# HG changeset patch
# User William Roberts <williamr@symbian.org>
# Date 1265719802 0
# Node ID 753418a2eb15b127dbb7a1425d98d2165ecaecf4
# Parent  b6c06a8333fba35acf5bb6d828e6cd6d8d787c0d
Switch convert_to_epl from Perl to Python, adding unicode file support and a "--check" option

diff -r b6c06a8333fb -r 753418a2eb15 williamr/convert_to_epl.pl
--- a/williamr/convert_to_epl.pl	Fri Feb 05 14:53:29 2010 +0000
+++ /dev/null	Thu Jan 01 00:00:00 1970 +0000
@@ -1,163 +0,0 @@
-#!/usr/bin/perl
-
-# Copyright (c) 2009 Symbian Foundation Ltd
-# This component and the accompanying materials are made available
-# under the terms of the License "Eclipse Public License v1.0"
-# which accompanies this distribution, and is available
-# at the URL "http://www.eclipse.org/legal/epl-v10.html".
-#
-# Initial Contributors:
-# Symbian Foundation Ltd - initial contribution.
-# 
-# Contributors:
-#
-# Description:
-# Map the SFL license to the EPL license, keeping a copy of the original file
-# in a parallel tree 
-
-use strict;
-use File::Copy;
-use File::Path;
-
-if (scalar @ARGV != 2)
-  {
-	print <<'EOF';
-Incorrect number of arguments
-
-Usage: perl convert_to_epl.pl workdir savedir
-
-Recursively processes workdir to examine all of the text files and convert
-all perfectly formed instances of the SFL copyright notice into EPL notices.
-
-If a file is modified, the original is first copied to the corresponding place
-under savedir. 
-
-It is safe to rerun this script if it stopped for any reason, as no converted 
-SFL notice will ever match on the second run through.
-EOF
-  exit 1;
-  }
-
-my $work_root = $ARGV[0];
-my $saved_root = $ARGV[1];
-
-$work_root =~ s/\\/\//g;    # convert to Unix separators please
-$saved_root =~ s/\\/\//g;
-
-print "* Processing $work_root, leaving the original of any modified file in $saved_root\n";
-
-my $debug = 0;
-
-my @oldtext = (
-  'terms of the License "Symbian Foundation License v1.0"',
-  'the URL "http://www.symbianfoundation.org/legal/sfl-v10.html"'
-);
-my @newtext = (
-  'terms of the License "Eclipse Public License v1.0"',
-  'the URL "http://www.eclipse.org/legal/epl-v10.html"'
-);
-
-my @errorfiles = ();
-my @multinoticefiles = ();
-
-sub map_epl($$$)
-  {
-  my ($file,$shadowdir,$name) = @_;
-  
-  open FILE, "<$file" or print "ERROR: Cannot open $file: $!\n" and return "Cannot open";
-  my @lines = <FILE>;
-  close FILE;
-  
-  my $updated = 0;
-  my @newlines = ();
-  while (my $line = shift @lines)
-    { 
-    # under the terms of the License "Symbian Foundation License v1.0"
-    # which accompanies this distribution, and is available
-    # at the URL "http://www.symbianfoundation.org/legal/sfl-v10.html".
-    my $pos1 = index $line, $oldtext[0];
-    if ($pos1 >= 0)
-      {
-      # be careful - oldtext is a prefix of newtext!
-      if (index($line, $newtext[0]) >= 0)
-        {
-        # line already converted - nothing to do
-        push @newlines, $line;
-        next;
-        }
-      my $midline = shift @lines;
-      my $urlline = shift @lines;
-      my $pos2 = index $urlline, $oldtext[1];
-      if ($pos2 >= 0)
-        {
-        # Found it - assume that there's only one instance
-        substr $line, $pos1, length($oldtext[0]), $newtext[0];
-        substr $urlline, $pos2, length($oldtext[1]), $newtext[1];
-        push @newlines, $line, $midline, $urlline;
-        $updated += 1;
-        next;
-        }
-      else
-        {
-        if(!$updated)
-          {
-          my $lineno = 1 + (scalar @newlines);
-          print STDERR "Problem in $file at $lineno: incorrectly formatted >\n$line$midline$urlline\n";
-          push @errorfiles, $file;
-          }	
-        last;
-        }
-      }
-    push @newlines, $line;
-    }
-
-  return if (!$updated);
-  
-  if ($updated > 1)
-    {
-    push @multinoticefiles, $file;
-    print "! found $updated SFL notices in $file\n";
-    }
- 
-  mkpath($shadowdir, {verbose=>0});
-  move($file, "$shadowdir/$name") or die("Cannot move $file to $shadowdir/$name: $!\n");
-  open NEWFILE, ">$file" or die("Cannot overwrite $file: $!\n");
-  print NEWFILE @newlines, @lines;
-  close NEWFILE or die("Failed to update $file: $!\n");
-  print "* updated $file\n";
-  }
-
-# Process tree
-
-sub scan_directory($$)
-  {
-  my ($path, $shadow) = @_;
-  
-  opendir DIR, $path;
-  my @files = grep !/^\.\.?$/, readdir DIR;
-  closedir DIR;
-  
-  foreach my $file (@files)
-    {
-    my $newpath = "$path/$file";
-    my $newshadow = "$shadow/$file";
-    
-    if (-d $newpath)
-      {
-      scan_directory($newpath, $newshadow);
-      next;
-      }
-    next if (-B $newpath);  # ignore binary files
-    
-    map_epl($newpath, $shadow, $file);
-    }
-  }
-
-scan_directory($work_root, $saved_root);
-
-printf "%d problem files\n", scalar @errorfiles;
-print "\t", join("\n\t", @errorfiles), "\n";
-
-printf "%d files with multiple notices\n", scalar @multinoticefiles;
-print "\t", join("\n\t", @multinoticefiles), "\n";
-
diff -r b6c06a8333fb -r 753418a2eb15 williamr/convert_to_epl.py
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/williamr/convert_to_epl.py	Tue Feb 09 12:50:02 2010 +0000
@@ -0,0 +1,160 @@
+#!/usr/bin/python
+# Copyright (c) 2009 Symbian Foundation.
+# All rights reserved.
+# This component and the accompanying materials are made available
+# under the terms of the License "Eclipse Public License v1.0"
+# which accompanies this distribution, and is available
+# at the URL "http://www.eclipse.org/legal/epl-v10.html".
+#
+# Initial Contributors:
+# Symbian Foundation - Initial contribution
+# 
+# Description:
+# Map the SFL license to the EPL license
+
+import os
+import os.path
+import re
+import codecs
+from optparse import OptionParser
+import sys
+
+oldtext0 = re.compile('terms of the License "Symbian Foundation License v1.0"(to Symbian Foundation)?')
+oldtext1 = re.compile('the URL "http:..www.symbianfoundation.org/legal/sfl-v10.html"')
+
+newtext = [
+  'terms of the License "Eclipse Public License v1.0"',
+  'the URL "http://www.eclipse.org/legal/epl-v10.html"'
+]
+
+errorfiles = []
+multinoticefiles = []
+shadowroot = 'shadow_epoc32'
+
+def file_type(file) :
+	f = open(file, 'r')
+	data = f.read(256)
+	f.close()
+	if len(data) < 2:
+		return None # too short to be worth bothering about anyway
+	if data[0] == chr(255) and data[1] == chr(254) :
+		return 'utf_16_le'
+	if data.find(chr(0)) >= 0 : 
+		return None	# zero byte implies binary file
+	return 'text'
+	
+def map_eula(dir, name, encoded) :
+	global oldtext0
+	global newtext1
+	global newtext
+	file = os.path.join(dir, name)
+	if encoded == 'text':
+		f = open(file, 'r')
+	else:
+		f = codecs.open(file, 'r', encoding=encoded)
+	lines = f.readlines()
+	# print ">> %s encoded as %s" % (file, f.encoding)
+	f.close()
+	
+	updated = 0
+	newlines = []
+	while len(lines) > 0:
+		line = lines.pop(0)
+		pos1 = oldtext0.search(line)
+		if pos1 != None:
+			# be careful - oldtext is a prefix of newtext
+			if pos1.group(1) != None:
+				# line already converted - nothing to do
+				newlines.append(line)
+				continue
+			midlines = []
+			midlinecount = 1
+			while len(lines) > 0:
+				nextline = lines.pop(0)
+				if not re.match('^\s$', nextline):
+					# non-blank line
+					if midlinecount == 0:
+						break
+					midlinecount -= 1
+				midlines.append(nextline)
+			urlline = nextline
+			pos2 = oldtext1.search(urlline)
+			if pos2 != None:
+				# found it - assume that there's only one instance
+				newline = oldtext0.sub(newtext[0], line)
+				newurl  = oldtext1.sub(newtext[1], urlline)
+				newlines.append(newline)
+				newlines.extend(midlines)
+				newlines.append(newurl)
+				updated += 1
+				continue
+			else:
+			  if updated != 0:
+			  	lineno = 1 + len(newlines)
+			  	print "Problem in " + file + " at " + lineno + ": incorrectly formatted >"
+			  	print line
+			  	print midlines
+			  	print urlline
+			  	global errorfiles
+			  	errorfiles.append(file)
+			  break
+		newlines.append(line)
+	
+	if updated == 0:
+		# print " = no change to " + file
+		return 0
+	
+	if updated > 1:
+	  global multinoticefiles
+	  multinoticefiles.append(file)
+	  print '! found %d SFL notices in %s' % (updated, file)
+	
+	# global shadowroot
+	# shadowdir = os.path.join(shadowroot, dir)
+	# if not os.path.exists(shadowdir) :
+	# 	os.makedirs(shadowdir)
+	# newfile = os.path.join(shadowroot,file)
+	# os.rename(file, newfile)
+	
+	global options
+	if not options.dryrun:
+		if encoded == 'text':
+			f = open(file, 'w')
+		else:
+			f = codecs.open(file, 'w', encoding=encoded)
+		f.writelines(newlines)
+		f.close()
+	print "* updated %s (encoding %s)" % (file, encoded)
+	return 1
+
+parser = OptionParser(version="%prog 0.2", usage="Usage: %prog [options]")
+parser.add_option("-n", "--check", action="store_true", dest="dryrun",
+	help="report the files which would be updated, but don't change anything")
+parser.set_defaults(dryrun=False)
+
+(options, args) = parser.parse_args()
+if len(args) != 0:
+	parser.error("Unexpected commandline arguments")
+
+# process tree
+
+update_count = 0
+for root, dirs, files in os.walk('.', topdown=True):
+	if '.hg' in dirs:
+			dirs.remove('.hg') # don't recurse into the Mercurial repository storage
+	for name in files:
+		encoding = file_type(os.path.join(root, name))
+		if encoding:
+			update_count += map_eula(root, name, encoding)
+	
+print '%d problem files' % len(errorfiles)
+print errorfiles
+
+print '%d files with multiple notices' % len(multinoticefiles)
+print multinoticefiles
+
+if options.dryrun and update_count > 0:
+	print "%d files need updating" % update_count
+	sys.exit(1)
+	
+