop == OP_NOT_UCP_WORD_BOUNDARY)
         {
-        if (fc == '_') cur_is_word = TRUE; else
-          {
-          int cat = UCD_CATEGORY(fc);
-          cur_is_word = (cat == ucp_L || cat == ucp_N);
-          }
+        int chartype = UCD_CHARTYPE(fc);
+        int category = PRIV(ucp_gentype)[chartype];
+        cur_is_word = (category == ucp_L || category == ucp_N ||
+          chartype == ucp_Mn || chartype == ucp_Pc);
         }
       else
 #endif  /* SUPPORT_UNICODE */

testdata/testinput4

@@ -2856,4 +2856,10 @@
 /[[:xdigit:]]+/utf,ucp
     --123ef\x{ff10}\x{ff19}\x{ff21}\x{ff26}\x{ff1a}\=no_jit
 
+/\b.+?\b/utf,ucp
+    --cafe\x{300}_au\x{203f}lait!\=no_jit
+
+/caf\B.+?\B/utf,ucp
+    --cafe\x{300}_au\x{203f}lait!\=no_jit
+
 # End of testinput4

testdata/testinput7

@@ -2478,4 +2478,10 @@
 /[\w]+/utf,ucp
     --cafe\x{300}_au\x{203f}lait!
 
+/\b.+?\b/utf,ucp
+    --cafe\x{300}_au\x{203f}lait!
+
+/caf\B.+?\B/utf,ucp
+    --cafe\x{300}_au\x{203f}lait!
+
 # End of testinput7

testdata/testoutput4

@@ -4567,4 +4567,12 @@ No match
     --123ef\x{ff10}\x{ff19}\x{ff21}\x{ff26}\x{ff1a}\=no_jit
  0: 123ef\x{ff10}\x{ff19}\x{ff21}\x{ff26}
 
+/\b.+?\b/utf,ucp
+    --cafe\x{300}_au\x{203f}lait!\=no_jit
+ 0: cafe\x{300}_au\x{203f}lait
+
+/caf\B.+?\B/utf,ucp
+    --cafe\x{300}_au\x{203f}lait!\=no_jit
+ 0: cafe
+
 # End of testinput4

testdata/testoutput7

@@ -4139,4 +4139,21 @@ No match
     --cafe\x{300}_au\x{203f}lait!
  0: cafe\x{300}_au\x{203f}lait
 
+/\b.+?\b/utf,ucp
+    --cafe\x{300}_au\x{203f}lait!
+ 0: cafe\x{300}_au\x{203f}lait
+
+/caf\B.+?\B/utf,ucp
+    --cafe\x{300}_au\x{203f}lait!
+ 0: cafe\x{300}_au\x{203f}lait!
+ 1: cafe\x{300}_au\x{203f}lai
+ 2: cafe\x{300}_au\x{203f}la
+ 3: cafe\x{300}_au\x{203f}l
+ 4: cafe\x{300}_au\x{203f}
+ 5: cafe\x{300}_au
+ 6: cafe\x{300}_a
+ 7: cafe\x{300}_
+ 8: cafe\x{300}
+ 9: cafe
+
 # End of testinput7