Add JSX support for message extraction

babel/messages/extract.py

@@ -461,6 +461,8 @@ def extract_javascript(fileobj, keywords, comment_tags, options):
     :param comment_tags: a list of translator tags to search for and include
                          in the results
     :param options: a dictionary of additional options (optional)
+                    Supported options are:
+                    * `jsx` -- set to false to disable JSX/E4X support.
     """
     from babel.messages.jslexer import tokenize, unquote_string
     funcname = message_lineno = None
@@ -472,7 +474,7 @@ def extract_javascript(fileobj, keywords, comment_tags, options):
     last_token = None
     call_stack = -1
 
-    for token in tokenize(fileobj.read().decode(encoding)):
+    for token in tokenize(fileobj.read().decode(encoding), jsx=options.get("jsx", True)):
         if token.type == 'operator' and token.value == '(':
             if funcname:
                 message_lineno = token.lineno

babel/messages/jslexer.py

@@ -36,6 +36,7 @@
         ([eE][-+]?\d+)? |
         (0x[a-fA-F0-9]+)
     )''')),
+    ('jsx_tag', re.compile(r'<(?:/?)\w+.+?>', re.I)),
     ('operator', re.compile(r'(%s)' % '|'.join(map(re.escape, operators)))),
     ('string', re.compile(r'''(?xs)(
         '(?:[^'\\]*(?:\\.[^'\\]*)*)'  |
@@ -127,8 +128,11 @@ def unquote_string(string):
     return u''.join(result)
 
 
-def tokenize(source):
-    """Tokenize a JavaScript source.  Returns a generator of tokens.
+def tokenize(source, jsx=True):
+    """
+    Tokenize JavaScript/JSX source.  Returns a generator of tokens.
+
+    :param jsx: Enable (limited) JSX parsing.
     """
     may_divide = False
     pos = 0
@@ -138,6 +142,8 @@ def tokenize(source):
     while pos < end:
         # handle regular rules first
         for token_type, rule in rules:
+            if not jsx and token_type and 'jsx' in token_type:
+                continue
             match = rule.match(source, pos)
             if match is not None:
                 break

tests/messages/test_extract.py

@@ -388,97 +388,6 @@ def test_extract_strip_comment_tags(self):
                           u'a prefix too'], messages[1][2])
 
 
-class ExtractJavaScriptTestCase(unittest.TestCase):
-
-    def test_simple_extract(self):
-        buf = BytesIO(b"""\
-msg1 = _('simple')
-msg2 = gettext('simple')
-msg3 = ngettext('s', 'p', 42)
-        """)
-        messages = \
-            list(extract.extract('javascript', buf, extract.DEFAULT_KEYWORDS,
-                                 [], {}))
-
-        self.assertEqual([(1, 'simple', [], None),
-                          (2, 'simple', [], None),
-                          (3, ('s', 'p'), [], None)], messages)
-
-    def test_various_calls(self):
-        buf = BytesIO(b"""\
-msg1 = _(i18n_arg.replace(/"/, '"'))
-msg2 = ungettext(i18n_arg.replace(/"/, '"'), multi_arg.replace(/"/, '"'), 2)
-msg3 = ungettext("Babel", multi_arg.replace(/"/, '"'), 2)
-msg4 = ungettext(i18n_arg.replace(/"/, '"'), "Babels", 2)
-msg5 = ungettext('bunny', 'bunnies', parseInt(Math.random() * 2 + 1))
-msg6 = ungettext(arg0, 'bunnies', rparseInt(Math.random() * 2 + 1))
-msg7 = _(hello.there)
-msg8 = gettext('Rabbit')
-msg9 = dgettext('wiki', model.addPage())
-msg10 = dngettext(domain, 'Page', 'Pages', 3)
-""")
-        messages = \
-            list(extract.extract('javascript', buf, extract.DEFAULT_KEYWORDS, [],
-                                 {}))
-        self.assertEqual([(5, (u'bunny', u'bunnies'), [], None),
-                          (8, u'Rabbit', [], None),
-                          (10, (u'Page', u'Pages'), [], None)], messages)
-
-    def test_message_with_line_comment(self):
-        buf = BytesIO(u"""\
-// NOTE: hello
-msg = _('Bonjour à tous')
-""".encode('utf-8'))
-        messages = list(extract.extract_javascript(buf, ('_',), ['NOTE:'], {}))
-        self.assertEqual(u'Bonjour à tous', messages[0][2])
-        self.assertEqual([u'NOTE: hello'], messages[0][3])
-
-    def test_message_with_multiline_comment(self):
-        buf = BytesIO(u"""\
-/* NOTE: hello
-   and bonjour
-     and servus */
-msg = _('Bonjour à tous')
-""".encode('utf-8'))
-        messages = list(extract.extract_javascript(buf, ('_',), ['NOTE:'], {}))
-        self.assertEqual(u'Bonjour à tous', messages[0][2])
-        self.assertEqual([u'NOTE: hello', 'and bonjour', '  and servus'], messages[0][3])
-
-    def test_ignore_function_definitions(self):
-        buf = BytesIO(b"""\
-function gettext(value) {
-    return translations[language][value] || value;
-}""")
-
-        messages = list(extract.extract_javascript(buf, ('gettext',), [], {}))
-        self.assertEqual(messages, [])
-
-    def test_misplaced_comments(self):
-        buf = BytesIO(b"""\
-/* NOTE: this won't show up */
-foo()
-
-/* NOTE: this will */
-msg = _('Something')
-
-// NOTE: this will show up
-// too.
-msg = _('Something else')
-
-// NOTE: but this won't
-bar()
-
-_('no comment here')
-""")
-        messages = list(extract.extract_javascript(buf, ('_',), ['NOTE:'], {}))
-        self.assertEqual(u'Something', messages[0][2])
-        self.assertEqual([u'NOTE: this will'], messages[0][3])
-        self.assertEqual(u'Something else', messages[1][2])
-        self.assertEqual([u'NOTE: this will show up', 'too.'], messages[1][3])
-        self.assertEqual(u'no comment here', messages[2][2])
-        self.assertEqual([], messages[2][3])
-
-
 class ExtractTestCase(unittest.TestCase):
 
     def test_invalid_filter(self):

tests/messages/test_js_extract.py

@@ -0,0 +1,124 @@
+# -- encoding: UTF-8 --
+import pytest
+from babel._compat import BytesIO
+from babel.messages import extract
+
+
+def test_simple_extract():
+    buf = BytesIO(b"""\
+msg1 = _('simple')
+msg2 = gettext('simple')
+msg3 = ngettext('s', 'p', 42)
+    """)
+    messages = \
+        list(extract.extract('javascript', buf, extract.DEFAULT_KEYWORDS,
+                             [], {}))
+
+    assert messages == [(1, 'simple', [], None),
+                        (2, 'simple', [], None),
+                        (3, ('s', 'p'), [], None)]
+
+
+def test_various_calls():
+    buf = BytesIO(b"""\
+msg1 = _(i18n_arg.replace(/"/, '"'))
+msg2 = ungettext(i18n_arg.replace(/"/, '"'), multi_arg.replace(/"/, '"'), 2)
+msg3 = ungettext("Babel", multi_arg.replace(/"/, '"'), 2)
+msg4 = ungettext(i18n_arg.replace(/"/, '"'), "Babels", 2)
+msg5 = ungettext('bunny', 'bunnies', parseInt(Math.random() * 2 + 1))
+msg6 = ungettext(arg0, 'bunnies', rparseInt(Math.random() * 2 + 1))
+msg7 = _(hello.there)
+msg8 = gettext('Rabbit')
+msg9 = dgettext('wiki', model.addPage())
+msg10 = dngettext(domain, 'Page', 'Pages', 3)
+""")
+    messages = \
+        list(extract.extract('javascript', buf, extract.DEFAULT_KEYWORDS, [],
+                             {}))
+    assert messages == [
+        (5, (u'bunny', u'bunnies'), [], None),
+        (8, u'Rabbit', [], None),
+        (10, (u'Page', u'Pages'), [], None)
+    ]
+
+
+def test_message_with_line_comment():
+    buf = BytesIO(u"""\
+// NOTE: hello
+msg = _('Bonjour à tous')
+""".encode('utf-8'))
+    messages = list(extract.extract_javascript(buf, ('_',), ['NOTE:'], {}))
+    assert messages[0][2] == u'Bonjour à tous'
+    assert messages[0][3] == [u'NOTE: hello']
+
+
+def test_message_with_multiline_comment():
+    buf = BytesIO(u"""\
+/* NOTE: hello
+and bonjour
+  and servus */
+msg = _('Bonjour à tous')
+""".encode('utf-8'))
+    messages = list(extract.extract_javascript(buf, ('_',), ['NOTE:'], {}))
+    assert messages[0][2] == u'Bonjour à tous'
+    assert messages[0][3] == [u'NOTE: hello', 'and bonjour', '  and servus']
+
+
+def test_ignore_function_definitions():
+    buf = BytesIO(b"""\
+function gettext(value) {
+return translations[language][value] || value;
+}""")
+
+    messages = list(extract.extract_javascript(buf, ('gettext',), [], {}))
+    assert not messages
+
+
+def test_misplaced_comments():
+    buf = BytesIO(b"""\
+/* NOTE: this won't show up */
+foo()
+
+/* NOTE: this will */
+msg = _('Something')
+
+// NOTE: this will show up
+// too.
+msg = _('Something else')
+
+// NOTE: but this won't
+bar()
+
+_('no comment here')
+""")
+    messages = list(extract.extract_javascript(buf, ('_',), ['NOTE:'], {}))
+    assert messages[0][2] == u'Something'
+    assert messages[0][3] == [u'NOTE: this will']
+    assert messages[1][2] == u'Something else'
+    assert messages[1][3] == [u'NOTE: this will show up', 'too.']
+    assert messages[2][2] == u'no comment here'
+    assert messages[2][3] == []
+
+
+JSX_SOURCE = b"""
+class Foo {
+    render() {
+        const value = gettext("hello");
+        return (
+            <option value="val1">{ i18n._('String1') }</option>
+            <option value="val2">{ i18n._('String 2') }</option>
+            <option value="val3">{ i18n._('String 3') }</option>
+        );
+    }
+"""
+EXPECTED_JSX_MESSAGES = ["hello", "String1", "String 2", "String 3"]
+
+
+@pytest.mark.parametrize("jsx_enabled", (False, True))
+def test_jsx_extraction(jsx_enabled):
+    buf = BytesIO(JSX_SOURCE)
+    messages = [m[2] for m in extract.extract_javascript(buf, ('_', 'gettext'), [], {"jsx": jsx_enabled})]
+    if jsx_enabled:
+        assert messages == EXPECTED_JSX_MESSAGES
+    else:
+        assert messages != EXPECTED_JSX_MESSAGES