Check for forbidden HTML tags

CMK-17499

checkmk_weblate_syncer/html_tags.py

@@ -0,0 +1,18 @@
+import re
+
+# keep in sync with tests/pylint/checker_localization.py:HTMLTagsChecker
+_TAG_PATTERN = re.compile("<.*?>")
+_ALLOWED_TAGS_PATTERN = re.compile(
+    r"</?(h1|h2|b|tt|i|u|hr|br(?: /)?|nobr(?: /)?|pre|sup|p|li|ul|ol|a|(a.*? href=.*?))>"
+)
+
+
+def forbidden_tags(text: str) -> set[str]:
+    return {
+        tag
+        for tag in re.findall(
+            _TAG_PATTERN,
+            text,
+        )
+        if not re.match(_ALLOWED_TAGS_PATTERN, tag)
+    }

checkmk_weblate_syncer/portable_object.py

@@ -2,6 +2,18 @@
 from pathlib import Path
 
 
+def remove_header(portable_object_content: str) -> str:
+    lines = portable_object_content.splitlines()
+    index_first_source_string_location = 0
+    for index, line in enumerate(lines):
+        if re.match(r"^#: .*?:\d+$", line):
+            index_first_source_string_location = index
+            break
+    return "\n".join(lines[index_first_source_string_location:]) + (
+        "\n" if portable_object_content.endswith("\n") else ""
+    )
+
+
 def make_soure_string_locations_relative(
     portable_object_content: str,
     relative_to: Path,

checkmk_weblate_syncer/update_sources.py

@@ -3,8 +3,9 @@
 
 from .config import UpdateSourcesConfig
 from .git import commit_and_push_files, repository_in_clean_state
+from .html_tags import forbidden_tags
 from .logging import LOGGER
-from .portable_object import make_soure_string_locations_relative
+from .portable_object import make_soure_string_locations_relative, remove_header
 
 
 def run(config: UpdateSourcesConfig) -> int:
@@ -30,6 +31,12 @@ def run(config: UpdateSourcesConfig) -> int:
         LOGGER.error("Generating pot file failed")
         raise e
 
+    LOGGER.info("Checking HTML tags")
+    if forbidden_html_tags := forbidden_tags(remove_header(pot_file_content)):
+        raise ValueError(
+            f"Found forbidden HTML tags: {', '.join(sorted(forbidden_html_tags))}"
+        )
+
     LOGGER.info("Making source string locations relative")
     pot_file_content = make_soure_string_locations_relative(
         portable_object_content=pot_file_content,

checkmk_weblate_syncer/update_translations.py

@@ -8,8 +8,13 @@
 
 from .config import PoFilePair, RepositoryConfig, UpdateTranslationsConfig
 from .git import commit_and_push_files, repository_in_clean_state
+from .html_tags import forbidden_tags
 from .logging import LOGGER
-from .portable_object import remove_last_translator, remove_source_string_locations
+from .portable_object import (
+    remove_header,
+    remove_last_translator,
+    remove_source_string_locations,
+)
 
 
 @dataclass(frozen=True)
@@ -105,6 +110,13 @@ def _process_po_file_pair(
             path=locale_po_file,
         )
 
+    LOGGER.info("Checking HTML tags")
+    if forbidden_html_tags := forbidden_tags(remove_header(po_file_content)):
+        return _Failure(
+            error_message=f"Found forbidden HTML tags: {', '.join(sorted(forbidden_html_tags))}",
+            path=locale_po_file,
+        )
+
     LOGGER.info("Stripping source string locations and Last-Translator")
     po_file_content = remove_source_string_locations(po_file_content)
     po_file_content = remove_last_translator(po_file_content)

tests/test_html_tags.py

@@ -0,0 +1,59 @@
+import pytest
+
+from checkmk_weblate_syncer.html_tags import forbidden_tags
+
+
+@pytest.mark.parametrize(
+    ["text", "expected_result"],
+    [
+        pytest.param(
+            "abc123",
+            frozenset(),
+        ),
+        pytest.param(
+            "<tt>bold</tt>",
+            frozenset(),
+        ),
+        pytest.param(
+            '* ? <a href="%s">%s</a>',
+            frozenset(),
+        ),
+        pytest.param(
+            '&copy; <a target="_blank" href="https://checkmk.com">Checkmk GmbH</a>',
+            frozenset(),
+        ),
+        pytest.param(
+            "123 <script>injection</script>",
+            frozenset(
+                ["<script>", "</script>"],
+            ),
+        ),
+        pytest.param(
+            # pylint: disable=line-too-long
+            """#: /home/weblate/checkmk_weblate_sync/git/checkmk/cmk/gui/wato/pages/host_rename.py:640
+#, python-format
+msgid " (%d times)"
+msgstr ""
+
+#: /home/weblate/checkmk_weblate_sync/git/checkmk/cmk/gui/visuals/_page_edit_visual.py:137
+msgid " (Copy)"
+msgstr ""
+
+#: /home/weblate/checkmk_weblate_sync/git/checkmk/cmk/gui/nodevis/topology.py:1814
+msgid " (Data incomplete, maximum number of nodes reached)"
+msgstr ""
+
+#: /home/weblate/checkmk_weblate_sync/git/checkmk/cmk/gui/backup/handler.py:969
+#, python-format
+msgid " (Duration: %s)"
+msgstr ""
+""",
+            frozenset(),
+        ),
+    ],
+)
+def test_html_tags_checker(
+    text: str,
+    expected_result: frozenset[str],
+) -> None:
+    assert forbidden_tags(text) == expected_result

tests/test_portable_object.py

@@ -2,11 +2,73 @@
 
 from checkmk_weblate_syncer.portable_object import (
     make_soure_string_locations_relative,
+    remove_header,
     remove_last_translator,
     remove_source_string_locations,
 )
 
 
+def test_remove_header() -> None:
+    assert (
+        remove_header(
+            """# Copyright (C) 2019 Checkmk GmbH - License: GNU General Public License v2
+# This file is part of Checkmk (https://checkmk.com). It is subject to the terms and
+# conditions defined in the file COPYING, which is part of this source code package.
+
+msgid ""
+msgstr ""
+"Project-Id-Version: Checkmk user interface translation 0.1\n"
+"Report-Msgid-Bugs-To: [email protected]\n"
+"POT-Creation-Date: 2011-05-13 09:42+0200\n"
+"PO-Revision-Date: YEAR-MO-DA HO:MI+ZONE\n"
+"Last-Translator: FULL NAME <EMAIL@ADDRESS>\n"
+"Language-Team: LANGUAGE <[email protected]>\n"
+"Language: LANGUAGE \n"
+"MIME-Version: 1.0\n"
+"Content-Type: text/plain; charset=utf-8\n"
+"Content-Transfer-Encoding: 8bit\n"
+
+#: /home/weblate/checkmk_weblate_sync/git/checkmk/cmk/gui/wato/pages/host_rename.py:640
+#, python-format
+msgid " (%d times)"
+msgstr ""
+
+#: /home/weblate/checkmk_weblate_sync/git/checkmk/cmk/gui/visuals/_page_edit_visual.py:137
+msgid " (Copy)"
+msgstr ""
+
+#: /home/weblate/checkmk_weblate_sync/git/checkmk/cmk/gui/nodevis/topology.py:1814
+msgid " (Data incomplete, maximum number of nodes reached)"
+msgstr ""
+
+#: /home/weblate/checkmk_weblate_sync/git/checkmk/cmk/gui/backup/handler.py:969
+#, python-format
+msgid " (Duration: %s)"
+msgstr ""
+"""
+        )
+        # pylint: disable=line-too-long
+        == """#: /home/weblate/checkmk_weblate_sync/git/checkmk/cmk/gui/wato/pages/host_rename.py:640
+#, python-format
+msgid " (%d times)"
+msgstr ""
+
+#: /home/weblate/checkmk_weblate_sync/git/checkmk/cmk/gui/visuals/_page_edit_visual.py:137
+msgid " (Copy)"
+msgstr ""
+
+#: /home/weblate/checkmk_weblate_sync/git/checkmk/cmk/gui/nodevis/topology.py:1814
+msgid " (Data incomplete, maximum number of nodes reached)"
+msgstr ""
+
+#: /home/weblate/checkmk_weblate_sync/git/checkmk/cmk/gui/backup/handler.py:969
+#, python-format
+msgid " (Duration: %s)"
+msgstr ""
+"""
+    )
+
+
 def test_make_soure_string_locations_relative() -> None:
     assert (
         make_soure_string_locations_relative(