Use a modified version of pybabel's code in normalize

Author: StanFromIreland

Lib/test/test_tools/i18n_data/messages.pot

@@ -5,7 +5,7 @@
 msgid ""
 msgstr ""
 "Project-Id-Version: PACKAGE VERSION\n"
-"POT-Creation-Date: 2000-01-01 00:00+0000\n"
+"POT-Creation-Date: 2025-03-01 09:36+0000\n"
 "PO-Revision-Date: YEAR-MO-DA HO:MI+ZONE\n"
 "Last-Translator: FULL NAME <EMAIL@ADDRESS>\n"
 "Language-Team: LANGUAGE <[email protected]>\n"
@@ -33,65 +33,71 @@ msgid ""
 "    multiline!\n"
 msgstr ""
 
-#: messages.py:46 messages.py:89 messages.py:90 messages.py:93 messages.py:94
-#: messages.py:99 messages.py:100 messages.py:101
+#: messages.py:32
+msgid ""
+"this is a very very very very very very very very very very very very very"
+"long string!"
+msgstr ""
+
+#: messages.py:49 messages.py:92 messages.py:93 messages.py:96 messages.py:97
+#: messages.py:102 messages.py:103 messages.py:104
 msgid "foo"
 msgid_plural "foos"
 msgstr[0] ""
 msgstr[1] ""
 
-#: messages.py:47
+#: messages.py:50
 msgid "something"
 msgstr ""
 
-#: messages.py:50
+#: messages.py:53
 msgid "Hello, {}!"
 msgstr ""
 
-#: messages.py:54
+#: messages.py:57
 msgid "1"
 msgstr ""
 
-#: messages.py:54
+#: messages.py:57
 msgid "2"
 msgstr ""
 
-#: messages.py:55 messages.py:56
+#: messages.py:58 messages.py:59
 msgid "A"
 msgstr ""
 
-#: messages.py:55 messages.py:56
+#: messages.py:58 messages.py:59
 msgid "B"
 msgstr ""
 
-#: messages.py:57
+#: messages.py:60
 msgid "set"
 msgstr ""
 
-#: messages.py:62 messages.py:63
+#: messages.py:65 messages.py:66
 msgid "nested string"
 msgstr ""
 
-#: messages.py:68
+#: messages.py:71
 msgid "baz"
 msgstr ""
 
-#: messages.py:71 messages.py:75
+#: messages.py:74 messages.py:78
 msgid "default value"
 msgstr ""
 
-#: messages.py:91 messages.py:92 messages.py:95 messages.py:96
+#: messages.py:94 messages.py:95 messages.py:98 messages.py:99
 msgctxt "context"
 msgid "foo"
 msgid_plural "foos"
 msgstr[0] ""
 msgstr[1] ""
 
-#: messages.py:102
+#: messages.py:105
 msgid "domain foo"
 msgstr ""
 
-#: messages.py:118 messages.py:119
+#: messages.py:121 messages.py:122
 msgid "world"
 msgid_plural "worlds"
 msgstr[0] ""

Lib/test/test_tools/i18n_data/messages.py

@@ -28,6 +28,9 @@
     multiline!
 """)
 
+# very long string that should be wrapped
+_("this is a very very very very very very very very very very very very very long string!")
+
 # Invalid arguments
 _()
 _(None)

Lib/test/test_tools/test_i18n.py

@@ -161,11 +161,6 @@ def test_POT_Creation_Date(self):
             # This will raise if the date format does not exactly match.
             datetime.strptime(creationDate, '%Y-%m-%d %H:%M%z')
 
-    def test_wrap_to_width(self):
-        msgid = self.extract_from_str(
-            '''_("thisisaveryveryveryveryveryveryveryveryveryveryveryveryveryveryveryveryveryverlongstring")''')
-        self.assertIn('\nlongstring', msgid[1])
-
     def test_funcdocstring(self):
         for doc in ('"""doc"""', "r'''doc'''", "R'doc'", 'u"doc"'):
             with self.subTest(doc):

Tools/i18n/pygettext.py

@@ -213,21 +213,39 @@ def escape_nonascii(s, encoding):
     return ''.join(escapes[b] for b in s.encode(encoding))
 
 
-def normalize(s, encoding):
+def normalize(s, encoding, options):
     # This converts the various Python string types into a format that is
-    # appropriate for .po files, namely much closer to C style.
-    lines = s.split('\n')
-    if len(lines) == 1:
-        s = '"' + escape(s, encoding) + '"'
-    else:
-        if not lines[-1]:
-            del lines[-1]
-            lines[-1] = lines[-1] + '\n'
-        for i in range(len(lines)):
-            lines[i] = escape(lines[i], encoding)
-        lineterm = '\\n"\n"'
-        s = '""\n"' + lineterm.join(lines) + '"'
-    return s
+    # appropriate for .po files, namely much closer to C style. While wrapping
+    # to options.width.
+    lines = []
+    for line in s.splitlines(True):
+        if len(escape(line, encoding)) > options.width:
+            words = line.split()
+            words.reverse()
+            while words:
+                buf = []
+                size = 2
+                while words:
+                    word = words[-1]
+                    escaped_word = escape(word, encoding)
+                    add_space = 1 if buf else 0
+                    if size + len(escaped_word) + add_space <= options.width:
+                        buf.append(words.pop())
+                        size += len(escaped_word) + add_space
+                    else:
+                        if not buf:
+                            buf.append(words.pop())
+                        break
+                lines.append(' '.join(buf))
+        else:
+            lines.append(line)
+    if len(lines) <= 1:
+        return '"' + escape(s, encoding) + '"'
+    if lines and not lines[-1]:
+        del lines[-1]
+        lines[-1] += '\n'
+    return '""\n' + '\n'.join(
+        [f'"{escape(line, encoding)}"' for line in lines])
 
 
 def containsAny(str, set):
@@ -618,28 +636,10 @@ def write_pot_file(messages, options, fp):
             # to skip translating some unimportant docstrings.
             print('#, docstring', file=fp)
         if msg.msgctxt is not None:
-            print('msgctxt', normalize(msg.msgctxt, encoding), file=fp)
-
-        # If msgid is longer than width wrap
-        msgid = normalize(msg.msgid, encoding)[1:-1] # normalize returns "msg"
-        if len(msgid) > options.width:
-            print('msgid ""', file=fp)
-            while msgid:
-                print(f'"{msgid[:options.width]}"', file=fp)
-                msgid = msgid[options.width:]
-        else:
-            print(f'msgid "{msgid}"', file=fp)
-
-        # If msgid_plural is longer than width wrap
+            print('msgctxt', normalize(msg.msgctxt, encoding, options), file=fp)
+        print('msgid', normalize(msg.msgid, encoding, options), file=fp)
         if msg.msgid_plural is not None:
-            msgid_plural = normalize(msg.msgid_plural, encoding)[1:-1]  # normalize returns "msg"
-            if len(msgid_plural) > options.width:
-                print('msgid_plural ""', file=fp)
-                while msgid_plural:
-                    print(f'"{msgid_plural[:options.width]}"', file=fp)
-                    msgid_plural = msgid_plural[options.width:]
-            else:
-                print(f'msgid_plural "{msgid_plural}"', file=fp)
+            print('msgid_plural', normalize(msg.msgid_plural, encoding, options), file=fp)
             print('msgstr[0] ""', file=fp)
             print('msgstr[1] ""\n', file=fp)
         else: