src/etc/tidy.py

   1 # Copyright 2010-2014 The Rust Project Developers. See the COPYRIGHT
   2 # file at the top-level directory of this distribution and at
   3 # http://rust-lang.org/COPYRIGHT.
   4 #
   5 # Licensed under the Apache License, Version 2.0 <LICENSE-APACHE or
   6 # http://www.apache.org/licenses/LICENSE-2.0> or the MIT license
   7 # <LICENSE-MIT or http://opensource.org/licenses/MIT>, at your
   8 # option. This file may not be copied, modified, or distributed
   9 # except according to those terms.
  10
  11 import sys
  12 import fileinput
  13 import subprocess
  14 import re
  15 import os
  16 from licenseck import *
  17 import snapshot
  18
  19 err = 0
  20 cols = 100
  21 cr_flag = "ignore-tidy-cr"
  22 tab_flag = "ignore-tidy-tab"
  23 linelength_flag = "ignore-tidy-linelength"
  24
  25 # Be careful to support Python 2.4, 2.6, and 3.x here!
  26 config_proc = subprocess.Popen(["git", "config", "core.autocrlf"],
  27                                stdout=subprocess.PIPE)
  28 result = config_proc.communicate()[0]
  29
  30 true = "true".encode('utf8')
  31 autocrlf = result.strip() == true if result is not None else False
  32
  33
  34 def report_error_name_no(name, no, s):
  35     global err
  36     print("%s:%d: %s" % (name, no, s))
  37     err = 1
  38
  39
  40 def report_err(s):
  41     report_error_name_no(fileinput.filename(), fileinput.filelineno(), s)
  42
  43
  44 def report_warn(s):
  45     print("%s:%d: %s" % (fileinput.filename(),
  46                          fileinput.filelineno(),
  47                          s))
  48
  49
  50 def do_license_check(name, contents):
  51     if not check_license(name, contents):
  52         report_error_name_no(name, 1, "incorrect license")
  53
  54 current_name = ""
  55 current_contents = ""
  56 check_tab = True
  57 check_cr = True
  58 check_linelength = True
  59
  60 if len(sys.argv) < 2:
  61     print "usage: tidy.py <src-dir>"
  62     sys.exit(1)
  63
  64 src_dir = sys.argv[1]
  65
  66 try:
  67     count_lines = 0
  68     count_non_blank_lines = 0
  69
  70     interesting_files = ['.rs', '.py', '.js', '.sh', '.c', '.h']
  71
  72     file_counts = {ext: 0 for ext in interesting_files}
  73     file_counts['other'] = 0
  74
  75     def update_counts(current_name):
  76         global file_counts
  77         _, ext = os.path.splitext(current_name)
  78
  79         if ext in file_counts:
  80             file_counts[ext] += 1
  81         else:
  82             file_counts['other'] += 1
  83
  84     all_paths = set()
  85
  86     for (dirpath, dirnames, filenames) in os.walk(src_dir):
  87
  88         # Skip some third-party directories
  89         skippable_dirs = {
  90             'src/jemalloc',
  91             'src/llvm',
  92             'src/gyp',
  93             'src/libbacktrace',
  94             'src/libuv',
  95             'src/compiler-rt',
  96             'src/rt/hoedown',
  97             'src/rustllvm',
  98             'src/rt/valgrind',
  99             'src/rt/msvc',
 100             'src/rust-installer'
 101         }
 102
 103         if any(d in dirpath for d in skippable_dirs):
 104             continue
 105
 106         def interesting_file(f):
 107             if "miniz.c" in f \
 108             or "jquery" in f \
 109             or "rust_android_dummy" in f:
 110                 return False
 111
 112             return any(os.path.splitext(f)[1] == ext for ext in interesting_files)
 113
 114         file_names = [os.path.join(dirpath, f) for f in filenames
 115                       if interesting_file(f)
 116                       and not f.endswith("_gen.rs")
 117                       and not ".#" is f]
 118
 119         if not file_names:
 120             continue
 121
 122         for line in fileinput.input(file_names,
 123                                 openhook=fileinput.hook_encoded("utf-8")):
 124
 125             filename = fileinput.filename()
 126
 127             if "tidy.py" not in filename:
 128                 if "TODO" in line:
 129                     report_err("TODO is deprecated; use FIXME")
 130                 match = re.match(r'^.*/(\*|/!?)\s*XXX', line)
 131                 if match:
 132                     report_err("XXX is no longer necessary, use FIXME")
 133                 match = re.match(r'^.*//\s*(NOTE.*)$', line)
 134                 if match and "TRAVIS" not in os.environ:
 135                     m = match.group(1)
 136                     if "snap" in m.lower():
 137                         report_warn(match.group(1))
 138                 match = re.match(r'^.*//\s*SNAP\s+(\w+)', line)
 139                 if match:
 140                     hsh = match.group(1)
 141                     date, rev = snapshot.curr_snapshot_rev()
 142                     if not hsh.startswith(rev):
 143                         report_err("snapshot out of date (" + date
 144                             + "): " + line)
 145                 else:
 146                     if "SNAP" in line:
 147                         report_warn("unmatched SNAP line: " + line)
 148
 149             if cr_flag in line:
 150                 check_cr = False
 151             if tab_flag in line:
 152                 check_tab = False
 153             if linelength_flag in line:
 154                 check_linelength = False
 155
 156             if check_tab and ('\t' in line and
 157                               "Makefile" not in filename):
 158                 report_err("tab character")
 159             if check_cr and not autocrlf and '\r' in line:
 160                 report_err("CR character")
 161             if line.endswith(" \n") or line.endswith("\t\n"):
 162                 report_err("trailing whitespace")
 163             line_len = len(line)-2 if autocrlf else len(line)-1
 164
 165             if check_linelength and line_len > cols:
 166                 report_err("line longer than %d chars" % cols)
 167
 168             if fileinput.isfirstline():
 169                 # This happens at the end of each file except the last.
 170                 if current_name != "":
 171                     update_counts(current_name)
 172                     assert len(current_contents) > 0
 173                     do_license_check(current_name, current_contents)
 174
 175                 current_name = filename
 176                 current_contents = ""
 177                 check_cr = True
 178                 check_tab = True
 179                 check_linelength = True
 180
 181             # Put a reasonable limit on the amount of header data we use for
 182             # the licenseck
 183             if len(current_contents) < 1000:
 184                 current_contents += line
 185
 186             count_lines += 1
 187             if line.strip():
 188                 count_non_blank_lines += 1
 189
 190     if current_name != "":
 191         update_counts(current_name)
 192         assert len(current_contents) > 0
 193         do_license_check(current_name, current_contents)
 194
 195 except UnicodeDecodeError as e:
 196     report_err("UTF-8 decoding error " + str(e))
 197
 198 print
 199 for ext in file_counts:
 200     print "* linted " + str(file_counts[ext]) + " " + ext + " files"
 201 print "* total lines of code: " + str(count_lines)
 202 print "* total non-blank lines of code: " + str(count_non_blank_lines)
 203 print
 204
 205 sys.exit(err)