1#! /usr/bin/python
2
3# Copyright (C) 2009-2011, International Business Machines Corporation, Google and Others.
4# All rights reserved.
5
6#
7#  Script to check and fix svn property settings for ICU source files.
8#  Also check for the correct line endings on files with svn:eol-style = native
9#
10#  THIS SCRIPT DOES NOT WORK ON WINDOWS
11#     It only works correctly on platforms where the native line ending is a plain \n
12#
13#  usage:
14#     icu-svnprops-check.py  [options]
15#
16#  options:
17#     -f | --fix     Fix any problems that are found
18#     -h | --help    Print a usage line and exit.
19#
20#  The tool operates recursively on the directory from which it is run.
21#  Only files from the svn repository are checked.
22#  No changes are made to the repository; only the working copy will be altered.
23
24import sys
25import os
26import os.path
27import re
28import getopt
29
30#
31#  svn autoprops definitions.
32#      Copy and paste here the ICU recommended auto-props from
33#      http://icu-project.org/docs/subversion_howto/index.html
34#
35#  This program will parse this autoprops string, and verify that files in
36#  the repository have the recommeded properties set.
37#
38svn_auto_props = """
39### Section for configuring automatic properties.
40[auto-props]
41### The format of the entries is:
42###   file-name-pattern = propname[=value][;propname[=value]...]
43### The file-name-pattern can contain wildcards (such as '*' and
44### '?').  All entries which match will be applied to the file.
45### Note that auto-props functionality must be enabled, which
46### is typically done by setting the 'enable-auto-props' option.
47*.c = svn:eol-style=native
48*.cc = svn:eol-style=native
49*.cpp = svn:eol-style=native
50*.h = svn:eol-style=native
51*.rc = svn:eol-style=native
52*.dsp = svn:eol-style=native
53*.dsw = svn:eol-style=native
54*.sln = svn:eol-style=native
55*.vcproj = svn:eol-style=native
56configure = svn:eol-style=native;svn:executable
57*.sh = svn:eol-style=native;svn:executable
58*.pl = svn:eol-style=native;svn:executable
59*.py = svn:eol-style=native;svn:executable
60*.txt = svn:mime-type=text/plain;svn:eol-style=native
61*.java = svn:eol-style=native;svn:mime-type=text/plain;;charset=utf-8
62*.ucm = svn:eol-style=native
63*.html = svn:eol-style=native;svn:mime-type=text/html
64*.htm = svn:eol-style=native;svn:mime-type=text/html
65*.xml = svn:eol-style=native
66Makefile = svn:eol-style=native
67*.in = svn:eol-style=native
68*.mak = svn:eol-style=native
69*.mk = svn:eol-style=native
70*.png = svn:mime-type=image/png
71*.jpeg = svn:mime-type=image/jpeg
72*.jpg = svn:mime-type=image/jpeg
73*.bin = svn:mime-type=application/octet-stream
74*.brk = svn:mime-type=application/octet-stream
75*.cnv = svn:mime-type=application/octet-stream
76*.dat = svn:mime-type=application/octet-stream
77*.icu = svn:mime-type=application/octet-stream
78*.res = svn:mime-type=application/octet-stream
79*.spp = svn:mime-type=application/octet-stream
80# new additions 2007-dec-5 srl
81*.rtf = mime-type=text/rtf
82*.pdf = mime-type=application/pdf
83# changed 2008-04-08: modified .txt, above, adding mime-type
84# changed 2010-11-09: modified .java, adding mime-type
85# Note: The escape syntax for semicolon (";;") is supported since subversion 1.6.1
86"""
87
88
89# file_types:  The parsed form of the svn auto-props specification.
90#              A list of file types - .cc, .cpp, .txt, etc.
91#              each element is a [type, proplist]
92#              "type" is a regular expression string that will match a file name
93#              prop list is another list, one element per property.
94#              Each property item is a two element list, [prop name, prop value]
95file_types = list()
96
97def parse_auto_props():
98    aprops = svn_auto_props.splitlines()
99    for propline in aprops:
100        if re.match("\s*(#.*)?$", propline):         # Match comment and blank lines
101            continue
102        if re.match("\s*\[auto-props\]", propline):  # Match the [auto-props] line.
103            continue
104        if not re.match("\s*[^\s]+\s*=", propline):  # minimal syntax check for <file-type> =
105            print "Bad line from autoprops definitions: " + propline
106            continue
107        file_type, string_proplist = propline.split("=", 1)
108
109        #transform the file type expression from autoprops into a normal regular expression.
110        #  e.g.  "*.cpp"  ==>  ".*\.cpp$"
111        file_type = file_type.strip()
112        file_type = file_type.replace(".", "\.")
113        file_type = file_type.replace("*", ".*")
114        file_type = file_type + "$"
115
116        # example string_proplist at this point: " svn:eol-style=native;svn:executable"
117        # split on ';' into a list of properties.  The negative lookahead and lookbehind
118        # in the split regexp are to prevent matching on ';;', which is an escaped ';'
119        # within a property value.
120        string_proplist = re.split("(?<!;);(?!;)", string_proplist)
121        proplist = list()
122        for prop in string_proplist:
123            if prop.find("=") >= 0:
124                prop_name, prop_val = prop.split("=", 1)
125            else:
126                # properties with no explicit value, e.g. svn:executable
127                prop_name, prop_val = prop, ""
128            prop_name = prop_name.strip()
129            prop_val = prop_val.strip()
130            # unescape any ";;" in a property value, e.g. the mime-type from
131            #    *.java = svn:eol-style=native;svn:mime-type=text/plain;;charset=utf-8
132            prop_val = prop_val.replace(";;", ";");
133            proplist.append((prop_name, prop_val))
134
135        file_types.append((file_type, proplist))
136    # print file_types
137
138
139def runCommand(cmd):
140    output_file = os.popen(cmd);
141    output_text = output_file.read();
142    exit_status = output_file.close();
143    if exit_status:
144        print >>sys.stderr, '"', cmd, '" failed.  Exiting.'
145        sys.exit(exit_status)
146    return output_text
147
148
149def usage():
150    print "usage: " + sys.argv[0] + " [-f | --fix] [-h | --help]"
151
152
153#
154#  UTF-8 file check.   For text files, add a charset to the mime-type if their contents are UTF-8
155#    file_name:        name of a text file.
156#    base_mime_type:   svn:mime-type property value from the auto-props file (no charset= part)
157#    actual_mime_type: existing svn:mime-type property value for the file.
158#    return:           svn:mime-type property value, with charset added when appropriate.
159#
160def check_utf8(file_name, base_mime_type, actual_mime_type):
161
162    # If the file already has a charset in its mime-type, don't make any change.
163
164    if actual_mime_type.find("charset=") > 0:
165        return actual_mime_type;
166
167    f = open(file_name, 'r')
168    bytes = f.read()
169    f.close()
170
171    if all(ord(byte) < 128 for byte in bytes):
172        # pure ASCII.
173        # print "Pure ASCII " + file_name
174        return base_mime_type
175
176    try:
177        bytes.decode("UTF-8")
178    except UnicodeDecodeError:
179        print "warning: %s: not ASCII, not UTF-8" % file_name
180        return base_mime_type
181
182    if ord(bytes[0]) != 0xef:
183      print "UTF-8 file with no BOM: " + file_name
184
185    # Append charset=utf-8.
186    return base_mime_type + ';charset=utf-8'
187
188
189def main(argv):
190    fix_problems = False;
191    try:
192        opts, args = getopt.getopt(argv, "fh", ("fix", "help"))
193    except getopt.GetoptError:
194        print "unrecognized option: " + argv[0]
195        usage()
196        sys.exit(2)
197    for opt, arg in opts:
198        if opt in ("-h", "--help"):
199            usage()
200            sys.exit()
201        if opt in ("-f", "--fix"):
202            fix_problems = True
203    if args:
204        print "unexpected command line argument"
205        usage()
206        sys.exit()
207
208    parse_auto_props()
209    output = runCommand("svn ls -R ");
210    file_list = output.splitlines()
211
212    for f in file_list:
213        if os.path.isdir(f):
214            # print "Skipping dir " + f
215            continue
216        if not os.path.isfile(f):
217            print "Repository file not in working copy: " + f
218            continue;
219
220        for file_pattern, props in file_types:
221            if re.match(file_pattern, f):
222                # print "doing " + f
223                for propname, propval in props:
224                    actual_propval = runCommand("svn propget --strict " + propname + " " + f)
225                    #print propname + ": " + actual_propval
226                    if propname == "svn:mime-type" and propval.find("text/") == 0:
227                        # check for UTF-8 text files, should have svn:mime-type=text/something; charset=utf8
228                        propval = check_utf8(f, propval, actual_propval)
229                    if not (propval == actual_propval or (propval == "" and actual_propval == "*")):
230                        print "svn propset %s '%s' %s" % (propname, propval, f)
231                        if fix_problems:
232                            os.system("svn propset %s '%s' %s" % (propname, propval, f))
233                    if propname == "svn:eol-style" and propval == "native":
234                        if os.system("grep -q -v \r " + f):
235                            if fix_problems:
236                                print f + ": Removing DOS CR characters."
237                                os.system("sed -i s/\r// " + f);
238                            else:
239                                print f + " contains DOS CR characters."
240
241
242if __name__ == "__main__":
243    main(sys.argv[1:])
244