Update POSIX basic regex conversion code.

src/pcre2_convert.c

@@ -74,6 +74,7 @@ POSSIBILITY OF SUCH DAMAGE.
 #define STR_DOT_STAR_LOOKBEHIND STR_DOT STR_ASTERISK STR_LEFT_PARENTHESIS STR_QUESTION_MARK STR_LESS_THAN_SIGN STR_EQUALS_SIGN
 #define STR_LOOKAHEAD_NOT_DOT STR_LEFT_PARENTHESIS STR_QUESTION_MARK STR_EXCLAMATION_MARK STR_BACKSLASH STR_DOT STR_RIGHT_PARENTHESIS
 #define STR_QUERY_s STR_LEFT_PARENTHESIS STR_QUESTION_MARK STR_s STR_RIGHT_PARENTHESIS
+#define STR_STAR_NUL STR_LEFT_PARENTHESIS STR_ASTERISK STR_N STR_U STR_L STR_RIGHT_PARENTHESIS
 
 /* States for range and POSIX processing */
 
@@ -101,12 +102,12 @@ static const char *pcre2_escaped_literals =
   STR_LEFT_SQUARE_BRACKET STR_RIGHT_SQUARE_BRACKET
   STR_LEFT_PARENTHESIS STR_RIGHT_PARENTHESIS;
 
-/* Recognized escapes in POSIX basic patterns. */
+/* Recognized escaped metacharacters in POSIX basic patterns. */
 
-static const char *posix_basic_escapes =
-  STR_QUESTION_MARK STR_PLUS STR_VERTICAL_LINE
+static const char *posix_meta_escapes =
   STR_LEFT_PARENTHESIS STR_RIGHT_PARENTHESIS
-  STR_0 STR_1 STR_2 STR_3 STR_4 STR_5 STR_6 STR_7 STR_8 STR_9;
+  STR_LEFT_CURLY_BRACKET STR_RIGHT_CURLY_BRACKET 
+  STR_1 STR_2 STR_3 STR_4 STR_5 STR_6 STR_7 STR_8 STR_9;
 
 
 
@@ -155,6 +156,7 @@ BOOL nextisliteral = FALSE;
 /* Initialize default for error offset as end of input. */
 
 *bufflenptr = plength;
+PUTCHARS(STR_STAR_NUL);
 
 /* Now scan the input. */
 
@@ -237,7 +239,9 @@ while (plength > 0)
     case CHAR_LEFT_SQUARE_BRACKET:
     PUTCHARS(STR_LEFT_SQUARE_BRACKET);
 
-    /* Handle special cases [[:<:]] and [[:>:]] (which PCRE does support) */
+#ifdef NEVER
+    /* We could handle special cases [[:<:]] and [[:>:]] (which PCRE does
+    support) but they are not part of POSIX 1003.1. */
 
     if (plength >= 6)
       {
@@ -257,8 +261,9 @@ while (plength > 0)
         continue;  /* With next character */
         }
       }
+#endif       
 
-    /* Handle "normal" character classes */
+    /* Handle start of "normal" character classes */
 
     posix_state = POSIX_CLASS_NOT_STARTED;
 
@@ -283,15 +288,17 @@ while (plength > 0)
 
     case CHAR_BACKSLASH:
     if (plength <= 0) return ERROR_END_BACKSLASH;
-    if (!extended && *posix < 127 &&
-          strchr(posix_basic_escapes, *posix) != NULL)
+    if (extended) nextisliteral = TRUE; else
       { 
-      if (isdigit(*posix)) PUTCHARS(STR_BACKSLASH);
-      if (p + 1 > endp) return PCRE2_ERROR_NOMEMORY;
-      lastspecial = *p++ = *posix++;
-      plength--;
+      if (*posix < 127 && strchr(posix_meta_escapes, *posix) != NULL)
+        {
+        if (isdigit(*posix)) PUTCHARS(STR_BACKSLASH);
+        if (p + 1 > endp) return PCRE2_ERROR_NOMEMORY;
+        lastspecial = *p++ = *posix++;
+        plength--;
+        }
+      else nextisliteral = TRUE;  
       }
-    else nextisliteral = TRUE;
     break;
 
     case CHAR_RIGHT_PARENTHESIS:
@@ -323,7 +330,8 @@ while (plength > 0)
     case CHAR_ASTERISK:
     if (lastspecial != CHAR_ASTERISK)
       {
-      if (!extended && posix_state < POSIX_NOT_BRACKET)
+      if (!extended && (posix_state < POSIX_NOT_BRACKET ||
+          lastspecial == CHAR_LEFT_PARENTHESIS))
         goto ESCAPE_LITERAL;
       goto COPY_SPECIAL;
       }

testdata/testinput24

@@ -323,9 +323,9 @@
 
 /a`*b/convert_glob_escape=x
 
-#pattern convert=unset:posix_extended
+# -------- Tests of extended POSIX conversion --------
 
-/a[[:>:]z/
+#pattern convert=unset:posix_extended
 
 /<[[:a[:digit:]b]>/
     <[>
@@ -338,8 +338,6 @@
 
 /a+\1b\\c|d[ab\c]/
 
-/a[[:<:]]b[[:>:]]/
-
 /<[]bc]>/
     <]>
     <b>
@@ -361,6 +359,8 @@
 
 /a***b/
 
+# -------- Tests of basic POSIX conversion --------
+
 #pattern convert=unset:posix_basic
 
 /a*b+c\+[def](ab)\(cd\)/
@@ -371,6 +371,9 @@
     a1b
 
 /how.to how\.to/
+    how\nto how.to
+\= Expect no match     
+    how\x{0}to how.to
 
 /^how to \^how to/
 
@@ -384,12 +387,10 @@
     X*abcY
     X**abcY
     
+/*ab\(*cd\)/ 
+
 /^b\(c^d\)\(^e^f\)/
 
 /a***b/
 
-#pattern convert=unset
-
-/abc/
-
 # End of testinput24

testdata/testoutput24

@@ -508,14 +508,12 @@ No match
 /a`*b/convert_glob_escape=x
 ** Invalid glob escape 'x'
 
+# -------- Tests of extended POSIX conversion --------
+
 #pattern convert=unset:posix_extended
 
-/a[[:>:]z/
-a[[:>:]z
-Failed: error 130 at offset 4: unknown POSIX class name
-
 /<[[:a[:digit:]b]>/
-<[[:a[:digit:]b]>
+(*NUL)<[[:a[:digit:]b]>
     <[>
  0: <[>
     <:>
@@ -531,13 +529,10 @@ Failed: error 130 at offset 4: unknown POSIX class name
 No match
 
 /a+\1b\\c|d[ab\c]/
-a+1b\\c|d[ab\\c]
-
-/a[[:<:]]b[[:>:]]/
-a[[:<:]]b[[:>:]]
+(*NUL)a+1b\\c|d[ab\\c]
 
 /<[]bc]>/
-<[]bc]>
+(*NUL)<[]bc]>
     <]>
  0: <]>
     <b>
@@ -546,7 +541,7 @@ a[[:<:]]b[[:>:]]
  0: <c>
 
 /<[^]bc]>/
-<[^]bc]>
+(*NUL)<[^]bc]>
     <.>
  0: <.>
 \= Expect no match
@@ -556,7 +551,7 @@ No match
 No match
 
 /(a)\1b/
-(a)1b
+(*NUL)(a)1b
     a1b
  0: a1b
  1: a
@@ -565,21 +560,23 @@ No match
 No match
 
 /(ab)c)d]/
-(ab)c\)d\]
+(*NUL)(ab)c\)d\]
     Xabc)d]Y
  0: abc)d]
  1: ab
 
 /a***b/
-a*b
+(*NUL)a*b
+
+# -------- Tests of basic POSIX conversion --------
 
 #pattern convert=unset:posix_basic
 
 /a*b+c\+[def](ab)\(cd\)/
-a*b\+c+[def]\(ab\)(cd)
+(*NUL)a*b\+c\+[def]\(ab\)(cd)
 
 /\(a\)\1b/
-(a)\1b
+(*NUL)(a)\1b
     aab
  0: aab
  1: a
@@ -588,21 +585,26 @@ a*b\+c+[def]\(ab\)(cd)
 No match
 
 /how.to how\.to/
-how.to how\.to
+(*NUL)how.to how\.to
+    how\nto how.to
+ 0: how\x0ato how.to
+\= Expect no match     
+    how\x{0}to how.to
+No match
 
 /^how to \^how to/
-^how to \^how to
+(*NUL)^how to \^how to
 
 /^*abc/
-^\*abc
+(*NUL)^\*abc
 
 /*abc/
-\*abc
+(*NUL)\*abc
     X*abcY
  0: *abc
 
 /**abc/
-\**abc
+(*NUL)\**abc
     XabcY
  0: abc
     X*abcY
@@ -610,14 +612,13 @@ how.to how\.to
     X**abcY
  0: **abc
     
+/*ab\(*cd\)/ 
+(*NUL)\*ab(\*cd)
+
 /^b\(c^d\)\(^e^f\)/
-^b(c\^d)(^e\^f)
+(*NUL)^b(c\^d)(^e\^f)
 
 /a***b/
-a*b
-
-#pattern convert=unset
-
-/abc/
+(*NUL)a*b
 
 # End of testinput24