Make webview_licenses.py script to use licensecheck.pl

android_webview/tools/webview_licenses.py

Index: android_webview/tools/webview_licenses.py
diff --git a/android_webview/tools/webview_licenses.py b/android_webview/tools/webview_licenses.py
index 0d2a7af9186baf90181f2f400cf369eee396d85b..a914b41c88d5b5a3781dca1ca5c81ad78d4c7c59 100755
--- a/android_webview/tools/webview_licenses.py
+++ b/android_webview/tools/webview_licenses.py
@@ -86,64 +86,45 @@ def _CheckLicenseHeaders(directory_list, whitelisted_files):
     whitelist contains no stale entries, otherwise false.
   """
 
-  # Matches one of ...
-  # - '[Cc]opyright', but not when followed by
-  #   ' 20[0-9][0-9] The Chromium Authors.', with optional (c) and date range
-  # - '([Cc]) (19|20)[0-9][0-9]', but not when preceeded by the word copyright,
-  #   as this is handled above
-  regex = '[Cc]opyright(?!( \(c\))? 20[0-9][0-9](-20[0-9][0-9])? ' \
-          'The Chromium Authors\. All rights reserved\.)' \
-          '|' \
-          '(?<!(pyright |opyright))\([Cc]\) (19|20)[0-9][0-9]'
-
-  args = ['grep',
-          '-rPlI',
-          '--exclude', '*.orig',
-          '--exclude', '*.rej',
-          '--exclude-dir', 'third_party',
-          '--exclude-dir', 'out',
-          '--exclude-dir', '.git',
-          '--exclude-dir', '.svn',
-          regex,
+  directory_list = [d for d in directory_list if not 'third_party' in d]
+  # This makes the ignore regexp shorter
+  directory_list.append('third_party')
+  # 'Copyright' appears in license agreements
+  directory_list.append('chrome/app/resources')
+  # Arm sysroot tools, doesn't exist in the snapshot
+  directory_list.append('arm-sysroot')
+
+  ignore_patterns = []
+  for d in directory_list:
+    ignore_patterns.append('(?:/' + d.replace('+', '\+') + '/)')
+
+  args = ['third_party/devscripts/licensecheck.pl',
+          '--lines', '99',
+          '--copyright',
+          '--machine',
+          '--recursive',
+          '--check', '\.(c(c|pp|xx)?|h(h|pp|xx)?|p(l|m)|xs|sh|php|py(|x)|rb' \

[Paweł Hajdan Jr., 2013/01/25 17:54:33]

Do you need to use --check? Why?

[mnaganov (inactive), 2013/01/28 14:51:54]

Because the pattern in licensecheck.pl apparently misses
"js|html|pac|mm|asm|idl"

[Paweł Hajdan Jr., 2013/01/28 16:28:29]

Why not update the pattern in licensecheck.pl then? :)

[mnaganov (inactive), 2013/01/28 17:12:38]

I was considering these additions to be Chromium-specific. Moved to
licensecheck.pl.

+            '|java|vala|el|sc(i|e)|cs|pas|inc|dtd|xsl|mod|m|tex|mli?|js|html' \
+            '|pac|mm|asm|idl)$',
+          '--ignore', '|'.join(ignore_patterns),
           '.']
   p = subprocess.Popen(args=args, cwd=REPOSITORY_ROOT, stdout=subprocess.PIPE)
-  files = p.communicate()[0].splitlines()
-
-  directory_list = directory_list[:]
-  # Ignore these tools.
-  directory_list.append('android_webview/tools/')
-  # This is a build intermediate directory.
-  directory_list.append('chrome/app/theme/google_chrome/')
-  # This is tests directory, doesn't exist in the snapshot
-  directory_list.append('content/test/data/')
-  # This is a test output directory.
-  directory_list.append('data/dom_perf/')
-  # This is a test output directory.
-  directory_list.append('data/page_cycler/')
-  # 'Copyright' appears in strings.
-  directory_list.append('chrome/app/resources/')
-  # This is a Chrome on Linux reference build, doesn't exist in the snapshot
-  directory_list.append('chrome/tools/test/reference_build/chrome_linux/')
-  # Remoting internal tools, doesn't exist in the snapshot
-  directory_list.append('remoting/appengine/')
-  # Histogram tools, doesn't exist in the snapshot
-  directory_list.append('tools/histograms/')
-  # Arm sysroot tools, doesn't exist in the snapshot
-  directory_list.append('arm-sysroot/')
-  # Windows-only
-  directory_list.append('tools/win/toolchain/7z/')
+  lines = p.communicate()[0].splitlines()
 
-  # Exclude files under listed directories and some known offenders.
   offending_files = []
-  for x in files:
-    x = os.path.normpath(x)
-    is_in_listed_directory = False
-    for y in directory_list:
-      if x.startswith(y):
-        is_in_listed_directory = True
+  allowed_copyrights = '^(?:\*No copyright\*' \
+      '|20[0-9][0-9](?:-20[0-9][0-9])? The Chromium Authors\. ' \
+      'All rights reserved.*)$'
+  allowed_copyrights_re = re.compile(allowed_copyrights)
+  for l in lines:
+    entries = l.split('\t')
+    if entries[1] == "GENERATED FILE":
+      continue
+    copyrights = entries[2].split(' / ')
+    for c in copyrights:
+      if c and not allowed_copyrights_re.match(c):
+        offending_files.append(os.path.normpath(entries[0]))
         break
-    if not is_in_listed_directory:
-      offending_files.append(x)
 
   all_files_valid = True
   unknown = set(offending_files) - set(whitelisted_files)