Simplify _publish_msgstr() (#13670)

Author: AA-Turner

sphinx/transforms/i18n.py

@@ -2,23 +2,21 @@
 
 from __future__ import annotations
 
-import contextlib
 from re import DOTALL, match
 from textwrap import indent
 from typing import TYPE_CHECKING, Any, TypeVar
 
+import docutils.utils
 from docutils import nodes
-from docutils.io import StringInput
 
 from sphinx import addnodes
 from sphinx.domains.std import make_glossary_term, split_term_classifiers
 from sphinx.errors import ConfigError
-from sphinx.io import SphinxBaseReader
 from sphinx.locale import __
 from sphinx.locale import init as init_locale
-from sphinx.transforms import AutoIndexUpgrader, DoctreeReadEvent, SphinxTransform
-from sphinx.transforms.references import SphinxDomains
+from sphinx.transforms import SphinxTransform
 from sphinx.util import get_filetype, logging
+from sphinx.util.docutils import LoggingReporter
 from sphinx.util.i18n import docname_to_domain
 from sphinx.util.index_entries import split_index_msg
 from sphinx.util.nodes import (
@@ -28,11 +26,12 @@
     extract_messages,
     traverse_translatable_index,
 )
-from sphinx.versioning import UIDTransform
 
 if TYPE_CHECKING:
     from collections.abc import Sequence
 
+    from docutils.frontend import Values
+
     from sphinx.application import Sphinx
     from sphinx.config import Config
     from sphinx.environment import BuildEnvironment
@@ -52,77 +51,49 @@
 N = TypeVar('N', bound=nodes.Node)
 
 
-class _SphinxI18nReader(SphinxBaseReader):
-    """A document reader for internationalisation (i18n).
-
-    This returns the source line number of the original text
-    as the current source line number to let users know where
-    the error happened, because the translated texts are
-    partial and they don't have correct line numbers.
-    """
-
-    def __init__(
-        self, *args: Any, registry: SphinxComponentRegistry, **kwargs: Any
-    ) -> None:
-        super().__init__(*args, **kwargs)
-        unused = frozenset({
-            PreserveTranslatableMessages,
-            Locale,
-            RemoveTranslatableInline,
-            AutoIndexUpgrader,
-            SphinxDomains,
-            DoctreeReadEvent,
-            UIDTransform,
-        })
-        transforms = self.transforms + registry.get_transforms()
-        self.transforms = [
-            transform for transform in transforms if transform not in unused
-        ]
-
-
-def publish_msgstr(
+def _publish_msgstr(
     source: str,
     source_path: str,
     source_line: int,
-    config: Config,
-    settings: Any,
     *,
+    config: Config,
     env: BuildEnvironment,
     registry: SphinxComponentRegistry,
+    settings: Values,
 ) -> nodes.Element:
     """Publish msgstr (single line) into docutils document
 
     :param str source: source text
     :param str source_path: source path for warning indication
     :param source_line: source line for warning indication
-    :param sphinx.config.Config config: sphinx config
     :param docutils.frontend.Values settings: docutils settings
-    :return: document
-    :rtype: docutils.nodes.document
+    :param sphinx.config.Config config: sphinx config
     :param sphinx.environment.BuildEnvironment env: sphinx environment
     :param sphinx.registry.SphinxComponentRegistry registry: sphinx registry
+    :return: document
+    :rtype: docutils.nodes.document
     """
+    filetype = get_filetype(config.source_suffix, source_path)
+    doc = docutils.utils.new_document(
+        f'{source_path}:{source_line}:<translated>', settings
+    )
+    doc.reporter = LoggingReporter.from_reporter(doc.reporter)
+
+    # clear rst_prolog temporarily
+    rst_prolog = config.rst_prolog
+    config.rst_prolog = None
     try:
-        # clear rst_prolog temporarily
-        rst_prolog = config.rst_prolog
-        config.rst_prolog = None
-
-        reader = _SphinxI18nReader(registry=registry)
-        filetype = get_filetype(config.source_suffix, source_path)
         parser = registry.create_source_parser(filetype, config=config, env=env)
-        doc = reader.read(
-            source=StringInput(
-                source=source, source_path=f'{source_path}:{source_line}:<translated>'
-            ),
-            parser=parser,
-            settings=settings,
-        )
-        with contextlib.suppress(IndexError):  # empty node
-            return doc[0]
-        return doc
+        parser.parse(source, doc)
+        doc.current_source = doc.current_line = None
     finally:
         config.rst_prolog = rst_prolog
 
+    try:
+        return doc[0]  # type: ignore[return-value]
+    except IndexError:  # empty node
+        return doc
+
 
 def parse_noqa(source: str) -> tuple[str, bool]:
     m = match(r'(.*)(?<!\\)#\s*noqa\s*$', source, DOTALL)
@@ -470,14 +441,14 @@ def apply(self, **kwargs: Any) -> None:
             if isinstance(node, LITERAL_TYPE_NODES):
                 msgstr = '::\n\n' + indent(msgstr, ' ' * 3)
 
-            patch = publish_msgstr(
+            patch = _publish_msgstr(
                 msgstr,
                 source,
                 node.line,  # type: ignore[arg-type]
-                self.config,
-                settings,
+                config=self.config,
                 env=self.env,
                 registry=self.env._registry,
+                settings=settings,
             )
             # FIXME: no warnings about inconsistent references in this part
             # XXX doctest and other block markup
@@ -491,14 +462,14 @@ def apply(self, **kwargs: Any) -> None:
             if isinstance(node, nodes.term):
                 for _id in node['ids']:
                     term, first_classifier = split_term_classifiers(msgstr)
-                    patch = publish_msgstr(
+                    patch = _publish_msgstr(
                         term or '',
                         source,
                         node.line,  # type: ignore[arg-type]
-                        self.config,
-                        settings,
+                        config=self.config,
                         env=self.env,
                         registry=self.env._registry,
+                        settings=settings,
                     )
                     updater.patch = make_glossary_term(
                         self.env,
@@ -569,14 +540,14 @@ def apply(self, **kwargs: Any) -> None:
                 # This generates: <section ...><title>msgstr</title></section>
                 msgstr = msgstr + '\n' + '=' * len(msgstr) * 2
 
-            patch = publish_msgstr(
+            patch = _publish_msgstr(
                 msgstr,
                 source,
                 node.line,  # type: ignore[arg-type]
-                self.config,
-                settings,
+                config=self.config,
                 env=self.env,
                 registry=self.env._registry,
+                settings=settings,
             )
             # Structural Subelements phase2
             if isinstance(node, nodes.title):