Mark idempotent translations as 'translated' (#11580)

sphinx/transforms/i18n.py

@@ -353,9 +353,16 @@ def apply(self, **kwargs: Any) -> None:
         if not has_catalog:
             return
 
+        catalogues = [getattr(catalog, '_catalog', None)]
+        while (catalog := catalog._fallback) is not None:  # type: ignore[attr-defined]
+            catalogues.append(getattr(catalog, '_catalog', None))
+        merged: dict[str, str] = {}
+        for catalogue in filter(None, reversed(catalogues)):  # type: dict[str, str]
+            merged |= catalogue
+
         # phase1: replace reference ids with translated names
         for node, msg in extract_messages(self.document):
-            msgstr = catalog.gettext(msg)
+            msgstr = merged.get(msg, '')
 
             # There is no point in having #noqa on literal blocks because
             # they cannot contain references.  Recognizing it would just
@@ -364,10 +371,14 @@ def apply(self, **kwargs: Any) -> None:
             if not isinstance(node, LITERAL_TYPE_NODES):
                 msgstr, _ = parse_noqa(msgstr)
 
-            if not msgstr or msgstr == msg or not msgstr.strip():
+            if msgstr.strip() == '':
                 # as-of-yet untranslated
                 node['translated'] = False
                 continue
+            if msgstr == msg:
+                # identical source and translated messages
+                node['translated'] = True
+                continue
 
             # Avoid "Literal block expected; none found." warnings.
             # If msgstr ends with '::' then it cause warning message at
@@ -416,7 +427,7 @@ def apply(self, **kwargs: Any) -> None:
             if node.setdefault('translated', False):  # to avoid double translation
                 continue  # skip if the node is already translated by phase1
 
-            msgstr = catalog.gettext(msg)
+            msgstr = merged.get(msg, '')
             noqa = False
 
             # See above.
@@ -510,7 +521,7 @@ def apply(self, **kwargs: Any) -> None:
                     msg_parts = split_index_msg(entry_type, value)
                     msgstr_parts = []
                     for part in msg_parts:
-                        msgstr = catalog.gettext(part)
+                        msgstr = merged.get(part, '')
                         if not msgstr:
                             msgstr = part
                         msgstr_parts.append(msgstr)

tests/roots/test-intl/translation_progress.txt

@@ -17,16 +17,20 @@ Desiring this man’s art and that man’s scope,
 
 With what I most enjoy contented least;
 
+.. idempotent translations (2 out of 14 lines):
+
 Yet in these thoughts myself almost despising,
 
 Haply I think on thee, and then my state,
 
-Like to the lark at break of day arising
+.. untranslated (2 out of 14 lines):
 
-.. untranslated (3 out of 14 lines):
+Like to the lark at break of day arising
 
 From sullen earth, sings hymns at heaven’s gate;
 
+.. translation missing (2 out of 14 lines):
+
 For thy sweet love remember’d such wealth brings
 
 That then I scorn to change my state with kings.

tests/roots/test-intl/xx/LC_MESSAGES/translation_progress.po

@@ -38,11 +38,20 @@ msgstr "DESIRING THIS MAN’S ART AND THAT MAN’S SCOPE,"
 msgid "With what I most enjoy contented least;"
 msgstr "WITH WHAT I MOST ENJOY CONTENTED LEAST;"
 
+# idempotent translations (2 out of 14 lines):
+
 msgid "Yet in these thoughts myself almost despising,"
-msgstr "YET IN THESE THOUGHTS MYSELF ALMOST DESPISING,"
+msgstr "Yet in these thoughts myself almost despising,"
 
 msgid "Haply I think on thee, and then my state,"
-msgstr "HAPLY I THINK ON THEE, AND THEN MY STATE,"
+msgstr "Haply I think on thee, and then my state,"
+
+# untranslated (2 out of 14 lines):
 
 msgid "Like to the lark at break of day arising"
-msgstr "LIKE TO THE LARK AT BREAK OF DAY ARISING"
+msgstr ""
+
+msgid "From sullen earth, sings hymns at heaven’s gate;"
+msgstr ""
+
+# translation missing (2 out of 14 lines):

tests/test_intl.py

@@ -612,24 +612,24 @@ def test_gettext_buildr_ignores_only_directive(app):
 
 @sphinx_intl
 def test_node_translated_attribute(app):
-    app.build()
+    app.builder.build_specific([app.srcdir / 'translation_progress.txt'])
 
     doctree = app.env.get_doctree('translation_progress')
 
     translated_nodes = sum(1 for _ in doctree.findall(NodeMatcher(translated=True)))
-    assert translated_nodes == 11 + 1  # 11 lines + title
+    assert translated_nodes == 10 + 1  # 10 lines + title
 
     untranslated_nodes = sum(1 for _ in doctree.findall(NodeMatcher(translated=False)))
-    assert untranslated_nodes == 3 + 1  # 3 lines + substitution reference
+    assert untranslated_nodes == 2 + 2 + 1  # 2 lines + 2 lines + substitution reference
 
 
 @sphinx_intl
 def test_translation_progress_substitution(app):
-    app.build()
+    app.builder.build_specific([app.srcdir / 'translation_progress.txt'])
 
     doctree = app.env.get_doctree('translation_progress')
 
-    assert doctree[0][17][0] == '75.00%'  # 12 out of 16 lines are translated
+    assert doctree[0][19][0] == '68.75%'  # 11 out of 16 lines are translated
 
 
 @pytest.mark.sphinx(testroot='intl', freshenv=True, confoverrides={
@@ -638,11 +638,14 @@ def test_translation_progress_substitution(app):
     'translation_progress_classes': True,
 })
 def test_translation_progress_classes_true(app):
-    app.build()
+    app.builder.build_specific([app.srcdir / 'translation_progress.txt'])
 
     doctree = app.env.get_doctree('translation_progress')
 
+    # title
     assert 'translated' in doctree[0][0]['classes']
+
+    # translated lines
     assert 'translated' in doctree[0][1]['classes']
     assert 'translated' in doctree[0][2]['classes']
     assert 'translated' in doctree[0][3]['classes']
@@ -651,21 +654,31 @@ def test_translation_progress_classes_true(app):
     assert 'translated' in doctree[0][6]['classes']
     assert 'translated' in doctree[0][7]['classes']
     assert 'translated' in doctree[0][8]['classes']
-    assert 'translated' in doctree[0][9]['classes']
+
+    assert doctree[0][9]['classes'] == []  # comment node
+
+    # idempotent
     assert 'translated' in doctree[0][10]['classes']
     assert 'translated' in doctree[0][11]['classes']
 
     assert doctree[0][12]['classes'] == []  # comment node
 
+    # untranslated
     assert 'untranslated' in doctree[0][13]['classes']
     assert 'untranslated' in doctree[0][14]['classes']
-    assert 'untranslated' in doctree[0][15]['classes']
 
-    assert doctree[0][16]['classes'] == []  # comment node
+    assert doctree[0][15]['classes'] == []  # comment node
 
+    # missing
+    assert 'untranslated' in doctree[0][16]['classes']
     assert 'untranslated' in doctree[0][17]['classes']
 
-    assert len(doctree[0]) == 18
+    assert doctree[0][18]['classes'] == []  # comment node
+
+    # substitution reference
+    assert 'untranslated' in doctree[0][19]['classes']
+
+    assert len(doctree[0]) == 20
 
 
 @sphinx_intl