[chromedriver] Fix mobile device embedding script

chrome/test/chromedriver/cpp_source.py

 # Copyright 2013 The Chromium Authors. All rights reserved.
 # Use of this source code is governed by a BSD-style license that can be
 # found in the LICENSE file.
 
 """Writes C++ header/cc source files for embedding resources into C++."""
 
 import datetime
 import os
 
+# 80-character max line length, minus 4 spaces, and 2 quotations.
+MAX_LINE_LEN = 74
+
 
 def WriteSource(base_name,
                 dir_from_src,
                 output_dir,
                 global_string_map):
   """Writes C++ header/cc source files for the given map of string variables.
 
   Args:
       base_name: The basename of the file, without the extension.
       dir_from_src: Path from src to the directory that will contain the file,
           using forward slashes.
       output_dir: Directory to output the sources to.
       global_string_map: Map of variable names to their string values. These
           variables will be available as globals.
   """
   copyright = '\n'.join([
       '// Copyright %s The Chromium Authors. All rights reserved.',
       '// Use of this source code is governed by a BSD-style license that '
           'can be',
       '// found in the LICENSE file.']) % datetime.date.today().year

[samuong, 2014/12/11 00:22:19]

Now that we're kicking off the generation of (some of) these files manually, I
think we should have a comment explaining that these are generated files and how
to regenerate them. Maybe something like:

// This file was generated at (datetime.datetime()) by running ('
'.join(sys.argv))

That would make it easier for future chromedriver devs to regenerate the mobile
device list if they find it is out of date.

[srawlins, 2014/12/12 23:41:28]

Done.

 
   # Write header file.
   externs = []
   for name in global_string_map.iterkeys():
     externs += ['extern const char %s[];' % name]
 
   temp = '_'.join(dir_from_src.split('/') + [base_name])
   define = temp.upper() + '_H_'
   header = '\n'.join([
       copyright,
@@ skipping 10 matching lines @@
     f.write(header)
 
   # Write cc file.
   def EscapeLine(line):
     return line.replace('\\', '\\\\').replace('"', '\\"')
 
   definitions = []
   for name, contents in global_string_map.iteritems():
     lines = []
     if '\n' not in contents:
-      lines = ['    "%s"' % EscapeLine(contents)]
+
+      contents = EscapeLine(contents)
+      while len(contents) > MAX_LINE_LEN:
+        if contents[MAX_LINE_LEN-1] == '\\':
+          lines.append('    "%s"' % contents[:MAX_LINE_LEN-1])
+          contents = contents[MAX_LINE_LEN-1:]
+        else:
+          lines.append('    "%s"' % contents[:MAX_LINE_LEN])
+          contents = contents[MAX_LINE_LEN:]
+
+      lines.append('    "%s"' % contents)

[samuong, 2014/12/11 00:22:19]

Maybe an easier way to do this is to just pump everything through clang-format?
This doesn't even really need to be automated. People could just do:

$ chrome/test/chromedriver/embed_mobile_devices_in_cpp.py
--directory=chrome/test/chromedriver/chrome/
third_party/WebKit/Source/devtools/front_end/toolbox/OverridesUI.js
$ clang-format -i chrome/test/chromedriver/chrome/mobile_device_list.cc

The nice thing about clang-format is that it breaks the lines between words,
which makes the generated code a bit easier to read, and also ensures that there
are no syntax errors in the generated file.

[srawlins, 2014/12/12 23:41:29]

Done.

     else:
       for line in contents.split('\n'):
         lines += ['    "%s\\n"' % EscapeLine(line)]
     definitions += ['const char %s[] =\n%s;' % (name, '\n'.join(lines))]
 
   cc = '\n'.join([
       copyright,
       '',
       '#include "%s"' % (dir_from_src + '/' + base_name + '.h'),
       '',
       '\n'.join(definitions)])
   cc += '\n'
 
   with open(os.path.join(output_dir, base_name + '.cc'), 'w') as f:
     f.write(cc)