Refactor sphinx.io; independ parser generation from reader class

sphinx/io.py

@@ -50,37 +50,6 @@ class SphinxBaseReader(standalone.Reader):
     """
     Add our source parsers
     """
-    def __init__(self, app, parsers={}, *args, **kwargs):
-        # type: (Sphinx, Dict[unicode, Parser], Any, Any) -> None
-        standalone.Reader.__init__(self, *args, **kwargs)
-        self.parser_map = {}  # type: Dict[unicode, Parser]
-        for suffix, parser_class in parsers.items():
-            if isinstance(parser_class, string_types):
-                parser_class = import_object(parser_class, 'source parser')  # type: ignore
-            parser = parser_class()
-            if hasattr(parser, 'set_application'):
-                parser.set_application(app)
-            self.parser_map[suffix] = parser
-
-    def read(self, source, parser, settings):
-        # type: (Input, Parser, Dict) -> nodes.document
-        self.source = source
-
-        for suffix in self.parser_map:
-            if source.source_path.endswith(suffix):
-                self.parser = self.parser_map[suffix]
-                break
-        else:
-            # use special parser for unknown file-extension '*' (if exists)
-            self.parser = self.parser_map.get('*')
-
-        if not self.parser:
-            self.parser = parser
-        self.settings = settings
-        self.input = self.source.read()
-        self.parse()
-        return self.document
-
     def get_transforms(self):
         # type: () -> List[Transform]
         return standalone.Reader.get_transforms(self) + self.transforms
@@ -114,17 +83,13 @@ class SphinxI18nReader(SphinxBaseReader):
     This class provides the correct line numbers when reporting.
     """
 
+    lineno = None  # type: int
     transforms = [ApplySourceWorkaround, ExtraTranslatableNodes, CitationReferences,
                   DefaultSubstitutions, MoveModuleTargets, HandleCodeBlocks,
                   AutoNumbering, SortIds, RemoveTranslatableInline,
                   FilterSystemMessages, RefOnlyBulletListTransform,
                   UnreferencedFootnotesDetector]
 
-    def __init__(self, *args, **kwargs):
-        # type: (Any, Any) -> None
-        SphinxBaseReader.__init__(self, *args, **kwargs)
-        self.lineno = None  # type: int
-
     def set_lineno_for_reporter(self, lineno):
         # type: (int) -> None
         self.lineno = lineno
@@ -216,11 +181,13 @@ class SphinxFileInput(FileInput):
 def read_doc(app, env, filename):
     # type: (Sphinx, BuildEnvironment, unicode) -> nodes.document
     """Parse a document and convert to doctree."""
-    reader = SphinxStandaloneReader(app, parsers=app.registry.get_source_parsers())
+    reader = SphinxStandaloneReader()
     source = SphinxFileInput(app, env, source=None, source_path=filename,
                              encoding=env.config.source_encoding)
+    parser = app.registry.create_source_parser(app, filename)
 
     pub = Publisher(reader=reader,
+                    parser=parser,
                     writer=SphinxDummyWriter(),
                     source_class=SphinxDummySourceClass,
                     destination=NullOutput())

sphinx/registry.py

@@ -13,15 +13,17 @@ from __future__ import print_function
 import traceback
 
 from pkg_resources import iter_entry_points
-from six import itervalues
+from six import iteritems, itervalues, string_types
 
 from sphinx.errors import ExtensionError, SphinxError, VersionRequirementError
 from sphinx.extension import Extension
 from sphinx.domains import ObjType
 from sphinx.domains.std import GenericObject, Target
 from sphinx.locale import __
+from sphinx.parsers import Parser as SphinxParser
 from sphinx.roles import XRefRole
 from sphinx.util import logging
+from sphinx.util import import_object
 from sphinx.util.docutils import directive_helper
 
 if False:
@@ -160,6 +162,25 @@ class SphinxComponentRegistry(object):
             raise ExtensionError(__('source_parser for %r is already registered') % suffix)
         self.source_parsers[suffix] = parser
 
+    def create_source_parser(self, app, filename):
+        # type: (Sphinx, unicode) -> Parser
+        for suffix, parser_class in iteritems(self.source_parsers):
+            if filename.endswith(suffix):
+                break
+        else:
+            # use special parser for unknown file-extension '*' (if exists)
+            parser_class = self.source_parsers.get('*')
+
+        if parser_class is None:
+            raise SphinxError(__('Source parser for %s not registered') % filename)
+        else:
+            if isinstance(parser_class, string_types):
+                parser_class = import_object(parser_class, 'source parser')  # type: ignore
+            parser = parser_class()
+            if isinstance(parser, SphinxParser):
+                parser.set_application(app)
+            return parser
+
     def get_source_parsers(self):
         # type: () -> Dict[unicode, Parser]
         return self.source_parsers

sphinx/transforms/i18n.py

@@ -50,15 +50,12 @@ def publish_msgstr(app, source, source_path, source_line, config, settings):
     :rtype: docutils.nodes.document
     """
     from sphinx.io import SphinxI18nReader
-    reader = SphinxI18nReader(
+    reader = SphinxI18nReader()
-        app=app,
-        parsers=app.registry.get_source_parsers(),
-        parser_name='restructuredtext',  # default parser
-    )
     reader.set_lineno_for_reporter(source_line)
+    parser = app.registry.create_source_parser(app, '')
     doc = reader.read(
         source=StringInput(source=source, source_path=source_path),
-        parser=reader.parser,
+        parser=parser,
         settings=settings,
     )
     try: