Fix extensions

cpplint.py

@@ -44,6 +44,7 @@
 import codecs
 import copy
 import getopt
+import itertools
 import math  # for log
 import os
 import re
@@ -52,15 +53,42 @@
 import sys
 import unicodedata
 
+# if empty, use defaults
+_header_extensions = set([])
+
+# if empty, use defaults
+_valid_extensions = set([])
+
+
+# Files with any of these extensions are considered to be
+# header files (and will undergo different style checks).
+# This set can be extended by using the --headers
+# option (also supported in CPPLINT.cfg)
+def GetHeaderExtensions():
+  if not _header_extensions:
+    return set(['h', 'hpp', 'hxx', 'h++', 'cuh'])
+  return _header_extensions
+
 # The allowed extensions for file names
-# This is set by --extensions flag.
-_valid_extensions = set(['c', 'cc', 'cpp', 'cxx', 'c++', 'h', 'hpp', 'hxx',
-    'h++'])
+# This is set by --extensions flag
+def GetAllExtensions():
+  if not _valid_extensions:
+    return GetHeaderExtensions().union(set(['c', 'cc', 'cpp', 'cxx', 'c++', 'cu']))
+  return _valid_extensions
+
+def GetNonHeaderExtensions():
+  return GetAllExtensions().difference(GetHeaderExtensions())
+
+
+# files with this suffix before the extension will be treated as test files
+_test_suffixes = set(['_unittest', '_test', '_regtest'])
 
 _USAGE = """
 Syntax: cpplint.py [--verbose=#] [--output=vs7] [--filter=-x,+y,...]
                    [--counting=total|toplevel|detailed] [--root=subdir]
                    [--linelength=digits]
+                   [--headers=ext1,ext2]
+                   [--extensions=ext1,ext2]
         <file> [file] ...
 
   The style guidelines this tries to follow are those in
@@ -137,7 +165,15 @@
       The allowed file extensions that cpplint will check
 
       Examples:
-        --extensions=hpp,cpp
+        --extensions=%s
+
+    headers=extension,extension,...
+      The allowed header extensions that cpplint will consider to be header files
+      (by default, only files with extensions %s
+      will be assumed to be headers)
+
+      Examples:
+        --headers=%s
 
     cpplint.py supports per-directory configurations specified in CPPLINT.cfg
     files. CPPLINT.cfg file can contain a number of key=value pairs.
@@ -173,7 +209,10 @@
     build/include_alpha as well as excludes all .cc from being
     processed by linter, in the current directory (where the .cfg
     file is located) and all sub-directories.
-""" % (list(_valid_extensions))
+""" % (list(GetAllExtensions()),
+       ','.join(list(GetAllExtensions())),
+       GetHeaderExtensions(),
+       ','.join(GetHeaderExtensions()))
 
 # We categorize each error message we print.  Here are the categories.
 # We want an explicit list so we can list them all in cpplint --filter=.
@@ -525,6 +564,7 @@ def u(x):
     itervalues = dict.values
     iteritems = dict.items
 
+
 def ParseNolintSuppressions(filename, raw_line, linenum, error):
   """Updates the global list of error-suppressions.
 
@@ -1074,7 +1114,7 @@ def BaseName(self):
     return self.Split()[1]
 
   def Extension(self):
-    """File extension - text following the final period."""
+    """File extension - text following the final period, includes that period."""
     return self.Split()[2]
 
   def NoExtension(self):
@@ -1083,7 +1123,7 @@ def NoExtension(self):
 
   def IsSource(self):
     """File has a source file extension."""
-    return self.Extension()[1:] in _valid_extensions
+    return self.Extension()[1:] in GetAllExtensions()
 
 
 def _ShouldPrintError(category, confidence, linenum):
@@ -1798,28 +1838,29 @@ def CheckForHeaderGuard(filename, clean_lines, error):
 
 
 def CheckHeaderFileIncluded(filename, include_state, error):
-  """Logs an error if a .cc file does not include its header."""
+  """Logs an error if a source file does not include its header."""
 
   # Do not check test files
-  if filename.endswith('_test.cc') or filename.endswith('_unittest.cc'):
+  if _IsTestFilename(filename):
     return
 
   fileinfo = FileInfo(filename)
-  headerfile = filename[0:len(filename) - 2] + 'h'
-  if not os.path.exists(headerfile):
-    return
-  headername = FileInfo(headerfile).RepositoryName()
-  first_include = 0
-  for section_list in include_state.include_list:
-    for f in section_list:
-      if headername in f[0] or f[0] in headername:
-        return
-      if not first_include:
-        first_include = f[1]
+  for ext in GetHeaderExtensions():
+      headerfile = filename[:filename.rfind('.') + 1] + ext
+      if not os.path.exists(headerfile):
+        continue
+      headername = FileInfo(headerfile).RepositoryName()
+      first_include = None
+      for section_list in include_state.include_list:
+        for f in section_list:
+          if headername in f[0] or f[0] in headername:
+            return
+          if not first_include:
+            first_include = f[1]
 
-  error(filename, first_include, 'build/include', 5,
-        '%s should include its header file %s' % (fileinfo.RepositoryName(),
-                                                  headername))
+      error(filename, first_include, 'build/include', 5,
+            '%s should include its header file %s' % (fileinfo.RepositoryName(),
+                                                      headername))
 
 
 def CheckForBadCharacters(filename, lines, error):
@@ -4460,7 +4501,7 @@ def CheckStyle(filename, clean_lines, linenum, file_extension, nesting_state,
 
   # Check if the line is a header guard.
   is_header_guard = False
-  if file_extension == 'h':
+  if file_extension in GetHeaderExtensions():
     cppvar = GetHeaderGuardCPPVariable(filename)
     if (line.startswith('#ifndef %s' % cppvar) or
         line.startswith('#define %s' % cppvar) or
@@ -4553,8 +4594,11 @@ def _DropCommonSuffixes(filename):
   Returns:
     The filename with the common suffix removed.
   """
-  for suffix in ('test.cc', 'regtest.cc', 'unittest.cc',
-                 'inl.h', 'impl.h', 'internal.h'):
+  for suffix in itertools.chain(
+      ('%s.%s' % (test_suffix.lstrip('_'), ext)
+       for test_suffix, ext in itertools.product(_test_suffixes, GetNonHeaderExtensions())),
+      ('%s.%s' % (suffix, ext)
+       for suffix, ext in itertools.product(['inl', 'imp', 'internal'], GetHeaderExtensions()))):
     if (filename.endswith(suffix) and len(filename) > len(suffix) and
         filename[-len(suffix) - 1] in ('-', '_')):
       return filename[:-len(suffix) - 1]
@@ -4570,12 +4614,10 @@ def _IsTestFilename(filename):
   Returns:
     True if 'filename' looks like a test, False otherwise.
   """
-  if (filename.endswith('_test.cc') or
-      filename.endswith('_unittest.cc') or
-      filename.endswith('_regtest.cc')):
-    return True
-  else:
-    return False
+  for test_suffix, ext in itertools.product(_test_suffixes, GetNonHeaderExtensions()):
+    if filename.endswith(test_suffix + '.' + ext):
+      return True
+  return False
 
 
 def _ClassifyInclude(fileinfo, include, is_system):
@@ -4679,11 +4721,16 @@ def CheckIncludeLine(filename, clean_lines, linenum, include_state, error):
       error(filename, linenum, 'build/include', 4,
             '"%s" already included at %s:%s' %
             (include, filename, duplicate_line))
-    elif (include.endswith('.cc') and
+      return
+
+    for extension in GetNonHeaderExtensions():
+      if (include.endswith('.' + extension) and
           os.path.dirname(fileinfo.RepositoryName()) != os.path.dirname(include)):
-      error(filename, linenum, 'build/include', 4,
-            'Do not include .cc files from other packages')
-    elif not _THIRD_PARTY_HEADERS_PATTERN.match(include):
+        error(filename, linenum, 'build/include', 4,
+              'Do not include .' + extension + ' files from other packages')
+        return
+
+    if not _THIRD_PARTY_HEADERS_PATTERN.match(include):
       include_state.include_list[-1].append((include, linenum))
 
       # We want to ensure that headers appear in the right order:
@@ -4833,7 +4880,7 @@ def CheckLanguage(filename, clean_lines, linenum, file_extension,
   CheckGlobalStatic(filename, clean_lines, linenum, error)
   CheckPrintf(filename, clean_lines, linenum, error)
 
-  if file_extension == 'h':
+  if file_extension in GetHeaderExtensions():
     # TODO(unknown): check that 1-arg constructors are explicit.
     #                How to tell it's a constructor?
     #                (handled in CheckForNonStandardConstructs for now)
@@ -4940,7 +4987,7 @@ def CheckLanguage(filename, clean_lines, linenum, file_extension,
   # Check for use of unnamed namespaces in header files.  Registration
   # macros are typically OK, so we allow use of "namespace {" on lines
   # that end with backslashes.
-  if (file_extension == 'h'
+  if (file_extension in GetHeaderExtensions()
       and Search(r'\bnamespace\s*{', line)
       and line[-1] != '\\'):
     error(filename, linenum, 'build/namespaces', 4,
@@ -5573,28 +5620,32 @@ def FilesBelongToSameModule(filename_cc, filename_h):
   some false positives. This should be sufficiently rare in practice.
 
   Args:
-    filename_cc: is the path for the .cc file
+    filename_cc: is the path for the source (e.g. .cc) file
     filename_h: is the path for the header path
 
   Returns:
     Tuple with a bool and a string:
     bool: True if filename_cc and filename_h belong to the same module.
     string: the additional prefix needed to open the header file.
   """
+  fileinfo_cc = FileInfo(filename_cc)
+  if not fileinfo_cc.Extension().lstrip('.') in GetNonHeaderExtensions():
+    return (False, '')
 
-  if not filename_cc.endswith('.cc'):
+  fileinfo_h = FileInfo(filename_h)
+  if not fileinfo_h.Extension().lstrip('.') in GetHeaderExtensions():
     return (False, '')
-  filename_cc = filename_cc[:-len('.cc')]
-  if filename_cc.endswith('_unittest'):
-    filename_cc = filename_cc[:-len('_unittest')]
-  elif filename_cc.endswith('_test'):
-    filename_cc = filename_cc[:-len('_test')]
+
+  filename_cc = filename_cc[:-(len(fileinfo_cc.Extension()))]
+  for suffix in _test_suffixes:
+    if filename_cc.endswith(suffix):
+      filename_cc = filename_cc[:-len(suffix)]
+      break
+
   filename_cc = filename_cc.replace('/public/', '/')
   filename_cc = filename_cc.replace('/internal/', '/')
 
-  if not filename_h.endswith('.h'):
-    return (False, '')
-  filename_h = filename_h[:-len('.h')]
+  filename_h = filename_h[:-(len(fileinfo_h.Extension()))]
   if filename_h.endswith('-inl'):
     filename_h = filename_h[:-len('-inl')]
   filename_h = filename_h.replace('/public/', '/')
@@ -5716,8 +5767,10 @@ def CheckForIncludeWhatYouUse(filename, clean_lines, include_state, error,
   # didn't include it in the .h file.
   # TODO(unknown): Do a better job of finding .h files so we are confident that
   # not having the .h file means there isn't one.
-  if filename.endswith('.cc') and not header_found:
-    return
+  if not header_found:
+    for extension in GetNonHeaderExtensions():
+      if filename.endswith('.' + extension):
+        return
 
   # All the lines have been processed, report the errors found.
   for required_header_unstripped in required:
@@ -6056,7 +6109,7 @@ def ProcessFileData(filename, file_extension, lines, error,
   RemoveMultiLineComments(filename, lines, error)
   clean_lines = CleansedLines(lines)
 
-  if file_extension == 'h':
+  if file_extension in GetHeaderExtensions():
     CheckForHeaderGuard(filename, clean_lines, error)
 
   for line in range(clean_lines.NumLines()):
@@ -6069,7 +6122,7 @@ def ProcessFileData(filename, file_extension, lines, error,
   CheckForIncludeWhatYouUse(filename, clean_lines, include_state, error)
 
   # Check that the .cc file has included its header if it exists.
-  if file_extension == 'cc':
+  if file_extension in GetNonHeaderExtensions():
     CheckHeaderFileIncluded(filename, include_state, error)
 
   # We check here rather than inside ProcessLine so that we see raw
@@ -6136,6 +6189,24 @@ def ProcessConfigOverrides(filename):
                 _line_length = int(val)
             except ValueError:
                 sys.stderr.write('Line length must be numeric.')
+          elif name == 'extensions':
+              global _valid_extensions
+              try:
+                  extensions = [ext.strip() for ext in val.split(',')]
+                  _valid_extensions = set(extensions)
+              except ValueError:
+                  sys.stderr.write('Extensions should be a comma-separated list of values;'
+                                   'for example: extensions=hpp,cpp\n'
+                                   'This could not be parsed: "%s"' % (val,))
+          elif name == 'headers':
+              global _header_extensions
+              try:
+                  extensions = [ext.strip() for ext in val.split(',')]
+                  _header_extensions = set(extensions)
+              except ValueError:
+                  sys.stderr.write('Extensions should be a comma-separated list of values;'
+                                   'for example: extensions=hpp,cpp\n'
+                                   'This could not be parsed: "%s"' % (val,))
           else:
             sys.stderr.write(
                 'Invalid configuration option (%s) in file %s\n' %
@@ -6213,9 +6284,9 @@ def ProcessFile(filename, vlevel, extra_check_functions=[]):
 
   # When reading from stdin, the extension is unknown, so no cpplint tests
   # should rely on the extension.
-  if filename != '-' and file_extension not in _valid_extensions:
+  if filename != '-' and file_extension not in GetAllExtensions():
     sys.stderr.write('Ignoring %s; not a valid file name '
-                     '(%s)\n' % (filename, ', '.join(_valid_extensions)))
+                     '(%s)\n' % (filename, ', '.join(GetAllExtensions())))
   else:
     ProcessFileData(filename, file_extension, lines, Error,
                     extra_check_functions)
@@ -6282,7 +6353,8 @@ def ParseArguments(args):
                                                  'filter=',
                                                  'root=',
                                                  'linelength=',
-                                                 'extensions='])
+                                                 'extensions=',
+                                                 'headers='])
   except getopt.GetoptError:
     PrintUsage('Invalid arguments.')
 
@@ -6323,6 +6395,12 @@ def ParseArguments(args):
           _valid_extensions = set(val.split(','))
       except ValueError:
           PrintUsage('Extensions must be comma seperated list.')
+    elif opt == '--headers':
+      global _header_extensions
+      try:
+          _header_extensions = set(val.split(','))
+      except ValueError:
+          PrintUsage('Extensions must be comma seperated list.')
 
   if not filenames:
     PrintUsage('No files were specified.')