Mercurial > cpython-lru_cache-weakref
annotate Lib/HTMLParser.py @ 17853:055fc046e080 legacy-trunk
A much improved HTML parser -- a replacement for sgmllib. The API is
derived from but not quite compatible with that of sgmllib, so it's a
new file. I suppose it needs documentation, and htmllib needs to be
changed to use this instead of sgmllib, and sgmllib needs to be
declared obsolete. But that can all be done later.
This code was first published as part of TAL (part of Zope Page
Templates), but that was strongly based on sgmllib anyway. Authors
are Fred drake and Guido van Rossum.
| author | Guido van Rossum <guido@python.org> |
|---|---|
| date | Fri, 18 May 2001 14:50:52 +0000 |
| parents | |
| children | ce7f0e380830 |
| rev | line source |
|---|---|
|
17853
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
1 """A parser for HTML.""" |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
2 |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
3 # This file is based on sgmllib.py, but the API is slightly different. |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
4 |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
5 # XXX There should be a way to distinguish between PCDATA (parsed |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
6 # character data -- the normal case), RCDATA (replaceable character |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
7 # data -- only char and entity references and end tags are special) |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
8 # and CDATA (character data -- only end tags are special). |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
9 |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
10 |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
11 import re |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
12 import string |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
13 |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
14 # Regular expressions used for parsing |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
15 |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
16 interesting_normal = re.compile('[&<]') |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
17 interesting_cdata = re.compile(r'<(/|\Z)') |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
18 incomplete = re.compile('&([a-zA-Z][-.a-zA-Z0-9]*|#[0-9]*)?') |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
19 |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
20 entityref = re.compile('&([a-zA-Z][-.a-zA-Z0-9]*)[^a-zA-Z0-9]') |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
21 charref = re.compile('&#([0-9]+)[^0-9]') |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
22 |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
23 starttagopen = re.compile('<[a-zA-Z]') |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
24 piopen = re.compile(r'<\?') |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
25 piclose = re.compile('>') |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
26 endtagopen = re.compile('</') |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
27 declopen = re.compile('<!') |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
28 special = re.compile('<![^<>]*>') |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
29 commentopen = re.compile('<!--') |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
30 commentclose = re.compile(r'--\s*>') |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
31 tagfind = re.compile('[a-zA-Z][-.a-zA-Z0-9:_]*') |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
32 attrfind = re.compile( |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
33 r'\s*([a-zA-Z_][-.:a-zA-Z_0-9]*)(\s*=\s*' |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
34 r'(\'[^\']*\'|"[^"]*"|[-a-zA-Z0-9./:;+*%?!&$\(\)_#=~]*))?') |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
35 |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
36 locatestarttagend = re.compile(r""" |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
37 <[a-zA-Z][-.a-zA-Z0-9:_]* # tag name |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
38 (?:\s+ # whitespace before attribute name |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
39 (?:[a-zA-Z_][-.:a-zA-Z0-9_]* # attribute name |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
40 (?:\s*=\s* # value indicator |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
41 (?:'[^']*' # LITA-enclosed value |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
42 |\"[^\"]*\" # LIT-enclosed value |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
43 |[^'\">\s]+ # bare value |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
44 ) |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
45 )? |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
46 ) |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
47 )* |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
48 \s* # trailing whitespace |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
49 """, re.VERBOSE) |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
50 endstarttag = re.compile(r"\s*/?>") |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
51 endendtag = re.compile('>') |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
52 endtagfind = re.compile('</\s*([a-zA-Z][-.a-zA-Z0-9:_]*)\s*>') |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
53 |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
54 declname = re.compile(r'[a-zA-Z][-_.a-zA-Z0-9]*\s*') |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
55 declstringlit = re.compile(r'(\'[^\']*\'|"[^"]*")\s*') |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
56 |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
57 |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
58 class HTMLParseError(Exception): |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
59 """Exception raised for all parse errors.""" |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
60 |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
61 def __init__(self, msg, position=(None, None)): |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
62 assert msg |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
63 self.msg = msg |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
64 self.lineno = position[0] |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
65 self.offset = position[1] |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
66 |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
67 def __str__(self): |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
68 result = self.msg |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
69 if self.lineno is not None: |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
70 result = result + ", at line %d" % self.lineno |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
71 if self.offset is not None: |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
72 result = result + ", column %d" % (self.offset + 1) |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
73 return result |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
74 |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
75 |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
76 # HTML parser class -- find tags and call handler functions. |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
77 # Usage: p = HTMLParser(); p.feed(data); ...; p.close(). |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
78 # The dtd is defined by deriving a class which defines methods |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
79 # with special names to handle tags: start_foo and end_foo to handle |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
80 # <foo> and </foo>, respectively, or do_foo to handle <foo> by itself. |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
81 # (Tags are converted to lower case for this purpose.) The data |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
82 # between tags is passed to the parser by calling self.handle_data() |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
83 # with some data as argument (the data may be split up in arbitrary |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
84 # chunks). Entity references are passed by calling |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
85 # self.handle_entityref() with the entity reference as argument. |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
86 |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
87 class HTMLParser: |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
88 |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
89 CDATA_CONTENT_ELEMENTS = ("script", "style") |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
90 |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
91 |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
92 # Interface -- initialize and reset this instance |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
93 def __init__(self): |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
94 self.reset() |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
95 |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
96 # Interface -- reset this instance. Loses all unprocessed data |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
97 def reset(self): |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
98 self.rawdata = '' |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
99 self.stack = [] |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
100 self.lasttag = '???' |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
101 self.lineno = 1 |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
102 self.offset = 0 |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
103 self.interesting = interesting_normal |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
104 |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
105 # Interface -- feed some data to the parser. Call this as |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
106 # often as you want, with as little or as much text as you |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
107 # want (may include '\n'). (This just saves the text, all the |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
108 # processing is done by goahead().) |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
109 def feed(self, data): |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
110 self.rawdata = self.rawdata + data |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
111 self.goahead(0) |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
112 |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
113 # Interface -- handle the remaining data |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
114 def close(self): |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
115 self.goahead(1) |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
116 |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
117 # Internal -- update line number and offset. This should be |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
118 # called for each piece of data exactly once, in order -- in other |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
119 # words the concatenation of all the input strings to this |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
120 # function should be exactly the entire input. |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
121 def updatepos(self, i, j): |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
122 if i >= j: |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
123 return j |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
124 rawdata = self.rawdata |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
125 nlines = string.count(rawdata, "\n", i, j) |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
126 if nlines: |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
127 self.lineno = self.lineno + nlines |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
128 pos = string.rindex(rawdata, "\n", i, j) # Should not fail |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
129 self.offset = j-(pos+1) |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
130 else: |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
131 self.offset = self.offset + j-i |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
132 return j |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
133 |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
134 # Interface -- return current line number and offset. |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
135 def getpos(self): |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
136 return self.lineno, self.offset |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
137 |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
138 __starttag_text = None |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
139 |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
140 # Interface -- return full source of start tag: "<...>" |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
141 def get_starttag_text(self): |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
142 return self.__starttag_text |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
143 |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
144 def set_cdata_mode(self): |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
145 self.interesting = interesting_cdata |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
146 |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
147 def clear_cdata_mode(self): |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
148 self.interesting = interesting_normal |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
149 |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
150 # Internal -- handle data as far as reasonable. May leave state |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
151 # and data to be processed by a subsequent call. If 'end' is |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
152 # true, force handling all data as if followed by EOF marker. |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
153 def goahead(self, end): |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
154 rawdata = self.rawdata |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
155 i = 0 |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
156 n = len(rawdata) |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
157 while i < n: |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
158 match = self.interesting.search(rawdata, i) # < or & |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
159 if match: |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
160 j = match.start() |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
161 else: |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
162 j = n |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
163 if i < j: self.handle_data(rawdata[i:j]) |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
164 i = self.updatepos(i, j) |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
165 if i == n: break |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
166 if rawdata[i] == '<': |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
167 if starttagopen.match(rawdata, i): # < + letter |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
168 k = self.parse_starttag(i) |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
169 elif endtagopen.match(rawdata, i): # </ |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
170 k = self.parse_endtag(i) |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
171 if k >= 0: |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
172 self.clear_cdata_mode() |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
173 elif commentopen.match(rawdata, i): # <!-- |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
174 k = self.parse_comment(i) |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
175 elif piopen.match(rawdata, i): # <? |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
176 k = self.parse_pi(i) |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
177 elif declopen.match(rawdata, i): # <! |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
178 k = self.parse_declaration(i) |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
179 else: |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
180 if i < n-1: |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
181 raise HTMLParseError( |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
182 "invalid '<' construct: %s" % `rawdata[i:i+2]`, |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
183 self.getpos()) |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
184 k = -1 |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
185 if k < 0: |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
186 if end: |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
187 raise HTMLParseError("EOF in middle of construct", |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
188 self.getpos()) |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
189 break |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
190 i = self.updatepos(i, k) |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
191 elif rawdata[i] == '&': |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
192 match = charref.match(rawdata, i) |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
193 if match: |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
194 name = match.group(1) |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
195 self.handle_charref(name) |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
196 k = match.end() |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
197 if rawdata[k-1] != ';': |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
198 k = k-1 |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
199 i = self.updatepos(i, k) |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
200 continue |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
201 match = entityref.match(rawdata, i) |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
202 if match: |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
203 name = match.group(1) |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
204 self.handle_entityref(name) |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
205 k = match.end() |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
206 if rawdata[k-1] != ';': |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
207 k = k-1 |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
208 i = self.updatepos(i, k) |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
209 continue |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
210 if incomplete.match(rawdata, i): |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
211 if end: |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
212 raise HTMLParseError( |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
213 "EOF in middle of entity or char ref", |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
214 self.getpos()) |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
215 return -1 # incomplete |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
216 raise HTMLParseError("'&' not part of entity or char ref", |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
217 self.getpos()) |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
218 else: |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
219 assert 0, "interesting.search() lied" |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
220 # end while |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
221 if end and i < n: |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
222 self.handle_data(rawdata[i:n]) |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
223 i = self.updatepos(i, n) |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
224 self.rawdata = rawdata[i:] |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
225 |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
226 # Internal -- parse comment, return end or -1 if not terminated |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
227 def parse_comment(self, i): |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
228 rawdata = self.rawdata |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
229 assert rawdata[i:i+4] == '<!--', 'unexpected call to parse_comment()' |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
230 match = commentclose.search(rawdata, i+4) |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
231 if not match: |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
232 return -1 |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
233 j = match.start() |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
234 self.handle_comment(rawdata[i+4: j]) |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
235 j = match.end() |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
236 return j |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
237 |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
238 # Internal -- parse declaration. |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
239 def parse_declaration(self, i): |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
240 # This is some sort of declaration; in "HTML as |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
241 # deployed," this should only be the document type |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
242 # declaration ("<!DOCTYPE html...>"). |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
243 rawdata = self.rawdata |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
244 j = i + 2 |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
245 assert rawdata[i:j] == "<!", "unexpected call to parse_declaration" |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
246 if rawdata[j:j+1] in ("-", ""): |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
247 # Start of comment followed by buffer boundary, |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
248 # or just a buffer boundary. |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
249 return -1 |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
250 # in practice, this should look like: ((name|stringlit) S*)+ '>' |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
251 n = len(rawdata) |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
252 while j < n: |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
253 c = rawdata[j] |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
254 if c == ">": |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
255 # end of declaration syntax |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
256 self.handle_decl(rawdata[i+2:j]) |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
257 return j + 1 |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
258 if c in "\"'": |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
259 m = declstringlit.match(rawdata, j) |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
260 if not m: |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
261 return -1 # incomplete |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
262 j = m.end() |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
263 elif c in "abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ": |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
264 m = declname.match(rawdata, j) |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
265 if not m: |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
266 return -1 # incomplete |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
267 j = m.end() |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
268 else: |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
269 raise HTMLParseError( |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
270 "unexpected char in declaration: %s" % `rawdata[j]`, |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
271 self.getpos()) |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
272 return -1 # incomplete |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
273 |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
274 # Internal -- parse processing instr, return end or -1 if not terminated |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
275 def parse_pi(self, i): |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
276 rawdata = self.rawdata |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
277 assert rawdata[i:i+2] == '<?', 'unexpected call to parse_pi()' |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
278 match = piclose.search(rawdata, i+2) # > |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
279 if not match: |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
280 return -1 |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
281 j = match.start() |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
282 self.handle_pi(rawdata[i+2: j]) |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
283 j = match.end() |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
284 return j |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
285 |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
286 # Internal -- handle starttag, return end or -1 if not terminated |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
287 def parse_starttag(self, i): |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
288 self.__starttag_text = None |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
289 endpos = self.check_for_whole_start_tag(i) |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
290 if endpos < 0: |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
291 return endpos |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
292 rawdata = self.rawdata |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
293 self.__starttag_text = rawdata[i:endpos] |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
294 |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
295 # Now parse the data between i+1 and j into a tag and attrs |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
296 attrs = [] |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
297 match = tagfind.match(rawdata, i+1) |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
298 assert match, 'unexpected call to parse_starttag()' |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
299 k = match.end() |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
300 self.lasttag = tag = string.lower(rawdata[i+1:k]) |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
301 |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
302 while k < endpos: |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
303 m = attrfind.match(rawdata, k) |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
304 if not m: |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
305 break |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
306 attrname, rest, attrvalue = m.group(1, 2, 3) |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
307 if not rest: |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
308 attrvalue = None |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
309 elif attrvalue[:1] == '\'' == attrvalue[-1:] or \ |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
310 attrvalue[:1] == '"' == attrvalue[-1:]: |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
311 attrvalue = attrvalue[1:-1] |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
312 attrvalue = self.unescape(attrvalue) |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
313 attrs.append((string.lower(attrname), attrvalue)) |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
314 k = m.end() |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
315 |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
316 end = string.strip(rawdata[k:endpos]) |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
317 if end not in (">", "/>"): |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
318 lineno, offset = self.getpos() |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
319 if "\n" in self.__starttag_text: |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
320 lineno = lineno + string.count(self.__starttag_text, "\n") |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
321 offset = len(self.__starttag_text) \ |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
322 - string.rfind(self.__starttag_text, "\n") |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
323 else: |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
324 offset = offset + len(self.__starttag_text) |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
325 raise HTMLParseError("junk characters in start tag: %s" |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
326 % `rawdata[k:endpos][:20]`, |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
327 (lineno, offset)) |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
328 if end[-2:] == '/>': |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
329 # XHTML-style empty tag: <span attr="value" /> |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
330 self.handle_startendtag(tag, attrs) |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
331 else: |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
332 self.handle_starttag(tag, attrs) |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
333 if tag in self.CDATA_CONTENT_ELEMENTS: |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
334 self.set_cdata_mode() |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
335 return endpos |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
336 |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
337 # Internal -- check to see if we have a complete starttag; return end |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
338 # or -1 if incomplete. |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
339 def check_for_whole_start_tag(self, i): |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
340 rawdata = self.rawdata |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
341 m = locatestarttagend.match(rawdata, i) |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
342 if m: |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
343 j = m.end() |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
344 next = rawdata[j:j+1] |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
345 if next == ">": |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
346 return j + 1 |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
347 if next == "/": |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
348 s = rawdata[j:j+2] |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
349 if s == "/>": |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
350 return j + 2 |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
351 if s == "/": |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
352 # buffer boundary |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
353 return -1 |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
354 # else bogus input |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
355 self.updatepos(i, j + 1) |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
356 raise HTMLParseError("malformed empty start tag", |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
357 self.getpos()) |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
358 if next == "": |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
359 # end of input |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
360 return -1 |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
361 if next in ("abcdefghijklmnopqrstuvwxyz=/" |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
362 "ABCDEFGHIJKLMNOPQRSTUVWXYZ"): |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
363 # end of input in or before attribute value, or we have the |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
364 # '/' from a '/>' ending |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
365 return -1 |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
366 self.updatepos(i, j) |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
367 raise HTMLParseError("malformed start tag", self.getpos()) |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
368 raise AssertionError("we should not gt here!") |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
369 |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
370 # Internal -- parse endtag, return end or -1 if incomplete |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
371 def parse_endtag(self, i): |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
372 rawdata = self.rawdata |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
373 assert rawdata[i:i+2] == "</", "unexpected call to parse_endtag" |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
374 match = endendtag.search(rawdata, i+1) # > |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
375 if not match: |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
376 return -1 |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
377 j = match.end() |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
378 match = endtagfind.match(rawdata, i) # </ + tag + > |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
379 if not match: |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
380 raise HTMLParseError("bad end tag: %s" % `rawdata[i:j]`, |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
381 self.getpos()) |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
382 tag = match.group(1) |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
383 self.handle_endtag(string.lower(tag)) |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
384 return j |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
385 |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
386 # Overridable -- finish processing of start+end tag: <tag.../> |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
387 def handle_startendtag(self, tag, attrs): |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
388 self.handle_starttag(tag, attrs) |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
389 self.handle_endtag(tag) |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
390 |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
391 # Overridable -- handle start tag |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
392 def handle_starttag(self, tag, attrs): |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
393 pass |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
394 |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
395 # Overridable -- handle end tag |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
396 def handle_endtag(self, tag): |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
397 pass |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
398 |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
399 # Overridable -- handle character reference |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
400 def handle_charref(self, name): |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
401 pass |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
402 |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
403 # Overridable -- handle entity reference |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
404 def handle_entityref(self, name): |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
405 pass |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
406 |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
407 # Overridable -- handle data |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
408 def handle_data(self, data): |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
409 pass |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
410 |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
411 # Overridable -- handle comment |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
412 def handle_comment(self, data): |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
413 pass |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
414 |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
415 # Overridable -- handle declaration |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
416 def handle_decl(self, decl): |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
417 pass |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
418 |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
419 # Overridable -- handle processing instruction |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
420 def handle_pi(self, data): |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
421 pass |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
422 |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
423 # Internal -- helper to remove special character quoting |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
424 def unescape(self, s): |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
425 if '&' not in s: |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
426 return s |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
427 s = string.replace(s, "<", "<") |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
428 s = string.replace(s, ">", ">") |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
429 s = string.replace(s, "'", "'") |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
430 s = string.replace(s, """, '"') |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
431 s = string.replace(s, "&", "&") # Must be last |
|
055fc046e080
A much improved HTML parser -- a replacement for sgmllib. The API is
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
432 return s |
