#!/usr/bin/env python
# -*- Mode: python coding: utf-8 -*-
#
# author diverKon <diverKon@gmail.com>
# version 0.2
# Copyright(C) KK.Kon
# licenes http://opensource.org/licenses/bsd-license.php New BSD License
# since 2012-01-13

import os

import sys

from svn import fs, repos, core, client, delta
from svn.core import SubversionException

if ( 2 > len(sys.argv) ):
  print 'Usage: repo_path'
  quit()

do_fix = False
if ( 3 <= len(sys.argv) ):
  if '--do-fix' == sys.argv[2]:
    do_fix = True

print 'do_fix=', do_fix

#print 'stdout.encoding=', sys.stdout.encoding

if 'win32' == sys.platform:
  if 'cp65001' == sys.stdout.encoding:
    print 'Not support codepage 65001!'
    quit()
else:
  import codecs
  sys.stdout = codecs.getwriter('utf_8')(sys.stdout)


myrepos = repos.open( sys.argv[1] )
myfs = repos.fs( myrepos )
myrev = fs.youngest_rev( myfs )

def conv_encoding( data, to_enc="utf_8"):
  lookups = ('utf_8', 'euc_jp', 'cp932', 'ascii' )
  for encoding in lookups:
    try:
      data = data.decode( encoding )
      break
    except:
      pass
  if isinstance( data, unicode ):
    return data.encode( to_enc )
  else:
    return data

for rev in range(1, myrev):
  rev_log = fs.revision_prop( myfs, rev, 'svn:log' )
  print 'revision=', rev
  if None == rev_log:
    continue

  #print 'rev_log=', rev_log
  svnlog_utf8 = conv_encoding( rev_log )

  need_fix = False
  if svnlog_utf8 != rev_log:
    need_fix = True
  else:
    need_fix = False

  if True == need_fix:
    print 'need fix'

  print 'rev_log=', unicode( svnlog_utf8, 'utf_8' )

  if True == do_fix:
    print 'fix...'
    fs.change_rev_prop( myfs, rev, 'svn:log', svnlog_utf8 )