Start implementing posix classes.

topazproject · Jul 7, 2013 · 61ba7f8 · 61ba7f8
1 parent 4108429
commit 61ba7f8
Show file tree

Hide file tree

Showing 2 changed files with 28 additions and 14 deletions.
diff --git a/tests/objects/test_regexpobject.py b/tests/objects/test_regexpobject.py
@@ -51,6 +51,7 @@ def test_compile_regexps(self, space):
         /(.*|.+)/
         /(?<=b)/
         /\\A|\\z/
+        /[[:alnum:]]/
         """)
 
     def test_regexp_syntax_errors(self, space):
@@ -204,6 +205,10 @@ def test_quantify_set(self, space):
         w_res = space.execute("return /([0-9]){3,5}?/ =~ 'ab12345'")
         assert space.int_w(w_res) == 2
 
+    def test_posix_class(self, space):
+        w_res = space.execute("return /[[:digit:]]/ =~ 'abc2'")
+        assert space.int_w(w_res) == 3
+
     def test_quantify(self, space):
         w_res = space.execute("return /a{2,4}/.match('aaaaaa').to_a")
         assert self.unwrap(space, w_res) == ["aaaa"]

diff --git a/topaz/utils/regexp.py b/topaz/utils/regexp.py
@@ -836,6 +836,7 @@ def compile(self, ctx):
 }
 PROPERTIES = {
     "digit": CATEGORY_DIGIT,
+    "alnum": CATEGORY_WORD,
 }
 
 
@@ -1247,31 +1248,39 @@ def _parse_property(source, info, positive, in_set):
     here = source.pos
     if source.match("{"):
         negate = source.match("^")
-        b = StringBuilder(5)
-        found = False
-        while True:
-            ch = source.get()
-            if ch == "}":
-                found = True
-                break
-            elif not ch:
-                break
-            else:
-                b.append(ch)
-        if found:
-            name = b.build()
+        prop_name, name = _parse_property_name(source)
+        if source.match("}"):
             if name in PROPERTIES:
                 return Property(PROPERTIES[name], positive != negate)
     source.pos = here
     return make_character(info, ord("p" if positive else "P"), in_set)
 
 
+def _parse_property_name(source):
+    b = StringBuilder(5)
+    here = source.pos
+    while True:
+        pos2 = source.pos
+        ch = source.get()
+        if ch.isalnum():
+            b.append(ch)
+        else:
+            source.pos = pos2
+            break
+    name = b.build()
+    return name, name
+
+
 def _parse_numeric_escape(source, info, ch, in_set):
     raise NotImplementedError("_parse_numeric_escape")
 
 
 def _parse_posix_class(source, info):
-    raise NotImplementedError("_parse_posix_class")
+    negate = source.match("^")
+    prop_name, name = _parse_property_name(source)
+    if not source.match(":]"):
+        raise ParseError
+    return Property(PROPERTIES[name], negate)
 
 
 def _compile_no_cache(pattern, flags):