Add and use batch forms of file matching calls.

src/python/pants/engine/legacy/address_mapper.py

@@ -46,20 +46,21 @@ def scan_build_files(self, base_path):
     return build_files_set
 
   @staticmethod
-  def is_declaring_file(address, file_path):
-    if not BuildFile._is_buildfile_name(os.path.basename(file_path)):
-      return False
-
+  def any_is_declaring_file(address, file_paths):
     try:
       # A precise check for BuildFileAddress
-      return address.rel_path == file_path
+      return address.rel_path in file_paths
     except AttributeError:
-      # NB: this will cause any BUILD file, whether it contains the address declaration or not to be
-      # considered the one that declared it. That's ok though, because the spec path should be enough
-      # information for debugging most of the time.
-      #
-      # TODO: remove this after https://github.com/pantsbuild/pants/issues/3925 lands
-      return os.path.dirname(file_path) == address.spec_path
+      pass
+    # NB: this will cause any BUILD file, whether it contains the address declaration or not to be
+    # considered the one that declared it. That's ok though, because the spec path should be enough
+    # information for debugging most of the time.
+    return any(address.spec_path == os.path.dirname(fp)
+               for fp in file_paths if BuildFile._is_buildfile_name(os.path.basename(fp)))
+
+  @staticmethod
+  def is_declaring_file(address, file_path):
+    return any_declaring_file(address, [file_path])
 
   def addresses_in_spec_path(self, spec_path):
     return self.scan_specs([SiblingAddresses(spec_path)])

src/python/pants/engine/legacy/source_mapper.py

@@ -14,7 +14,7 @@
 from pants.build_graph.source_mapper import SourceMapper
 from pants.engine.legacy.address_mapper import LegacyAddressMapper
 from pants.engine.legacy.graph import HydratedTargets
-from pants.source.filespec import matches_filespec
+from pants.source.filespec import any_matches_filespec
 
 
 def iter_resolve_and_parse_specs(rel_path, specs):
@@ -50,23 +50,26 @@ def _unique_dirs_for_sources(self, sources):
   def target_addresses_for_source(self, source):
     return list(self.iter_target_addresses_for_sources([source]))
 
-  def _match_source(self, source, fileset):
-    return fileset.matches(source) or matches_filespec(source, fileset.filespec)
+  def _match_sources(self, sources_set, fileset):
+    # NB: Deleted files can only be matched against the 'filespec' (ie, `PathGlobs`) for a target,
+    # so we don't actually call `fileset.matches` here.
+    # TODO: This call should be pushed down into the engine one way or another.
+    return any_matches_filespec(sources_set, fileset.filespec)
 
-  def _owns_source(self, source, legacy_target):
+  def _owns_any_source(self, sources_set, legacy_target):
     """Given a `HydratedTarget` instance, check if it owns the given source file."""
     target_kwargs = legacy_target.adaptor.kwargs()
 
     # Handle targets like `python_binary` which have a singular `source='main.py'` declaration.
     target_source = target_kwargs.get('source')
     if target_source:
       path_from_build_root = os.path.join(legacy_target.adaptor.address.spec_path, target_source)
-      if path_from_build_root == source:
+      if path_from_build_root in sources_set:
         return True
 
     # Handle `sources`-declaring targets.
     target_sources = target_kwargs.get('sources', [])
-    if target_sources and self._match_source(source, target_sources):
+    if target_sources and self._match_sources(sources_set, target_sources):
       return True
 
     return False
@@ -86,6 +89,6 @@ def iter_target_addresses_for_sources(self, sources):
 
     for hydrated_target, legacy_address in six.iteritems(hydrated_target_to_address):
       # Handle BUILD files.
-      if (any(LegacyAddressMapper.is_declaring_file(legacy_address, f) for f in sources_set) or
-          any(self._owns_source(source, hydrated_target) for source in sources_set)):
+      if (LegacyAddressMapper.any_is_declaring_file(legacy_address, sources_set) or
+          self._owns_any_source(sources_set, hydrated_target)):
         yield legacy_address

src/python/pants/source/filespec.py

@@ -44,16 +44,28 @@ def glob_to_regex(pattern):
   return ''.join(out)
 
 
-def globs_matches(path, patterns):
-  return any(re.match(glob_to_regex(pattern), path) for pattern in patterns)
+def globs_matches(paths, patterns, exclude_patterns):
+  def excluded(path):
+    if excluded.regexes is None:
+      excluded.regexes = [re.compile(glob_to_regex(ex)) for ex in exclude_patterns]
+    return any(ex.match(path) for ex in excluded.regexes)
+  excluded.regexes = None
+  for pattern in patterns:
+    regex = re.compile(glob_to_regex(pattern))
+    for path in paths:
+      if regex.match(path) and not excluded(path):
+        return True
+  return False
 
 
 def matches_filespec(path, spec):
-  if spec is None:
-    return False
-  if not globs_matches(path, spec.get('globs', [])):
+  return any_matches_filespec([path], spec)
+
+
+def any_matches_filespec(paths, spec):
+  if not paths or not spec:
     return False
-  for spec in spec.get('exclude', []):
-    if matches_filespec(path, spec):
-      return False
-  return True
+  exclude_patterns = []
+  for exclude_spec in spec.get('exclude', []):
+    exclude_patterns.extend(exclude_spec.get('globs', []))
+  return globs_matches(paths, spec.get('globs', []), exclude_patterns)