element is not legal in XHTML 1.0 Transitional."
% element)
return
# Check that the element may legally appear at this point in the
# document.
if (parent and legal_elements.has_key(parent)
and element not in legal_elements[parent]):
self.err(2, "Element <%s> appears in <%s> (not allowed)."
% (element, parent))
if not parent and element != 'html':
self.err(3, "Top-level element is <%s> (should be )."
% element)
if illegal_ancestors.has_key(element):
for (ancestor, line) in self.stack2[0:-1]:
if ancestor in illegal_ancestors[element]:
self.err(16, "Element <%s> appears below <%s> on line %d "
"(not allowed)." % (element, ancestor, line))
# Check attributes.
for (attr, value) in attrs.items():
# Check that the attributes is legal for the element.
if not legal_attributes[element].has_key(attr):
self.err(4, "Attribute '%s' is not allowed in <%s>."
% (attr, element))
continue
# Check that the value for the attribute is legal.
type, disposition = legal_attributes[element][attr]
if isinstance(type, types.ListType):
if value not in type:
self.err(10, "Attribute '%s' for element <%s> has value "
"'%s' (must be one of [%s])."
% (attr, element, value,
string.join(type, ', ')))
elif attribute_checkers.has_key(type):
if not attribute_checkers[type](value):
self.err(11, "Attribute '%s' for element <%s> has illegal "
"value '%s' (not a %s)."
% (attr, element, value, type))
# Check that id is unique (ignoring case) within the document.
if attr == 'id':
if self.nocase_ids.has_key(string.lower(value)):
self.err(12, "Duplicate id '%s' (original on line %d)."
% (value, self.nocase_ids[string.lower(value)]))
self.ids[value] = 1
self.nocase_ids[string.lower(value)] = self.locator.getLineNumber()
# Remember IDREF and IDREFS for checking later.
if type in ['IDREF', 'IDREFS']:
idrefs = re.split(" +", value)
for i in idrefs:
self.idrefs.append((i, self.locator.getLineNumber()))
# Check that required attributes are present.
for (attr, (type, disposition)) in legal_attributes[element].items():
if disposition == 1 and attr not in attrs.keys():
self.err(13, "Attribute '%s' is required for <%s> but not "
"present." % (attr, element))
# Check that recommended attributes are present.
if recommended_attributes.has_key(element):
for a in recommended_attributes[element]:
if not attrs.has_key(a):
self.err(26, "Attribute '%s' is recommended for <%s> but "
"not present." % (a, element))
# | is deprecated: should set same valign for all cells
# in the row using | .
if element in ['td','th'] and attrs.has_key('valign'):
self.err(20, "<%s> has valign attribute: better in the
."
% element)
# Test against rule xhtml/id.
if element == 'a' and attrs.has_key('id') and not attrs.has_key('name'):
self.err(5, "<%s> element has 'id' attribute but no 'name' "
"attribute." % element)
if element == 'a' and attrs.has_key('name') and not attrs.has_key('id'):
self.err(6, "<%s> element has 'name' attribute but no 'id' "
"attribute." % element)
if (attrs.has_key('name') and attrs.has_key('id')
and attrs['name'] != attrs['id']):
self.err(7, "<%s> element has id '%s' but name '%s'."
% (element, attrs['id'], attrs['name']))
# Test against rule xhtml/section. Remember current section number.
if element == 'a' and parent in ['h2', 'h3', 'h4', 'h5', 'h6']:
level = int(re.match('h([0-9])$', parent).group(1))
if not attrs.has_key('id'):
self.err(17, "Section anchor has no 'id' attribute.")
elif re.match("section-", attrs['id']):
ref_re = ("section-(\\w+" + "\\.[0-9]+"*max(0,level-2) + ")$")
ref_example = "section-1" + ".1" * max(0,level-2)
match = re.match(ref_re, attrs['id'])
if not match:
self.err(18, "Anchor for <%s> has id '%s': should look "
"like '%s'." % (parent, attrs['id'], ref_example))
else:
self.current_section = match.group(1)
# Check (some of) rule xhtml/ref-id.
if (self.current_section == 'A' and len(self.stack2) >= 4
and element == 'a' and self.stack2[-2][0] == 'td'
and self.stack2[-3][0] == 'tr' and self.stack2[-4][0] == 'table'):
if attrs.has_key('id') and not re.match("ref-", attrs['id']):
self.err(21, "Reference anchor has id '%s': should start with "
"'ref-'." % attrs['id'])
# Remember the target of the reference so that we can check the
# rule xhtml/ref-link in the endDocument() method.
if attrs.has_key('id') and attrs.has_key('href'):
self.refs[attrs['id']] = attrs['href']
# Remember cross-refs for later checking against the set of ids.
if (element == 'a' and attrs.has_key('href') and attrs['href']
and attrs['href'][0] == '#'):
self.cross_refs.append((attrs['href'][1:],
self.locator.getLineNumber()))
# Check ordering of sections ( can't follow ).
if element in elt_heading:
level = int(re.match('h([0-9])$', element).group(1))
if level > self.current_heading_level + 1:
self.err(25, "<%s> follows ."
% (element, self.current_heading_level))
self.current_heading_level = level
def characters(self, content):
if len(self.stack2) > 0:
element = self.stack2[-1][0]
# Check that non-blank character data is part of an element only
# when #PCDATA is legal contents.
if (legal_elements.has_key(element)
and '#PCDATA' not in legal_elements[element]
and not re.match('^\\s+$', content)):
self.err(8, "<%s> element contains character data '%s'."
% (element, content))
def endElement(self, element):
contents = self.stack[-1]
# Check that required contents are present.
if not contents and element in nonempty_elements:
self.err(9, "<%s> element is empty." % element)
# Extra contraints. See [2.4].
if contents and constraints.has_key(element):
if not re.match(constraints[element][1],
string.join(contents,'<') + '<'):
self.err(15, "Contents of <%s> element [%s] doesn't match "
"XHTML specification %s."
% (element, string.join(contents, ', '),
constraints[element][0]))
# Section headings (except ) must have anchors so they can be
# referred to (rule xhtml/section).
if element in ['h2', 'h3', 'h4', 'h5', 'h6'] and 'a' not in contents:
self.err(19, "<%s> has no section anchor." % element)
# Pop the stacks.
self.stack2 = self.stack2[0:-1]
self.stack = self.stack[0:-1]
assert self.stack[-1][-1] == element
def endDocument(self):
# Check that cross-references have a target.
for (target, line) in self.cross_refs:
if not self.ids.has_key(target):
self.write_error(22, "Cross-reference '#%s' has no target."
% target, line)
# Check the xhtml/ref-link rule.
elif self.refs.has_key(target):
self.write_error(23, "Cross-reference '#%s' to references "
"section should link to target '%s' instead."
% (target, self.refs[target]), line)
# Check that IDREFs have a target.
for (idref, line) in self.idrefs:
if not self.ids.has_key(idref):
self.write_error(27, "IDREF '%s' has no target." % idref,
line)
def error(self, exception):
line = exception.getLineNumber()
message = exception.getMessage()
# This unfortunately depends on the XML parser producing this
# error.
if message == 'mismatched tag' and self.stack2:
message = ("Mismatched closing tag (opening tag was <%s> "
"at line %d)." % self.stack2[-1])
self.error_stream.write("%s(%d) %s\n" % (self.doc, line, message))
def fatalError(self, exception):
self.error(exception)
raise self.exception
def warning(self, exception):
self.error(exception)
def check(self, path_or_stream):
try:
xml.sax.parse(path_or_stream, self, self)
except self.exception:
pass
# 4. CHECK DIRECTORIES AND FILES
class checker:
noerrors = []
skip = []
def __init__(self, skip = [], noerrors = [],
error_stream = sys.stderr):
self.skip = skip
self.noerrors = noerrors
self.error_stream = error_stream
def check(self, path):
if os.path.isdir(path):
for f in dircache.listdir(path):
if f not in self.skip:
self.check(os.path.join(path, f))
elif (os.path.isfile(path) and re.search("\\.html$", path)
and re.match("<\\?xml", open(path).readline())):
handler(path, self.noerrors, self.error_stream).check(path)
def check_stream(self, name, stream):
handler(name, self.noerrors, self.error_stream).check(stream)
# 5. COMMAND-LINE INTERFACE
def run():
opts, paths = getopt.getopt(sys.argv[1:], 's:n:', ['skip=', 'noerror='])
skip = []
noerrors = []
for (o, a) in opts:
if o in ('-s', '--skip'):
skip.extend(string.split(a, ','))
if o in ('-n', '--noerror'):
noerrors.extend(map(int, string.split(a, ',')))
c = checker(skip, noerrors)
for p in paths:
c.check(p)
if __name__ == "__main__":
run()
# A. REFERENCES
#
# [Chisholm 2000-11-06] "HTML Techniques for Web Content Accessibility
# Guidelines 1.0"; Wendy Chisholm, Gregg Vanderheiden, Ian Jacobs;
# ; 2000-11-06.
#
# [HTML 4.01] "HTML 4.01 Specification"; World Wide Web Consortium;
# ; 1999-12-24.
#
# [Jacobs 2001-04-09] "User Agent Accessibility Guidelines 1.0" (W3C Working
# Draft); Ian Jacobs, Jon Gunderson, Eric Hansen;
# ; 2001-04-09.
#
# [Rules] "Rules"; Gareth Rees; Ravenbrook Limited;
# ; 2001-04-22.
#
# [SAX] "xml.sax -- Support for SAX2 parsers"; Python;
# .
#
# [XHTML 1.0] "XHTML 1.0: The Extensible HyperText Markup Language"; World Wide
# Web Consortium; ; 2000-01-26.
#
# [XHTML 1.0 DTD] "XHTML 1.0 Transitional Document Type Definition"; World Wide
# Web Consortium;
# ; 2000-01-26.
#
#
# B. DOCUMENT HISTORY
#
# 2001-04-29 GDR Created.
#
# 2001-04-30 GDR Added checking of attributes to the check_xhtml.py
# script. Added Ravenbrook document format checks for section headers (rule
# xhtml/section) and reference ids (rule xhtml/ref-id).
#
# 2001-05-01 GDR Added checks for attribute values, cross-references, links to
# the references section, ordering of sections, XHTML constraints for ,
# ,