Merge pull request #256 from facelessuser/enhance/lang

Changes to `:lang()` as defined in recent CSS spec update

Author: facelessuser

docs/src/dictionary/en-custom.txt

@@ -73,6 +73,7 @@ tuples
 un
 unmatchable
 unpickle
+untagged
 unvisited
 whitespace
 wildcard

docs/src/markdown/about/changelog.md

@@ -1,6 +1,11 @@
 # Changelog
 
-## 2.4.0
+## 2.5
+
+- **NEW**: Update to support changes related to `:lang()` in the official CSS spec. `:lang("")` should match unspecified
+  languages, e.g. `lang=""`, but not `lang=und`.
+
+## 2.4
 
 - **NEW**: `:nth-child()` and `:nth-last-child()` will forgive irregular comma usage.
 - **NEW**: Formally drop Python 3.6.

soupsieve/__meta__.py

@@ -193,5 +193,5 @@ def parse_version(ver: str) -> Version:
     return Version(major, minor, micro, release, pre, post, dev)
 
 
-__version_info__ = Version(2, 4, 0, "final", post=1)
+__version_info__ = Version(2, 5, 0, "final", post=1)
 __version__ = __version_info__._get_canonical()

soupsieve/css_match.py

@@ -601,13 +601,18 @@ def extended_language_filter(self, lang_range: str, lang_tag: str) -> bool:
         ranges = lang_range.split('-')
         subtags = lang_tag.lower().split('-')
         length = len(ranges)
+        slength = len(subtags)
         rindex = 0
         sindex = 0
         r = ranges[rindex]
         s = subtags[sindex]
 
+        # Empty specified language should match unspecified language attributes
+        if length == 1 and slength == 1 and not r and r == s:
+            return True
+
         # Primary tag needs to match
-        if r != '*' and r != s:
+        if (r != '*' and r != s) or (r == '*' and slength == 1 and not s):
             match = False
 
         rindex += 1
@@ -1184,7 +1189,7 @@ def match_lang(self, el: bs4.Tag, langs: tuple[ct.SelectorLang, ...]) -> bool:
                 break
 
         # Use cached meta language.
-        if not found_lang and self.cached_meta_lang:
+        if found_lang is None and self.cached_meta_lang:
             for cache in self.cached_meta_lang:
                 if root is cache[0]:
                     found_lang = cache[1]
@@ -1218,13 +1223,13 @@ def match_lang(self, el: bs4.Tag, langs: tuple[ct.SelectorLang, ...]) -> bool:
                                 found_lang = content
                                 self.cached_meta_lang.append((cast(str, root), cast(str, found_lang)))
                                 break
-                    if found_lang:
+                    if found_lang is not None:
                         break
-                if not found_lang:
+                if found_lang is None:
                     self.cached_meta_lang.append((cast(str, root), ''))
 
         # If we determined a language, compare.
-        if found_lang:
+        if found_lang is not None:
             for patterns in langs:
                 match = False
                 for pattern in patterns:

tests/test_level4/test_lang.py

@@ -155,6 +155,44 @@ def test_avoid_implicit_language(self):
             flags=util.HTML
         )
 
+    def test_language_und(self):
+        """Test that undefined language can be matched by `*`."""
+
+        markup = """
+        <div id="1" lang=""></div>
+        <div id="2" lang="und"></div>
+        <div id="3" lang=>
+            <div id="4"></div>
+        </div>
+        <div id="5"></div>
+        """
+
+        self.assert_selector(
+            markup,
+            "div:lang('*')",
+            ['2'],
+            flags=util.HTML
+        )
+
+    def test_language_empty_string(self):
+        """Test that an empty string language will only match untagged languages `lang=""`."""
+
+        markup = """
+        <div id="1" lang=""></div>
+        <div id="2" lang="und"></div>
+        <div id="3" lang=>
+            <div id="4"></div>
+        </div>
+        <div id="5"></div>
+        """
+
+        self.assert_selector(
+            markup,
+            "div:lang('')",
+            ['1', '3', '4'],
+            flags=util.HTML
+        )
+
     def test_language_list(self):
         """Test language list."""