chrome/browser/web_dev_style/css_checker.py

   1 # Copyright (c) 2012 The Chromium Authors. All rights reserved.
   2 # Use of this source code is governed by a BSD-style license that can be
   3 # found in the LICENSE file.
   4
   5 """Presubmit script for Chromium WebUI resources.
   6
   7 See http://dev.chromium.org/developers/how-tos/depottools/presubmit-scripts
   8 for more details about the presubmit API built into gcl/git cl, and see
   9 http://www.chromium.org/developers/web-development-style-guide for the rules
  10 we're checking against here.
  11 """
  12
  13 # TODO(dbeam): Real CSS parser? https://github.com/danbeam/css-py/tree/css3
  14
  15 class CSSChecker(object):
  16   def __init__(self, input_api, output_api, file_filter=None):
  17     self.input_api = input_api
  18     self.output_api = output_api
  19     self.file_filter = file_filter
  20
  21   def RunChecks(self):
  22     # We use this a lot, so make a nick name variable.
  23     re = self.input_api.re
  24
  25     def _collapseable_hex(s):
  26       return (len(s) == 6 and s[0] == s[1] and s[2] == s[3] and s[4] == s[5])
  27
  28     def _is_gray(s):
  29       return s[0] == s[1] == s[2] if len(s) == 3 else s[0:2] == s[2:4] == s[4:6]
  30
  31     def _remove_all(s):
  32       return _remove_grit(_remove_ats(_remove_comments(s)))
  33
  34     def _remove_ats(s):
  35       at_reg = re.compile(r"""
  36           @\w+[^'"]*?{  # @at-keyword selector junk {
  37           (.*{.*?})+    # inner { curly } blocks, rules, and selector junk
  38           .*?}          # stuff up to the first end curly }""",
  39           re.DOTALL | re.VERBOSE)
  40       return at_reg.sub('\\1', s)
  41
  42     def _remove_comments(s):
  43       return re.sub(re.compile(r'/\*.*?\*/', re.DOTALL), '', s)
  44
  45     def _remove_grit(s):
  46       grit_reg = re.compile(r"""
  47           <if[^>]+>.*?<\s*/\s*if[^>]*>|  # <if> contents </if>
  48           <include[^>]+>                 # <include>""",
  49           re.DOTALL | re.VERBOSE)
  50       return re.sub(grit_reg, '', s)
  51
  52     def _rgb_from_hex(s):
  53       if len(s) == 3:
  54         r, g, b = s[0] + s[0], s[1] + s[1], s[2] + s[2]
  55       else:
  56         r, g, b = s[0:2], s[2:4], s[4:6]
  57       return int(r, base=16), int(g, base=16), int(b, base=16)
  58
  59     def _strip_prefix(s):
  60       return re.sub(r'^-(?:o|ms|moz|khtml|webkit)-', '', s)
  61
  62     def alphabetize_props(contents):
  63       errors = []
  64       for rule in re.finditer(r'{(.*?)}', contents, re.DOTALL):
  65         semis = map(lambda t: t.strip(), rule.group(1).split(';'))[:-1]
  66         rules = filter(lambda r: ': ' in r, semis)
  67         props = map(lambda r: r[0:r.find(':')], rules)
  68         if props != sorted(props):
  69           errors.append('    %s;\n' % (';\n    '.join(rules)))
  70       return errors
  71
  72     def braces_have_space_before_and_nothing_after(line):
  73       brace_space_reg = re.compile(r"""
  74           (?:^|\S){|  # selector{ or selector\n{ or
  75           {\s*\S+\s*  # selector { with stuff after it
  76           $           # must be at the end of a line""",
  77           re.VERBOSE)
  78       return brace_space_reg.search(line)
  79
  80     def classes_use_dashes(line):
  81       # Intentionally dumbed down version of CSS 2.1 grammar for class without
  82       # non-ASCII, escape chars, or whitespace.
  83       class_reg = re.compile(r"""
  84           \.(-?[\w-]+).*  # ., then maybe -, then alpha numeric and -
  85           [,{]\s*$        # selectors should end with a , or {""",
  86           re.VERBOSE)
  87       m = class_reg.search(line)
  88       if not m:
  89         return False
  90       class_name = m.group(1)
  91       return class_name.lower() != class_name or '_' in class_name
  92
  93     def close_brace_on_new_line(line):
  94       # Ignore single frames in a @keyframe, i.e. 0% { margin: 50px; }
  95       frame_reg = re.compile(r"""
  96           \s*\d+%\s*{       # 50% {
  97           \s*[\w-]+:        # rule:
  98           (\s*[\w-]+)+\s*;  # value;
  99           \s*}\s*           # }""",
 100           re.VERBOSE)
 101       return ('}' in line and re.search(r'[^ }]', line) and
 102               not frame_reg.match(line))
 103
 104     def colons_have_space_after(line):
 105       colon_space_reg = re.compile(r"""
 106           (?<!data)    # ignore data URIs
 107           :(?!//)      # ignore url(http://), etc.
 108           \S[^;]+;\s*  # only catch one-line rules for now""",
 109           re.VERBOSE)
 110       return colon_space_reg.search(line)
 111
 112     def favor_single_quotes(line):
 113       return '"' in line
 114
 115     # Shared between hex_could_be_shorter and rgb_if_not_gray.
 116     hex_reg = re.compile(r"""
 117         \#([a-fA-F0-9]{3}|[a-fA-F0-9]{6})  # pound followed by 3 or 6 hex digits
 118         (?=[^\w-]|$)                       # no more alphanum chars or at EOL
 119         (?!.*(?:{.*|,\s*)$)                # not in a selector""",
 120         re.VERBOSE)
 121
 122     def hex_could_be_shorter(line):
 123       m = hex_reg.search(line)
 124       return (m and _is_gray(m.group(1)) and _collapseable_hex(m.group(1)))
 125
 126     def rgb_if_not_gray(line):
 127       m = hex_reg.search(line)
 128       return (m and not _is_gray(m.group(1)))
 129
 130     small_seconds_reg = re.compile(r"""
 131         (?:^|[^\w-])   # start of a line or a non-alphanumeric char
 132         (0?\.[0-9]+)s  # 1.0s
 133         (?!-?[\w-])    # no following - or alphanumeric chars""",
 134         re.VERBOSE)
 135
 136     def milliseconds_for_small_times(line):
 137       return small_seconds_reg.search(line)
 138
 139     def suggest_ms_from_s(line):
 140       ms = int(float(small_seconds_reg.search(line).group(1)) * 1000)
 141       return ' (replace with %dms)' % ms
 142
 143     def no_data_uris_in_source_files(line):
 144       return re.search(r'\(\s*\'?\s*data:', line)
 145
 146     def one_rule_per_line(line):
 147       one_rule_reg = re.compile(r"""
 148           [\w-](?<!data):  # a rule: but no data URIs
 149           (?!//)[^;]+;     # value; ignoring colons in protocols://
 150           \s*[^ }]\s*      # any non-space after the end colon""",
 151           re.VERBOSE)
 152       return one_rule_reg.search(line)
 153
 154     def pseudo_elements_double_colon(contents):
 155       pseudo_elements = ['after',
 156                          'before',
 157                          'calendar-picker-indicator',
 158                          'color-swatch',
 159                          'color-swatch-wrapper',
 160                          'date-and-time-container',
 161                          'date-and-time-value',
 162                          'datetime-edit',
 163                          'datetime-edit-ampm-field',
 164                          'datetime-edit-day-field',
 165                          'datetime-edit-hour-field',
 166                          'datetime-edit-millisecond-field',
 167                          'datetime-edit-minute-field',
 168                          'datetime-edit-month-field',
 169                          'datetime-edit-second-field',
 170                          'datetime-edit-text',
 171                          'datetime-edit-week-field',
 172                          'datetime-edit-year-field',
 173                          'details-marker',
 174                          'file-upload-button',
 175                          'first-letter',
 176                          'first-line',
 177                          'inner-spin-button',
 178                          'input-placeholder',
 179                          'input-speech-button',
 180                          'keygen-select',
 181                          'media-slider-container',
 182                          'media-slider-thumb',
 183                          'meter-bar',
 184                          'meter-even-less-good-value',
 185                          'meter-inner-element',
 186                          'meter-optimum-value',
 187                          'meter-suboptimum-value',
 188                          'progress-bar',
 189                          'progress-inner-element',
 190                          'progress-value',
 191                          'resizer',
 192                          'scrollbar',
 193                          'scrollbar-button',
 194                          'scrollbar-corner',
 195                          'scrollbar-thumb',
 196                          'scrollbar-track',
 197                          'scrollbar-track-piece',
 198                          'search-cancel-button',
 199                          'search-decoration',
 200                          'search-results-button',
 201                          'search-results-decoration',
 202                          'selection',
 203                          'slider-container',
 204                          'slider-runnable-track',
 205                          'slider-thumb',
 206                          'textfield-decoration-container',
 207                          'validation-bubble',
 208                          'validation-bubble-arrow',
 209                          'validation-bubble-arrow-clipper',
 210                          'validation-bubble-heading',
 211                          'validation-bubble-message',
 212                          'validation-bubble-text-block']
 213       pseudo_reg = re.compile(r"""
 214           (?<!:):       # a single colon, i.e. :after but not ::after
 215           ([a-zA-Z-]+)  # a pseudo element, class, or function
 216           (?=[^{}]+?{)  # make sure a selector, not inside { rules }""",
 217           re.MULTILINE | re.VERBOSE)
 218       errors = []
 219       for p in re.finditer(pseudo_reg, contents):
 220         pseudo = p.group(1).strip().splitlines()[0]
 221         if _strip_prefix(pseudo.lower()) in pseudo_elements:
 222           errors.append('    :%s (should be ::%s)' % (pseudo, pseudo))
 223       return errors
 224
 225     def one_selector_per_line(contents):
 226       any_reg = re.compile(r"""
 227           :(?:-webkit-)?any\(.*?\)  # :-webkit-any(a, b, i) selector""",
 228           re.DOTALL | re.VERBOSE)
 229       multi_sels_reg = re.compile(r"""
 230           (?:}\s*)?            # ignore 0% { blah: blah; }, from @keyframes
 231           ([^,]+,(?=[^{}]+?{)  # selector junk {, not in a { rule }
 232           .*[,{])\s*$          # has to end with , or {""",
 233           re.MULTILINE | re.VERBOSE)
 234       errors = []
 235       for b in re.finditer(multi_sels_reg, re.sub(any_reg, '', contents)):
 236         errors.append('    ' + b.group(1).strip().splitlines()[-1:][0])
 237       return errors
 238
 239     def suggest_rgb_from_hex(line):
 240       suggestions = ['rgb(%d, %d, %d)' % _rgb_from_hex(h.group(1))
 241           for h in re.finditer(hex_reg, line)]
 242       return ' (replace with %s)' % ', '.join(suggestions)
 243
 244     def suggest_short_hex(line):
 245       h = hex_reg.search(line).group(1)
 246       return ' (replace with #%s)' % (h[0] + h[2] + h[4])
 247
 248     webkit_before_or_after_reg = re.compile(r'-webkit-(\w+-)(after|before):')
 249
 250     def suggest_top_or_bottom(line):
 251       prop, pos = webkit_before_or_after_reg.search(line).groups()
 252       top_or_bottom = 'top' if pos == 'before' else 'bottom'
 253       return ' (replace with %s)' % (prop + top_or_bottom)
 254
 255     def webkit_before_or_after(line):
 256       return webkit_before_or_after_reg.search(line)
 257
 258     def zero_length_values(contents):
 259       hsl_reg = re.compile(r"""
 260           hsl\([^\)]*       # hsl(<maybe stuff>
 261           (?:[, ]|(?<=\())  # a comma or space not followed by a (
 262           (?:0?\.?)?0%      # some equivalent to 0%""",
 263           re.VERBOSE)
 264       zeros_reg = re.compile(r"""
 265           ^.*(?:^|[^0-9.])                                 # start/non-number
 266           (?:\.0|0(?:\.0?                                  # .0, 0, or 0.0
 267           |px|em|%|in|cm|mm|pc|pt|ex|deg|g?rad|m?s|k?hz))  # a length unit
 268           (?:\D|$)                                         # non-number/end
 269           (?=[^{}]+?}).*$                                  # only { rules }""",
 270           re.MULTILINE | re.VERBOSE)
 271       errors = []
 272       for z in re.finditer(zeros_reg, contents):
 273         first_line = z.group(0).strip().splitlines()[0]
 274         if not hsl_reg.search(first_line):
 275           errors.append('    ' + first_line)
 276       return errors
 277
 278     # NOTE: Currently multi-line checks don't support 'after'. Instead, add
 279     # suggestions while parsing the file so another pass isn't necessary.
 280     added_or_modified_files_checks = [
 281         { 'desc': 'Alphabetize properties and list vendor specific (i.e. '
 282                   '-webkit) above standard.',
 283           'test': alphabetize_props,
 284           'multiline': True,
 285         },
 286         { 'desc': 'Start braces ({) end a selector, have a space before them '
 287                   'and no rules after.',
 288           'test': braces_have_space_before_and_nothing_after,
 289         },
 290         { 'desc': 'Classes use .dash-form.',
 291           'test': classes_use_dashes,
 292         },
 293         { 'desc': 'Always put a rule closing brace (}) on a new line.',
 294           'test': close_brace_on_new_line,
 295         },
 296         { 'desc': 'Colons (:) should have a space after them.',
 297           'test': colons_have_space_after,
 298         },
 299         { 'desc': 'Use single quotes (\') instead of double quotes (") in '
 300                   'strings.',
 301           'test': favor_single_quotes,
 302         },
 303         { 'desc': 'Use abbreviated hex (#rgb) when in form #rrggbb.',
 304           'test': hex_could_be_shorter,
 305           'after': suggest_short_hex,
 306         },
 307         { 'desc': 'Use milliseconds for time measurements under 1 second.',
 308           'test': milliseconds_for_small_times,
 309           'after': suggest_ms_from_s,
 310         },
 311         { 'desc': "Don't use data URIs in source files. Use grit instead.",
 312           'test': no_data_uris_in_source_files,
 313         },
 314         { 'desc': 'One rule per line (what not to do: color: red; margin: 0;).',
 315           'test': one_rule_per_line,
 316         },
 317         { 'desc': 'One selector per line (what not to do: a, b {}).',
 318           'test': one_selector_per_line,
 319           'multiline': True,
 320         },
 321         { 'desc': 'Pseudo-elements should use double colon (i.e. ::after).',
 322           'test': pseudo_elements_double_colon,
 323           'multiline': True,
 324         },
 325         { 'desc': 'Use rgb() over #hex when not a shade of gray (like #333).',
 326           'test': rgb_if_not_gray,
 327           'after': suggest_rgb_from_hex,
 328         },
 329         { 'desc': 'Use *-top/bottom instead of -webkit-*-before/after.',
 330           'test': webkit_before_or_after,
 331           'after': suggest_top_or_bottom,
 332         },
 333         { 'desc': 'Make all zero length terms (i.e. 0px) 0 unless inside of '
 334                   'hsl() or part of @keyframe.',
 335           'test': zero_length_values,
 336           'multiline': True,
 337         },
 338     ]
 339
 340     results = []
 341     affected_files = self.input_api.AffectedFiles(include_deletes=False,
 342                                                   file_filter=self.file_filter)
 343     files = []
 344     for f in affected_files:
 345       # Remove all /*comments*/, @at-keywords, and grit <if|include> tags; we're
 346       # not using a real parser. TODO(dbeam): Check alpha in <if> blocks.
 347       file_contents = _remove_all('\n'.join(f.NewContents()))
 348       files.append((f.LocalPath(), file_contents))
 349
 350     # Only look at CSS files for now.
 351     for f in filter(lambda f: f[0].endswith('.css'), files):
 352       file_errors = []
 353       for check in added_or_modified_files_checks:
 354         # If the check is multiline, it receieves the whole file and gives us
 355         # back a list of things wrong. If the check isn't multiline, we pass it
 356         # each line and the check returns something truthy if there's an issue.
 357         if ('multiline' in check and check['multiline']):
 358           assert not 'after' in check
 359           check_errors = check['test'](f[1])
 360           if len(check_errors) > 0:
 361             file_errors.append('- %s\n%s' %
 362                 (check['desc'], '\n'.join(check_errors).rstrip()))
 363         else:
 364           check_errors = []
 365           lines = f[1].splitlines()
 366           for lnum, line in enumerate(lines):
 367             if check['test'](line):
 368               error = '    ' + line.strip()
 369               if 'after' in check:
 370                 error += check['after'](line)
 371               check_errors.append(error)
 372           if len(check_errors) > 0:
 373             file_errors.append('- %s\n%s' %
 374                 (check['desc'], '\n'.join(check_errors)))
 375       if file_errors:
 376         results.append(self.output_api.PresubmitPromptWarning(
 377             '%s:\n%s' % (f[0], '\n\n'.join(file_errors))))
 378
 379     if results:
 380       # Add your name if you're here often mucking around in the code.
 381       authors = ['dbeam@chromium.org']
 382       results.append(self.output_api.PresubmitNotifyResult(
 383           'Was the CSS checker useful? Send feedback or hate mail to %s.' %
 384           ', '.join(authors)))
 385
 386     return results