| 1 | """Shared support for scanning document type declarations in HTML and XHTML. |
| 2 | |
| 3 | This module is used as a foundation for the HTMLParser and sgmllib |
| 4 | modules (indirectly, for htmllib as well). It has no documented |
| 5 | public API and should not be used directly. |
| 6 | |
| 7 | """ |
| 8 | |
| 9 | import re |
| 10 | |
| 11 | _declname_match = re.compile(r'[a-zA-Z][-_.a-zA-Z0-9]*\s*').match |
| 12 | _declstringlit_match = re.compile(r'(\'[^\']*\'|"[^"]*")\s*').match |
| 13 | _commentclose = re.compile(r'--\s*>') |
| 14 | _markedsectionclose = re.compile(r']\s*]\s*>') |
| 15 | |
| 16 | # An analysis of the MS-Word extensions is available at |
| 17 | # http://www.planetpublish.com/xmlarena/xap/Thursday/WordtoXML.pdf |
| 18 | |
| 19 | _msmarkedsectionclose = re.compile(r']\s*>') |
| 20 | |
| 21 | del re |
| 22 | |
| 23 | |
| 24 | class ParserBase: |
| 25 | """Parser base class which provides some common support methods used |
| 26 | by the SGML/HTML and XHTML parsers.""" |
| 27 | |
| 28 | def __init__(self): |
| 29 | if self.__class__ is ParserBase: |
| 30 | raise RuntimeError( |
| 31 | "markupbase.ParserBase must be subclassed") |
| 32 | |
| 33 | def error(self, message): |
| 34 | raise NotImplementedError( |
| 35 | "subclasses of ParserBase must override error()") |
| 36 | |
| 37 | def reset(self): |
| 38 | self.lineno = 1 |
| 39 | self.offset = 0 |
| 40 | |
| 41 | def getpos(self): |
| 42 | """Return current line number and offset.""" |
| 43 | return self.lineno, self.offset |
| 44 | |
| 45 | # Internal -- update line number and offset. This should be |
| 46 | # called for each piece of data exactly once, in order -- in other |
| 47 | # words the concatenation of all the input strings to this |
| 48 | # function should be exactly the entire input. |
| 49 | def updatepos(self, i, j): |
| 50 | if i >= j: |
| 51 | return j |
| 52 | rawdata = self.rawdata |
| 53 | nlines = rawdata.count("\n", i, j) |
| 54 | if nlines: |
| 55 | self.lineno = self.lineno + nlines |
| 56 | pos = rawdata.rindex("\n", i, j) # Should not fail |
| 57 | self.offset = j-(pos+1) |
| 58 | else: |
| 59 | self.offset = self.offset + j-i |
| 60 | return j |
| 61 | |
| 62 | _decl_otherchars = '' |
| 63 | |
| 64 | # Internal -- parse declaration (for use by subclasses). |
| 65 | def parse_declaration(self, i): |
| 66 | # This is some sort of declaration; in "HTML as |
| 67 | # deployed," this should only be the document type |
| 68 | # declaration ("<!DOCTYPE html...>"). |
| 69 | # ISO 8879:1986, however, has more complex |
| 70 | # declaration syntax for elements in <!...>, including: |
| 71 | # --comment-- |
| 72 | # [marked section] |
| 73 | # name in the following list: ENTITY, DOCTYPE, ELEMENT, |
| 74 | # ATTLIST, NOTATION, SHORTREF, USEMAP, |
| 75 | # LINKTYPE, LINK, IDLINK, USELINK, SYSTEM |
| 76 | rawdata = self.rawdata |
| 77 | j = i + 2 |
| 78 | assert rawdata[i:j] == "<!", "unexpected call to parse_declaration" |
| 79 | if rawdata[j:j+1] in ("-", ""): |
| 80 | # Start of comment followed by buffer boundary, |
| 81 | # or just a buffer boundary. |
| 82 | return -1 |
| 83 | # A simple, practical version could look like: ((name|stringlit) S*) + '>' |
| 84 | n = len(rawdata) |
| 85 | if rawdata[j:j+1] == '--': #comment |
| 86 | # Locate --.*-- as the body of the comment |
| 87 | return self.parse_comment(i) |
| 88 | elif rawdata[j] == '[': #marked section |
| 89 | # Locate [statusWord [...arbitrary SGML...]] as the body of the marked section |
| 90 | # Where statusWord is one of TEMP, CDATA, IGNORE, INCLUDE, RCDATA |
| 91 | # Note that this is extended by Microsoft Office "Save as Web" function |
| 92 | # to include [if...] and [endif]. |
| 93 | return self.parse_marked_section(i) |
| 94 | else: #all other declaration elements |
| 95 | decltype, j = self._scan_name(j, i) |
| 96 | if j < 0: |
| 97 | return j |
| 98 | if decltype == "doctype": |
| 99 | self._decl_otherchars = '' |
| 100 | while j < n: |
| 101 | c = rawdata[j] |
| 102 | if c == ">": |
| 103 | # end of declaration syntax |
| 104 | data = rawdata[i+2:j] |
| 105 | if decltype == "doctype": |
| 106 | self.handle_decl(data) |
| 107 | else: |
| 108 | self.unknown_decl(data) |
| 109 | return j + 1 |
| 110 | if c in "\"'": |
| 111 | m = _declstringlit_match(rawdata, j) |
| 112 | if not m: |
| 113 | return -1 # incomplete |
| 114 | j = m.end() |
| 115 | elif c in "abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ": |
| 116 | name, j = self._scan_name(j, i) |
| 117 | elif c in self._decl_otherchars: |
| 118 | j = j + 1 |
| 119 | elif c == "[": |
| 120 | # this could be handled in a separate doctype parser |
| 121 | if decltype == "doctype": |
| 122 | j = self._parse_doctype_subset(j + 1, i) |
| 123 | elif decltype in ("attlist", "linktype", "link", "element"): |
| 124 | # must tolerate []'d groups in a content model in an element declaration |
| 125 | # also in data attribute specifications of attlist declaration |
| 126 | # also link type declaration subsets in linktype declarations |
| 127 | # also link attribute specification lists in link declarations |
| 128 | self.error("unsupported '[' char in %s declaration" % decltype) |
| 129 | else: |
| 130 | self.error("unexpected '[' char in declaration") |
| 131 | else: |
| 132 | self.error( |
| 133 | "unexpected %r char in declaration" % rawdata[j]) |
| 134 | if j < 0: |
| 135 | return j |
| 136 | return -1 # incomplete |
| 137 | |
| 138 | # Internal -- parse a marked section |
| 139 | # Override this to handle MS-word extension syntax <![if word]>content<![endif]> |
| 140 | def parse_marked_section( self, i, report=1 ): |
| 141 | rawdata= self.rawdata |
| 142 | assert rawdata[i:i+3] == '<![', "unexpected call to parse_marked_section()" |
| 143 | sectName, j = self._scan_name( i+3, i ) |
| 144 | if j < 0: |
| 145 | return j |
| 146 | if sectName in ("temp", "cdata", "ignore", "include", "rcdata"): |
| 147 | # look for standard ]]> ending |
| 148 | match= _markedsectionclose.search(rawdata, i+3) |
| 149 | elif sectName in ("if", "else", "endif"): |
| 150 | # look for MS Office ]> ending |
| 151 | match= _msmarkedsectionclose.search(rawdata, i+3) |
| 152 | else: |
| 153 | self.error('unknown status keyword %r in marked section' % rawdata[i+3:j]) |
| 154 | if not match: |
| 155 | return -1 |
| 156 | if report: |
| 157 | j = match.start(0) |
| 158 | self.unknown_decl(rawdata[i+3: j]) |
| 159 | return match.end(0) |
| 160 | |
| 161 | # Internal -- parse comment, return length or -1 if not terminated |
| 162 | def parse_comment(self, i, report=1): |
| 163 | rawdata = self.rawdata |
| 164 | if rawdata[i:i+4] != '<!--': |
| 165 | self.error('unexpected call to parse_comment()') |
| 166 | match = _commentclose.search(rawdata, i+4) |
| 167 | if not match: |
| 168 | return -1 |
| 169 | if report: |
| 170 | j = match.start(0) |
| 171 | self.handle_comment(rawdata[i+4: j]) |
| 172 | return match.end(0) |
| 173 | |
| 174 | # Internal -- scan past the internal subset in a <!DOCTYPE declaration, |
| 175 | # returning the index just past any whitespace following the trailing ']'. |
| 176 | def _parse_doctype_subset(self, i, declstartpos): |
| 177 | rawdata = self.rawdata |
| 178 | n = len(rawdata) |
| 179 | j = i |
| 180 | while j < n: |
| 181 | c = rawdata[j] |
| 182 | if c == "<": |
| 183 | s = rawdata[j:j+2] |
| 184 | if s == "<": |
| 185 | # end of buffer; incomplete |
| 186 | return -1 |
| 187 | if s != "<!": |
| 188 | self.updatepos(declstartpos, j + 1) |
| 189 | self.error("unexpected char in internal subset (in %r)" % s) |
| 190 | if (j + 2) == n: |
| 191 | # end of buffer; incomplete |
| 192 | return -1 |
| 193 | if (j + 4) > n: |
| 194 | # end of buffer; incomplete |
| 195 | return -1 |
| 196 | if rawdata[j:j+4] == "<!--": |
| 197 | j = self.parse_comment(j, report=0) |
| 198 | if j < 0: |
| 199 | return j |
| 200 | continue |
| 201 | name, j = self._scan_name(j + 2, declstartpos) |
| 202 | if j == -1: |
| 203 | return -1 |
| 204 | if name not in ("attlist", "element", "entity", "notation"): |
| 205 | self.updatepos(declstartpos, j + 2) |
| 206 | self.error( |
| 207 | "unknown declaration %r in internal subset" % name) |
| 208 | # handle the individual names |
| 209 | meth = getattr(self, "_parse_doctype_" + name) |
| 210 | j = meth(j, declstartpos) |
| 211 | if j < 0: |
| 212 | return j |
| 213 | elif c == "%": |
| 214 | # parameter entity reference |
| 215 | if (j + 1) == n: |
| 216 | # end of buffer; incomplete |
| 217 | return -1 |
| 218 | s, j = self._scan_name(j + 1, declstartpos) |
| 219 | if j < 0: |
| 220 | return j |
| 221 | if rawdata[j] == ";": |
| 222 | j = j + 1 |
| 223 | elif c == "]": |
| 224 | j = j + 1 |
| 225 | while j < n and rawdata[j].isspace(): |
| 226 | j = j + 1 |
| 227 | if j < n: |
| 228 | if rawdata[j] == ">": |
| 229 | return j |
| 230 | self.updatepos(declstartpos, j) |
| 231 | self.error("unexpected char after internal subset") |
| 232 | else: |
| 233 | return -1 |
| 234 | elif c.isspace(): |
| 235 | j = j + 1 |
| 236 | else: |
| 237 | self.updatepos(declstartpos, j) |
| 238 | self.error("unexpected char %r in internal subset" % c) |
| 239 | # end of buffer reached |
| 240 | return -1 |
| 241 | |
| 242 | # Internal -- scan past <!ELEMENT declarations |
| 243 | def _parse_doctype_element(self, i, declstartpos): |
| 244 | name, j = self._scan_name(i, declstartpos) |
| 245 | if j == -1: |
| 246 | return -1 |
| 247 | # style content model; just skip until '>' |
| 248 | rawdata = self.rawdata |
| 249 | if '>' in rawdata[j:]: |
| 250 | return rawdata.find(">", j) + 1 |
| 251 | return -1 |
| 252 | |
| 253 | # Internal -- scan past <!ATTLIST declarations |
| 254 | def _parse_doctype_attlist(self, i, declstartpos): |
| 255 | rawdata = self.rawdata |
| 256 | name, j = self._scan_name(i, declstartpos) |
| 257 | c = rawdata[j:j+1] |
| 258 | if c == "": |
| 259 | return -1 |
| 260 | if c == ">": |
| 261 | return j + 1 |
| 262 | while 1: |
| 263 | # scan a series of attribute descriptions; simplified: |
| 264 | # name type [value] [#constraint] |
| 265 | name, j = self._scan_name(j, declstartpos) |
| 266 | if j < 0: |
| 267 | return j |
| 268 | c = rawdata[j:j+1] |
| 269 | if c == "": |
| 270 | return -1 |
| 271 | if c == "(": |
| 272 | # an enumerated type; look for ')' |
| 273 | if ")" in rawdata[j:]: |
| 274 | j = rawdata.find(")", j) + 1 |
| 275 | else: |
| 276 | return -1 |
| 277 | while rawdata[j:j+1].isspace(): |
| 278 | j = j + 1 |
| 279 | if not rawdata[j:]: |
| 280 | # end of buffer, incomplete |
| 281 | return -1 |
| 282 | else: |
| 283 | name, j = self._scan_name(j, declstartpos) |
| 284 | c = rawdata[j:j+1] |
| 285 | if not c: |
| 286 | return -1 |
| 287 | if c in "'\"": |
| 288 | m = _declstringlit_match(rawdata, j) |
| 289 | if m: |
| 290 | j = m.end() |
| 291 | else: |
| 292 | return -1 |
| 293 | c = rawdata[j:j+1] |
| 294 | if not c: |
| 295 | return -1 |
| 296 | if c == "#": |
| 297 | if rawdata[j:] == "#": |
| 298 | # end of buffer |
| 299 | return -1 |
| 300 | name, j = self._scan_name(j + 1, declstartpos) |
| 301 | if j < 0: |
| 302 | return j |
| 303 | c = rawdata[j:j+1] |
| 304 | if not c: |
| 305 | return -1 |
| 306 | if c == '>': |
| 307 | # all done |
| 308 | return j + 1 |
| 309 | |
| 310 | # Internal -- scan past <!NOTATION declarations |
| 311 | def _parse_doctype_notation(self, i, declstartpos): |
| 312 | name, j = self._scan_name(i, declstartpos) |
| 313 | if j < 0: |
| 314 | return j |
| 315 | rawdata = self.rawdata |
| 316 | while 1: |
| 317 | c = rawdata[j:j+1] |
| 318 | if not c: |
| 319 | # end of buffer; incomplete |
| 320 | return -1 |
| 321 | if c == '>': |
| 322 | return j + 1 |
| 323 | if c in "'\"": |
| 324 | m = _declstringlit_match(rawdata, j) |
| 325 | if not m: |
| 326 | return -1 |
| 327 | j = m.end() |
| 328 | else: |
| 329 | name, j = self._scan_name(j, declstartpos) |
| 330 | if j < 0: |
| 331 | return j |
| 332 | |
| 333 | # Internal -- scan past <!ENTITY declarations |
| 334 | def _parse_doctype_entity(self, i, declstartpos): |
| 335 | rawdata = self.rawdata |
| 336 | if rawdata[i:i+1] == "%": |
| 337 | j = i + 1 |
| 338 | while 1: |
| 339 | c = rawdata[j:j+1] |
| 340 | if not c: |
| 341 | return -1 |
| 342 | if c.isspace(): |
| 343 | j = j + 1 |
| 344 | else: |
| 345 | break |
| 346 | else: |
| 347 | j = i |
| 348 | name, j = self._scan_name(j, declstartpos) |
| 349 | if j < 0: |
| 350 | return j |
| 351 | while 1: |
| 352 | c = self.rawdata[j:j+1] |
| 353 | if not c: |
| 354 | return -1 |
| 355 | if c in "'\"": |
| 356 | m = _declstringlit_match(rawdata, j) |
| 357 | if m: |
| 358 | j = m.end() |
| 359 | else: |
| 360 | return -1 # incomplete |
| 361 | elif c == ">": |
| 362 | return j + 1 |
| 363 | else: |
| 364 | name, j = self._scan_name(j, declstartpos) |
| 365 | if j < 0: |
| 366 | return j |
| 367 | |
| 368 | # Internal -- scan a name token and the new position and the token, or |
| 369 | # return -1 if we've reached the end of the buffer. |
| 370 | def _scan_name(self, i, declstartpos): |
| 371 | rawdata = self.rawdata |
| 372 | n = len(rawdata) |
| 373 | if i == n: |
| 374 | return None, -1 |
| 375 | m = _declname_match(rawdata, i) |
| 376 | if m: |
| 377 | s = m.group() |
| 378 | name = s.strip() |
| 379 | if (i + len(s)) == n: |
| 380 | return None, -1 # end of buffer |
| 381 | return name.lower(), m.end() |
| 382 | else: |
| 383 | self.updatepos(declstartpos, i) |
| 384 | self.error("expected name token at %r" |
| 385 | % rawdata[declstartpos:declstartpos+20]) |
| 386 | |
| 387 | # To be overridden -- handlers for unknown objects |
| 388 | def unknown_decl(self, data): |
| 389 | pass |