Script to automatically update copyright notices in C/C++ source code.

Created to implement http://codereview.appspot.com/4816058/ ('Automatic update of all copyright notices to reflect new license terms.')

We can also use this to periodically clean up our code headers.
Review URL: http://codereview.appspot.com/4800055

git-svn-id: http://skia.googlecode.com/svn/trunk@1983 2bbb7eff-a529-9590-31e7-b0007b416f81

2 files changed, 199 insertions, 0 deletions

diff --git a/tools/copyright/fileparser.py b/tools/copyright/fileparser.py
new file mode 100644
index 0000000000..a4051e25b9
--- /dev/null
+++ b/tools/copyright/fileparser.py
@@ -0,0 +1,92 @@
+'''
+Copyright 2011 Google Inc.
+
+Use of this source code is governed by a BSD-style license that can be
+found in the LICENSE file.
+'''
+
+import datetime
+import re
+
+def CreateParser(filepath):
+    """Returns a Parser as appropriate for the file at this filepath.
+    """
+    if (filepath.endswith('.cpp') or
+        filepath.endswith('.h') or
+        filepath.endswith('.c')):
+        return CParser()
+    else:
+        return None
+
+
+class Parser(object):
+    """Base class for all language-specific parsers.
+    """
+
+    def __init__(self):
+        self._copyright_pattern = re.compile('copyright', re.IGNORECASE)
+        self._attribute_pattern = re.compile(
+            'copyright.*\D(\d{4})\W*(\w.*[\w.])', re.IGNORECASE)
+
+    def FindCopyrightBlock(self, comment_blocks):
+        """Given a list of comment block strings, return the one that seems
+        like the most likely copyright block.
+
+        Returns None if comment_blocks was empty, or if we couldn't find
+        a comment block that contains copyright info."""
+        if not comment_blocks:
+            return None
+        for block in comment_blocks:
+            if self._copyright_pattern.search(block):
+                return block
+
+    def GetCopyrightBlockAttributes(self, comment_block):
+        """Given a comment block, return a tuple of attributes: (year, holder).
+
+        If comment_block is None, or none of the attributes are found,
+        this will return (None, None)."""
+        if not comment_block:
+            return (None, None)
+        matches = self._attribute_pattern.findall(comment_block)
+        if not matches:
+            return (None, None)
+        first_match = matches[0]
+        return (first_match[0], first_match[1])
+
+
+class CParser(Parser):
+    """Parser that knows how to parse C/C++ files.
+    """
+
+    DEFAULT_YEAR = datetime.date.today().year
+    DEFAULT_HOLDER = 'Google Inc.'
+    COPYRIGHT_BLOCK_FORMAT = '''
+/*
+ * Copyright %s %s
+ *
+ * Use of this source code is governed by a BSD-style license that can be
+ * found in the LICENSE file.
+ */
+'''
+
+    def __init__(self):
+        super(CParser, self).__init__()
+        self._comment_pattern = re.compile('/\*.*?\*/', re.DOTALL)
+
+    def FindAllCommentBlocks(self, file_contents):
+        """Returns a list of all comment blocks within these file contents.
+        """
+        return self._comment_pattern.findall(file_contents)
+
+    def CreateCopyrightBlock(self, year, holder):
+        """Returns a copyright block suitable for this language, with the
+        given attributes.
+
+        @param year year in which to hold copyright (defaults to DEFAULT_YEAR)
+        @param holder holder of copyright (defaults to DEFAULT_HOLDER)
+        """
+        if not year:
+            year = self.DEFAULT_YEAR
+        if not holder:
+            holder = self.DEFAULT_HOLDER
+        return self.COPYRIGHT_BLOCK_FORMAT % (year, holder)
diff --git a/tools/copyright/main.py b/tools/copyright/main.py
new file mode 100644
index 0000000000..24969a7ac5
--- /dev/null
+++ b/tools/copyright/main.py
@@ -0,0 +1,107 @@
+'''
+Copyright 2011 Google Inc.
+
+Use of this source code is governed by a BSD-style license that can be
+found in the LICENSE file.
+'''
+
+'''
+Updates all copyright headers within our code:
+- For files that already have a copyright header, the header is modified
+  while keeping the year and holder intact.
+- For files that don't have a copyright header, we add one with the current
+  year and default holder.
+
+@author: epoger@google.com
+'''
+
+import os
+import sys
+
+import fileparser
+
+# Only modify copyright stanzas if the copyright holder is one of these.
+ALLOWED_COPYRIGHT_HOLDERS = [
+    'Google Inc.',
+    'Skia',
+    'The Android Open Source Project',
+]
+
+def Main(root_directory):
+    """Run everything.
+
+    @param root_directory root directory within which to modify all files
+    """
+    filepaths = GetAllFilepaths(root_directory)
+    for filepath in filepaths:
+        parser = fileparser.CreateParser(filepath)
+        if not parser:
+            ReportWarning('cannot find a parser for file %s, skipping...' %
+                          filepath)
+            continue
+        old_file_contents = ReadFileIntoString(filepath)
+        comment_blocks = parser.FindAllCommentBlocks(old_file_contents)
+        if not comment_blocks:
+            ReportWarning('cannot find any comment blocks in file %s' %
+                          filepath)
+        old_copyright_block = parser.FindCopyrightBlock(comment_blocks)
+        if not old_copyright_block:
+            ReportWarning('cannot find copyright block in file %s' % filepath)
+        (year, holder) = parser.GetCopyrightBlockAttributes(old_copyright_block)
+        if holder and not ConfirmAllowedCopyrightHolder(holder):
+            ReportWarning(
+                'unrecognized copyright holder "%s" in file %s, skipping...' % (
+                    holder, filepath))
+            continue
+        new_copyright_block = parser.CreateCopyrightBlock(year, holder)
+        if old_copyright_block:
+            new_file_contents = old_file_contents.replace(
+                old_copyright_block, new_copyright_block, 1)
+        else:
+            new_file_contents = new_copyright_block + old_file_contents
+        WriteStringToFile(new_file_contents, filepath)
+
+def GetAllFilepaths(root_directory):
+    """Return a list of all files (absolute path for each one) within a tree.
+
+    @param root_directory root directory within which to find all files
+    """
+    path_list = []
+    for dirpath, dirnames, filenames in os.walk(root_directory):
+        for filename in filenames:
+            path_list.append(os.path.abspath(os.path.join(dirpath, filename)))
+    return path_list
+
+def ReportWarning(text):
+    """Report a warning, but continue.
+    """
+    print 'warning: %s' % text
+
+def ReportError(text):
+    """Report an error and raise an exception.
+    """
+    raise IOError(text)
+
+def ReadFileIntoString(filepath):
+    """Returns the full contents of this file as a string.
+    """
+    with open(filepath, 'r') as file_handle:
+        contents = file_handle.read()
+    return contents
+
+def WriteStringToFile(string, filepath):
+    """Writes this string out to filepath, replacing the file if it already
+    exists.
+    """
+    with open(filepath, 'w') as file_handle:
+        file_handle.write(string)
+
+def ConfirmAllowedCopyrightHolder(holder):
+    """Returns True if this is one of our allowed copyright holders.
+
+    @param holder copyright holder as a string
+    """
+    return holder in ALLOWED_COPYRIGHT_HOLDERS
+
+
+Main(sys.argv[1])