1 # Copyright 2015 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.
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.
12 htmldocck.py is a custom checker script for Rustdoc HTML outputs.
16 The principle is simple: This script receives a path to generated HTML
17 documentation and a "template" script, which has a series of check
18 commands like `@has` or `@matches`. Each command can be used to check if
19 some pattern is present or not present in the particular file or in
20 the particular node of HTML tree. In many cases, the template script
21 happens to be a source code given to rustdoc.
23 While it indeed is possible to test in smaller portions, it has been
24 hard to construct tests in this fashion and major rendering errors were
25 discovered much later. This script is designed for making the black-box
26 and regression testing of Rustdoc easy. This does not preclude the needs
27 for unit testing, but can be used to complement related tests by quickly
28 showing the expected renderings.
30 In order to avoid one-off dependencies for this task, this script uses
31 a reasonably working HTML parser and the existing XPath implementation
32 from Python 2's standard library. Hopefully we won't render
37 Commands start with an `@` followed by a command name (letters and
38 hyphens), and zero or more arguments separated by one or more whitespace
39 and optionally delimited with single or double quotes. The `@` mark
40 cannot be preceded by a non-whitespace character. Other lines (including
41 every text up to the first `@`) are ignored, but it is recommended to
42 avoid the use of `@` in the template file.
44 There are a number of supported commands:
46 * `@has PATH` checks for the existence of given file.
48 `PATH` is relative to the output directory. It can be given as `-`
49 which repeats the most recently used `PATH`.
51 * `@has PATH PATTERN` and `@matches PATH PATTERN` checks for
52 the occurrence of given `PATTERN` in the given file. Only one
53 occurrence of given pattern is enough.
55 For `@has`, `PATTERN` is a whitespace-normalized (every consecutive
56 whitespace being replaced by one single space character) string.
57 The entire file is also whitespace-normalized including newlines.
59 For `@matches`, `PATTERN` is a Python-supported regular expression.
60 The file remains intact but the regexp is matched with no `MULTILINE`
61 and `IGNORECASE` option. You can still use a prefix `(?m)` or `(?i)`
62 to override them, and `\A` and `\Z` for definitely matching
63 the beginning and end of the file.
65 (The same distinction goes to other variants of these commands.)
67 * `@has PATH XPATH PATTERN` and `@matches PATH XPATH PATTERN` checks for
68 the presence of given `XPATH` in the given HTML file, and also
69 the occurrence of given `PATTERN` in the matching node or attribute.
70 Only one occurrence of given pattern in the match is enough.
72 `PATH` should be a valid and well-formed HTML file. It does *not*
73 accept arbitrary HTML5; it should have matching open and close tags
74 and correct entity references at least.
76 `XPATH` is an XPath expression to match. This is fairly limited:
77 `tag`, `*`, `.`, `//`, `..`, `[@attr]`, `[@attr='value']`, `[tag]`,
78 `[POS]` (element located in given `POS`), `[last()-POS]`, `text()`
79 and `@attr` (both as the last segment) are supported. Some examples:
81 - `//pre` or `.//pre` matches any element with a name `pre`.
82 - `//a[@href]` matches any element with an `href` attribute.
83 - `//*[@class="impl"]//code` matches any element with a name `code`,
84 which is an ancestor of some element which `class` attr is `impl`.
85 - `//h1[@class="fqn"]/span[1]/a[last()]/@class` matches a value of
86 `class` attribute in the last `a` element (can be followed by more
87 elements that are not `a`) inside the first `span` in the `h1` with
88 a class of `fqn`. Note that there cannot be no additional elements
89 between them due to the use of `/` instead of `//`.
91 Do not try to use non-absolute paths, it won't work due to the flawed
92 ElementTree implementation. The script rejects them.
94 For the text matches (i.e. paths not ending with `@attr`), any
95 subelements are flattened into one string; this is handy for ignoring
96 highlights for example. If you want to simply check the presence of
97 given node or attribute, use an empty string (`""`) as a `PATTERN`.
99 * `@count PATH XPATH COUNT' checks for the occurrence of given XPath
100 in the given file. The number of occurrences must match the given count.
102 All conditions can be negated with `!`. `@!has foo/type.NoSuch.html`
103 checks if the given file does not exist, for example.
111 from collections import namedtuple
112 from HTMLParser import HTMLParser
113 from xml.etree import cElementTree as ET
115 # ⇤/⇥ are not in HTML 4 but are in HTML 5
116 from htmlentitydefs import entitydefs
117 entitydefs['larrb'] = u'\u21e4'
118 entitydefs['rarrb'] = u'\u21e5'
120 # "void elements" (no closing tag) from the HTML Standard section 12.1.2
121 VOID_ELEMENTS = set(['area', 'base', 'br', 'col', 'embed', 'hr', 'img', 'input', 'keygen',
122 'link', 'menuitem', 'meta', 'param', 'source', 'track', 'wbr'])
125 class CustomHTMLParser(HTMLParser):
126 """simplified HTML parser.
128 this is possible because we are dealing with very regular HTML from
129 rustdoc; we only have to deal with i) void elements and ii) empty
131 def __init__(self, target=None):
132 HTMLParser.__init__(self)
133 self.__builder = target or ET.TreeBuilder()
135 def handle_starttag(self, tag, attrs):
136 attrs = dict((k, v or '') for k, v in attrs)
137 self.__builder.start(tag, attrs)
138 if tag in VOID_ELEMENTS:
139 self.__builder.end(tag)
141 def handle_endtag(self, tag):
142 self.__builder.end(tag)
144 def handle_startendtag(self, tag, attrs):
145 attrs = dict((k, v or '') for k, v in attrs)
146 self.__builder.start(tag, attrs)
147 self.__builder.end(tag)
149 def handle_data(self, data):
150 self.__builder.data(data)
152 def handle_entityref(self, name):
153 self.__builder.data(entitydefs[name])
155 def handle_charref(self, name):
156 code = int(name[1:], 16) if name.startswith(('x', 'X')) else int(name, 10)
157 self.__builder.data(unichr(code).encode('utf-8'))
160 HTMLParser.close(self)
161 return self.__builder.close()
163 Command = namedtuple('Command', 'negated cmd args lineno')
166 def concat_multi_lines(f):
167 """returns a generator out of the file object, which
168 - removes `\\` then `\n` then a shared prefix with the previous line then
170 - keeps a line number (starting from 0) of the first line being
172 lastline = None # set to the last line when the last line has a backslash
175 for lineno, line in enumerate(f):
176 line = line.rstrip('\r\n')
178 # strip the common prefix from the current line if needed
179 if lastline is not None:
181 for i in xrange(min(len(line), len(lastline))):
182 if line[i] != lastline[i]:
185 line = line[maxprefix:].lstrip()
187 firstlineno = firstlineno or lineno
188 if line.endswith('\\'):
191 catenated += line[:-1]
193 yield firstlineno, catenated + line
198 if lastline is not None:
199 raise RuntimeError('Trailing backslash in the end of file')
201 LINE_PATTERN = re.compile(r'''
202 (?<=(?<!\S)@)(?P<negated>!?)
203 (?P<cmd>[A-Za-z]+(?:-[A-Za-z]+)*)
208 def get_commands(template):
209 with open(template, 'rUb') as f:
210 for lineno, line in concat_multi_lines(f):
211 m = LINE_PATTERN.search(line)
215 negated = (m.group('negated') == '!')
217 args = m.group('args')
218 if args and not args[:1].isspace():
219 raise RuntimeError('Invalid template syntax at line {}'.format(lineno+1))
220 args = shlex.split(args)
221 yield Command(negated=negated, cmd=cmd, args=args, lineno=lineno+1)
224 def _flatten(node, acc):
226 acc.append(node.text)
239 def normalize_xpath(path):
240 if path.startswith('//'):
241 return '.' + path # avoid warnings
242 elif path.startswith('.//'):
245 raise RuntimeError('Non-absolute XPath is not supported due to \
246 the implementation issue.')
249 class CachedFiles(object):
250 def __init__(self, root):
254 self.last_path = None
256 def resolve_path(self, path):
258 path = os.path.normpath(path)
259 self.last_path = path
261 elif self.last_path is None:
262 raise RuntimeError('Tried to use the previous path in the first command')
264 return self.last_path
266 def get_file(self, path):
267 path = self.resolve_path(path)
269 return self.files[path]
272 with open(os.path.join(self.root, path)) as f:
274 except Exception as e:
275 raise RuntimeError('Cannot open file {!r}: {}'.format(path, e))
277 self.files[path] = data
280 def get_tree(self, path):
281 path = self.resolve_path(path)
283 return self.trees[path]
286 f = open(os.path.join(self.root, path))
287 except Exception as e:
288 raise RuntimeError('Cannot open file {!r}: {}'.format(path, e))
291 tree = ET.parse(f, CustomHTMLParser())
292 except Exception as e:
293 raise RuntimeError('Cannot parse an HTML file {!r}: {}'.format(path, e))
295 self.trees[path] = tree
296 return self.trees[path]
299 def check_string(data, pat, regexp):
301 return True # special case a presence testing
303 return re.search(pat, data) is not None
305 data = ' '.join(data.split())
306 pat = ' '.join(pat.split())
310 def check_tree_attr(tree, path, attr, pat, regexp):
311 path = normalize_xpath(path)
313 for e in tree.findall(path):
315 value = e.attrib[attr]
319 ret = check_string(value, pat, regexp)
325 def check_tree_text(tree, path, pat, regexp):
326 path = normalize_xpath(path)
328 for e in tree.findall(path):
334 ret = check_string(value, pat, regexp)
340 def check_tree_count(tree, path, count):
341 path = normalize_xpath(path)
342 return len(tree.findall(path)) == count
345 def check(target, commands):
346 cache = CachedFiles(target)
348 if c.cmd == 'has' or c.cmd == 'matches': # string test
349 regexp = (c.cmd == 'matches')
350 if len(c.args) == 1 and not regexp: # @has <path> = file existence
352 cache.get_file(c.args[0])
356 elif len(c.args) == 2: # @has/matches <path> <pat> = string test
357 ret = check_string(cache.get_file(c.args[0]), c.args[1], regexp)
358 elif len(c.args) == 3: # @has/matches <path> <pat> <match> = XML tree test
359 tree = cache.get_tree(c.args[0])
360 pat, sep, attr = c.args[1].partition('/@')
362 ret = check_tree_attr(cache.get_tree(c.args[0]), pat, attr, c.args[2], regexp)
363 else: # normalized text
365 if pat.endswith('/text()'):
367 ret = check_tree_text(cache.get_tree(c.args[0]), pat, c.args[2], regexp)
369 raise RuntimeError('Invalid number of @{} arguments \
370 at line {}'.format(c.cmd, c.lineno))
372 elif c.cmd == 'count': # count test
373 if len(c.args) == 3: # @count <path> <pat> <count> = count test
374 ret = check_tree_count(cache.get_tree(c.args[0]), c.args[1], int(c.args[2]))
376 raise RuntimeError('Invalid number of @{} arguments \
377 at line {}'.format(c.cmd, c.lineno))
379 elif c.cmd == 'valid-html':
380 raise RuntimeError('Unimplemented @valid-html at line {}'.format(c.lineno))
382 elif c.cmd == 'valid-links':
383 raise RuntimeError('Unimplemented @valid-links at line {}'.format(c.lineno))
386 raise RuntimeError('Unrecognized @{} at line {}'.format(c.cmd, c.lineno))
389 raise RuntimeError('@{}{} check failed at line {}'.format('!' if c.negated else '',
392 if __name__ == '__main__':
393 if len(sys.argv) < 3:
394 print >>sys.stderr, 'Usage: {} <doc dir> <template>'.format(sys.argv[0])
397 check(sys.argv[1], get_commands(sys.argv[2]))