genshi/genshi-test: markup/input.py annotate

annotate markup/input.py @ 66:822089ae65ce

Switch copyright to Edgewall and URLs to markup.edgewall.org.

author	cmlenz
date	Sun, 09 Jul 2006 17:46:12 +0000
parents	b8456279c444
children	e9a3930f8823

rev	line source
1 821114ec4f69 Initial import. cmlenz parents: diff changeset	1 # -- coding: utf-8 --
821114ec4f69 Initial import. cmlenz parents: diff changeset	2 #
66 822089ae65ce Switch copyright to Edgewall and URLs to markup.edgewall.org. cmlenz parents: 27 diff changeset	3 # Copyright (C) 2006 Edgewall Software
1 821114ec4f69 Initial import. cmlenz parents: diff changeset	4 # All rights reserved.
821114ec4f69 Initial import. cmlenz parents: diff changeset	5 #
821114ec4f69 Initial import. cmlenz parents: diff changeset	6 # This software is licensed as described in the file COPYING, which
821114ec4f69 Initial import. cmlenz parents: diff changeset	7 # you should have received as part of this distribution. The terms
66 822089ae65ce Switch copyright to Edgewall and URLs to markup.edgewall.org. cmlenz parents: 27 diff changeset	8 # are also available at http://markup.edgewall.org/wiki/License.
1 821114ec4f69 Initial import. cmlenz parents: diff changeset	9 #
821114ec4f69 Initial import. cmlenz parents: diff changeset	10 # This software consists of voluntary contributions made by many
821114ec4f69 Initial import. cmlenz parents: diff changeset	11 # individuals. For the exact contribution history, see the revision
66 822089ae65ce Switch copyright to Edgewall and URLs to markup.edgewall.org. cmlenz parents: 27 diff changeset	12 # history and logs, available at http://markup.edgewall.org/log/.
1 821114ec4f69 Initial import. cmlenz parents: diff changeset	13
821114ec4f69 Initial import. cmlenz parents: diff changeset	14 from xml.parsers import expat
821114ec4f69 Initial import. cmlenz parents: diff changeset	15 try:
821114ec4f69 Initial import. cmlenz parents: diff changeset	16 frozenset
821114ec4f69 Initial import. cmlenz parents: diff changeset	17 except NameError:
821114ec4f69 Initial import. cmlenz parents: diff changeset	18 from sets import ImmutableSet as frozenset
821114ec4f69 Initial import. cmlenz parents: diff changeset	19 import HTMLParser as html
821114ec4f69 Initial import. cmlenz parents: diff changeset	20 import htmlentitydefs
821114ec4f69 Initial import. cmlenz parents: diff changeset	21 from StringIO import StringIO
821114ec4f69 Initial import. cmlenz parents: diff changeset	22
821114ec4f69 Initial import. cmlenz parents: diff changeset	23 from markup.core import Attributes, Markup, QName, Stream
821114ec4f69 Initial import. cmlenz parents: diff changeset	24
821114ec4f69 Initial import. cmlenz parents: diff changeset	25
21 eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	26 class ParseError(Exception):
eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	27 """Exception raised when fatal syntax errors are found in the input being
eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	28 parsed."""
eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	29
eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	30 def __init__(self, message, filename='<string>', lineno=-1, offset=-1):
eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	31 Exception.__init__(self, message)
eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	32 self.filename = filename
eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	33 self.lineno = lineno
eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	34 self.offset = offset
eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	35
eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	36
1 821114ec4f69 Initial import. cmlenz parents: diff changeset	37 class XMLParser(object):
821114ec4f69 Initial import. cmlenz parents: diff changeset	38 """Generator-based XML parser based on roughly equivalent code in
26 039fc5b87405 * Split out the XPath tests into a separate `unittest`-based file. cmlenz parents: 21 diff changeset	39 Kid/ElementTree.
039fc5b87405 * Split out the XPath tests into a separate `unittest`-based file. cmlenz parents: 21 diff changeset	40
039fc5b87405 * Split out the XPath tests into a separate `unittest`-based file. cmlenz parents: 21 diff changeset	41 The parsing is initiated by iterating over the parser object:
039fc5b87405 * Split out the XPath tests into a separate `unittest`-based file. cmlenz parents: 21 diff changeset	42
039fc5b87405 * Split out the XPath tests into a separate `unittest`-based file. cmlenz parents: 21 diff changeset	43 >>> parser = XMLParser(StringIO('<root id="2"><child>Foo</child></root>'))
039fc5b87405 * Split out the XPath tests into a separate `unittest`-based file. cmlenz parents: 21 diff changeset	44 >>> for kind, data, pos in parser:
039fc5b87405 * Split out the XPath tests into a separate `unittest`-based file. cmlenz parents: 21 diff changeset	45 ... print kind, data
039fc5b87405 * Split out the XPath tests into a separate `unittest`-based file. cmlenz parents: 21 diff changeset	46 START (u'root', [(u'id', u'2')])
039fc5b87405 * Split out the XPath tests into a separate `unittest`-based file. cmlenz parents: 21 diff changeset	47 START (u'child', [])
039fc5b87405 * Split out the XPath tests into a separate `unittest`-based file. cmlenz parents: 21 diff changeset	48 TEXT Foo
039fc5b87405 * Split out the XPath tests into a separate `unittest`-based file. cmlenz parents: 21 diff changeset	49 END child
039fc5b87405 * Split out the XPath tests into a separate `unittest`-based file. cmlenz parents: 21 diff changeset	50 END root
039fc5b87405 * Split out the XPath tests into a separate `unittest`-based file. cmlenz parents: 21 diff changeset	51 """
1 821114ec4f69 Initial import. cmlenz parents: diff changeset	52
21 eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	53 def __init__(self, source, filename=None):
26 039fc5b87405 * Split out the XPath tests into a separate `unittest`-based file. cmlenz parents: 21 diff changeset	54 """Initialize the parser for the given XML text.
039fc5b87405 * Split out the XPath tests into a separate `unittest`-based file. cmlenz parents: 21 diff changeset	55
039fc5b87405 * Split out the XPath tests into a separate `unittest`-based file. cmlenz parents: 21 diff changeset	56 @param source: the XML text as a file-like object
039fc5b87405 * Split out the XPath tests into a separate `unittest`-based file. cmlenz parents: 21 diff changeset	57 @param filename: the name of the file, if appropriate
039fc5b87405 * Split out the XPath tests into a separate `unittest`-based file. cmlenz parents: 21 diff changeset	58 """
1 821114ec4f69 Initial import. cmlenz parents: diff changeset	59 self.source = source
21 eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	60 self.filename = filename
1 821114ec4f69 Initial import. cmlenz parents: diff changeset	61
821114ec4f69 Initial import. cmlenz parents: diff changeset	62 # Setup the Expat parser
821114ec4f69 Initial import. cmlenz parents: diff changeset	63 parser = expat.ParserCreate('utf-8', '}')
821114ec4f69 Initial import. cmlenz parents: diff changeset	64 parser.buffer_text = True
821114ec4f69 Initial import. cmlenz parents: diff changeset	65 parser.returns_unicode = True
821114ec4f69 Initial import. cmlenz parents: diff changeset	66 parser.StartElementHandler = self._handle_start
821114ec4f69 Initial import. cmlenz parents: diff changeset	67 parser.EndElementHandler = self._handle_end
821114ec4f69 Initial import. cmlenz parents: diff changeset	68 parser.CharacterDataHandler = self._handle_data
821114ec4f69 Initial import. cmlenz parents: diff changeset	69 parser.XmlDeclHandler = self._handle_prolog
821114ec4f69 Initial import. cmlenz parents: diff changeset	70 parser.StartDoctypeDeclHandler = self._handle_doctype
821114ec4f69 Initial import. cmlenz parents: diff changeset	71 parser.StartNamespaceDeclHandler = self._handle_start_ns
821114ec4f69 Initial import. cmlenz parents: diff changeset	72 parser.EndNamespaceDeclHandler = self._handle_end_ns
821114ec4f69 Initial import. cmlenz parents: diff changeset	73 parser.ProcessingInstructionHandler = self._handle_pi
821114ec4f69 Initial import. cmlenz parents: diff changeset	74 parser.CommentHandler = self._handle_comment
821114ec4f69 Initial import. cmlenz parents: diff changeset	75 parser.DefaultHandler = self._handle_other
821114ec4f69 Initial import. cmlenz parents: diff changeset	76
821114ec4f69 Initial import. cmlenz parents: diff changeset	77 # Location reporting is only support in Python >= 2.4
821114ec4f69 Initial import. cmlenz parents: diff changeset	78 if not hasattr(parser, 'CurrentLineNumber'):
21 eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	79 self._getpos = self._getpos_unknown
1 821114ec4f69 Initial import. cmlenz parents: diff changeset	80
821114ec4f69 Initial import. cmlenz parents: diff changeset	81 self.expat = parser
21 eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	82 self._queue = []
1 821114ec4f69 Initial import. cmlenz parents: diff changeset	83
821114ec4f69 Initial import. cmlenz parents: diff changeset	84 def __iter__(self):
21 eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	85 try:
eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	86 bufsize = 4 * 1024 # 4K
eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	87 done = False
eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	88 while True:
eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	89 while not done and len(self._queue) == 0:
eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	90 data = self.source.read(bufsize)
eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	91 if data == '': # end of data
eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	92 if hasattr(self, 'expat'):
eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	93 self.expat.Parse('', True)
eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	94 del self.expat # get rid of circular references
eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	95 done = True
eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	96 else:
eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	97 self.expat.Parse(data, False)
eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	98 for event in self._queue:
eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	99 yield event
eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	100 self._queue = []
eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	101 if done:
eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	102 break
eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	103 except expat.ExpatError, e:
eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	104 msg = str(e)
eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	105 if self.filename:
eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	106 msg += ', in ' + self.filename
eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	107 raise ParseError(msg, self.filename, e.lineno, e.offset)
1 821114ec4f69 Initial import. cmlenz parents: diff changeset	108
26 039fc5b87405 * Split out the XPath tests into a separate `unittest`-based file. cmlenz parents: 21 diff changeset	109 def _enqueue(self, kind, data, pos=None):
039fc5b87405 * Split out the XPath tests into a separate `unittest`-based file. cmlenz parents: 21 diff changeset	110 if pos is None:
039fc5b87405 * Split out the XPath tests into a separate `unittest`-based file. cmlenz parents: 21 diff changeset	111 pos = self._getpos()
039fc5b87405 * Split out the XPath tests into a separate `unittest`-based file. cmlenz parents: 21 diff changeset	112 self._queue.append((kind, data, pos))
039fc5b87405 * Split out the XPath tests into a separate `unittest`-based file. cmlenz parents: 21 diff changeset	113
1 821114ec4f69 Initial import. cmlenz parents: diff changeset	114 def _getpos_unknown(self):
21 eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	115 return (self.filename or '<string>', -1, -1)
1 821114ec4f69 Initial import. cmlenz parents: diff changeset	116
21 eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	117 def _getpos(self):
eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	118 return (self.filename or '<string>', self.expat.CurrentLineNumber,
eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	119 self.expat.CurrentColumnNumber)
1 821114ec4f69 Initial import. cmlenz parents: diff changeset	120
821114ec4f69 Initial import. cmlenz parents: diff changeset	121 def _handle_start(self, tag, attrib):
26 039fc5b87405 * Split out the XPath tests into a separate `unittest`-based file. cmlenz parents: 21 diff changeset	122 self._enqueue(Stream.START, (QName(tag), Attributes(attrib.items())))
1 821114ec4f69 Initial import. cmlenz parents: diff changeset	123
821114ec4f69 Initial import. cmlenz parents: diff changeset	124 def _handle_end(self, tag):
26 039fc5b87405 * Split out the XPath tests into a separate `unittest`-based file. cmlenz parents: 21 diff changeset	125 self._enqueue(Stream.END, QName(tag))
1 821114ec4f69 Initial import. cmlenz parents: diff changeset	126
821114ec4f69 Initial import. cmlenz parents: diff changeset	127 def _handle_data(self, text):
26 039fc5b87405 * Split out the XPath tests into a separate `unittest`-based file. cmlenz parents: 21 diff changeset	128 self._enqueue(Stream.TEXT, text)
1 821114ec4f69 Initial import. cmlenz parents: diff changeset	129
821114ec4f69 Initial import. cmlenz parents: diff changeset	130 def _handle_prolog(self, version, encoding, standalone):
26 039fc5b87405 * Split out the XPath tests into a separate `unittest`-based file. cmlenz parents: 21 diff changeset	131 self._enqueue(Stream.PROLOG, (version, encoding, standalone))
1 821114ec4f69 Initial import. cmlenz parents: diff changeset	132
821114ec4f69 Initial import. cmlenz parents: diff changeset	133 def _handle_doctype(self, name, sysid, pubid, has_internal_subset):
26 039fc5b87405 * Split out the XPath tests into a separate `unittest`-based file. cmlenz parents: 21 diff changeset	134 self._enqueue(Stream.DOCTYPE, (name, pubid, sysid))
1 821114ec4f69 Initial import. cmlenz parents: diff changeset	135
821114ec4f69 Initial import. cmlenz parents: diff changeset	136 def _handle_start_ns(self, prefix, uri):
26 039fc5b87405 * Split out the XPath tests into a separate `unittest`-based file. cmlenz parents: 21 diff changeset	137 self._enqueue(Stream.START_NS, (prefix or '', uri))
1 821114ec4f69 Initial import. cmlenz parents: diff changeset	138
821114ec4f69 Initial import. cmlenz parents: diff changeset	139 def _handle_end_ns(self, prefix):
26 039fc5b87405 * Split out the XPath tests into a separate `unittest`-based file. cmlenz parents: 21 diff changeset	140 self._enqueue(Stream.END_NS, prefix or '')
1 821114ec4f69 Initial import. cmlenz parents: diff changeset	141
821114ec4f69 Initial import. cmlenz parents: diff changeset	142 def _handle_pi(self, target, data):
26 039fc5b87405 * Split out the XPath tests into a separate `unittest`-based file. cmlenz parents: 21 diff changeset	143 self._enqueue(Stream.PI, (target, data))
1 821114ec4f69 Initial import. cmlenz parents: diff changeset	144
821114ec4f69 Initial import. cmlenz parents: diff changeset	145 def _handle_comment(self, text):
26 039fc5b87405 * Split out the XPath tests into a separate `unittest`-based file. cmlenz parents: 21 diff changeset	146 self._enqueue(Stream.COMMENT, text)
1 821114ec4f69 Initial import. cmlenz parents: diff changeset	147
821114ec4f69 Initial import. cmlenz parents: diff changeset	148 def _handle_other(self, text):
821114ec4f69 Initial import. cmlenz parents: diff changeset	149 if text.startswith('&'):
821114ec4f69 Initial import. cmlenz parents: diff changeset	150 # deal with undefined entities
821114ec4f69 Initial import. cmlenz parents: diff changeset	151 try:
821114ec4f69 Initial import. cmlenz parents: diff changeset	152 text = unichr(htmlentitydefs.name2codepoint[text[1:-1]])
26 039fc5b87405 * Split out the XPath tests into a separate `unittest`-based file. cmlenz parents: 21 diff changeset	153 self._enqueue(Stream.TEXT, text)
1 821114ec4f69 Initial import. cmlenz parents: diff changeset	154 except KeyError:
21 eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	155 lineno, offset = self._getpos()
1 821114ec4f69 Initial import. cmlenz parents: diff changeset	156 raise expat.error("undefined entity %s: line %d, column %d" %
821114ec4f69 Initial import. cmlenz parents: diff changeset	157 (text, lineno, offset))
821114ec4f69 Initial import. cmlenz parents: diff changeset	158
821114ec4f69 Initial import. cmlenz parents: diff changeset	159
821114ec4f69 Initial import. cmlenz parents: diff changeset	160 def XML(text):
821114ec4f69 Initial import. cmlenz parents: diff changeset	161 return Stream(list(XMLParser(StringIO(text))))
821114ec4f69 Initial import. cmlenz parents: diff changeset	162
821114ec4f69 Initial import. cmlenz parents: diff changeset	163
21 eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	164 class HTMLParser(html.HTMLParser, object):
1 821114ec4f69 Initial import. cmlenz parents: diff changeset	165 """Parser for HTML input based on the Python `HTMLParser` module.
821114ec4f69 Initial import. cmlenz parents: diff changeset	166
821114ec4f69 Initial import. cmlenz parents: diff changeset	167 This class provides the same interface for generating stream events as
821114ec4f69 Initial import. cmlenz parents: diff changeset	168 `XMLParser`, and attempts to automatically balance tags.
26 039fc5b87405 * Split out the XPath tests into a separate `unittest`-based file. cmlenz parents: 21 diff changeset	169
039fc5b87405 * Split out the XPath tests into a separate `unittest`-based file. cmlenz parents: 21 diff changeset	170 The parsing is initiated by iterating over the parser object:
039fc5b87405 * Split out the XPath tests into a separate `unittest`-based file. cmlenz parents: 21 diff changeset	171
039fc5b87405 * Split out the XPath tests into a separate `unittest`-based file. cmlenz parents: 21 diff changeset	172 >>> parser = HTMLParser(StringIO('<UL compact><LI>Foo</UL>'))
039fc5b87405 * Split out the XPath tests into a separate `unittest`-based file. cmlenz parents: 21 diff changeset	173 >>> for kind, data, pos in parser:
039fc5b87405 * Split out the XPath tests into a separate `unittest`-based file. cmlenz parents: 21 diff changeset	174 ... print kind, data
039fc5b87405 * Split out the XPath tests into a separate `unittest`-based file. cmlenz parents: 21 diff changeset	175 START (u'ul', [(u'compact', u'compact')])
039fc5b87405 * Split out the XPath tests into a separate `unittest`-based file. cmlenz parents: 21 diff changeset	176 START (u'li', [])
039fc5b87405 * Split out the XPath tests into a separate `unittest`-based file. cmlenz parents: 21 diff changeset	177 TEXT Foo
039fc5b87405 * Split out the XPath tests into a separate `unittest`-based file. cmlenz parents: 21 diff changeset	178 END li
039fc5b87405 * Split out the XPath tests into a separate `unittest`-based file. cmlenz parents: 21 diff changeset	179 END ul
1 821114ec4f69 Initial import. cmlenz parents: diff changeset	180 """
821114ec4f69 Initial import. cmlenz parents: diff changeset	181
821114ec4f69 Initial import. cmlenz parents: diff changeset	182 _EMPTY_ELEMS = frozenset(['area', 'base', 'basefont', 'br', 'col', 'frame',
821114ec4f69 Initial import. cmlenz parents: diff changeset	183 'hr', 'img', 'input', 'isindex', 'link', 'meta',
821114ec4f69 Initial import. cmlenz parents: diff changeset	184 'param'])
821114ec4f69 Initial import. cmlenz parents: diff changeset	185
21 eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	186 def __init__(self, source, filename=None):
1 821114ec4f69 Initial import. cmlenz parents: diff changeset	187 html.HTMLParser.__init__(self)
821114ec4f69 Initial import. cmlenz parents: diff changeset	188 self.source = source
21 eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	189 self.filename = filename
eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	190 self._queue = []
1 821114ec4f69 Initial import. cmlenz parents: diff changeset	191 self._open_tags = []
821114ec4f69 Initial import. cmlenz parents: diff changeset	192
821114ec4f69 Initial import. cmlenz parents: diff changeset	193 def __iter__(self):
21 eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	194 try:
eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	195 bufsize = 4 * 1024 # 4K
eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	196 done = False
eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	197 while True:
eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	198 while not done and len(self._queue) == 0:
eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	199 data = self.source.read(bufsize)
eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	200 if data == '': # end of data
eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	201 self.close()
eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	202 done = True
eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	203 else:
eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	204 self.feed(data)
eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	205 for kind, data, pos in self._queue:
eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	206 yield kind, data, pos
eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	207 self._queue = []
eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	208 if done:
eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	209 open_tags = self._open_tags
eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	210 open_tags.reverse()
eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	211 for tag in open_tags:
eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	212 yield Stream.END, QName(tag), pos
eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	213 break
eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	214 except html.HTMLParseError, e:
eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	215 msg = '%s: line %d, column %d' % (e.msg, e.lineno, e.offset)
eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	216 if self.filename:
eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	217 msg += ', in %s' % self.filename
eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	218 raise ParseError(msg, self.filename, e.lineno, e.offset)
eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	219
26 039fc5b87405 * Split out the XPath tests into a separate `unittest`-based file. cmlenz parents: 21 diff changeset	220 def _enqueue(self, kind, data, pos=None):
039fc5b87405 * Split out the XPath tests into a separate `unittest`-based file. cmlenz parents: 21 diff changeset	221 if pos is None:
039fc5b87405 * Split out the XPath tests into a separate `unittest`-based file. cmlenz parents: 21 diff changeset	222 pos = self._getpos()
039fc5b87405 * Split out the XPath tests into a separate `unittest`-based file. cmlenz parents: 21 diff changeset	223 self._queue.append((kind, data, pos))
039fc5b87405 * Split out the XPath tests into a separate `unittest`-based file. cmlenz parents: 21 diff changeset	224
21 eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	225 def _getpos(self):
eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	226 lineno, column = self.getpos()
eca77129518a * Include paths are now interpreted relative to the path of the including template. Closes #3. cmlenz parents: 1 diff changeset	227 return (self.filename, lineno, column)
1 821114ec4f69 Initial import. cmlenz parents: diff changeset	228
821114ec4f69 Initial import. cmlenz parents: diff changeset	229 def handle_starttag(self, tag, attrib):
26 039fc5b87405 * Split out the XPath tests into a separate `unittest`-based file. cmlenz parents: 21 diff changeset	230 fixed_attrib = []
039fc5b87405 * Split out the XPath tests into a separate `unittest`-based file. cmlenz parents: 21 diff changeset	231 for name, value in attrib: # Fixup minimized attributes
039fc5b87405 * Split out the XPath tests into a separate `unittest`-based file. cmlenz parents: 21 diff changeset	232 if value is None:
039fc5b87405 * Split out the XPath tests into a separate `unittest`-based file. cmlenz parents: 21 diff changeset	233 value = name
039fc5b87405 * Split out the XPath tests into a separate `unittest`-based file. cmlenz parents: 21 diff changeset	234 fixed_attrib.append((name, unicode(value)))
039fc5b87405 * Split out the XPath tests into a separate `unittest`-based file. cmlenz parents: 21 diff changeset	235
039fc5b87405 * Split out the XPath tests into a separate `unittest`-based file. cmlenz parents: 21 diff changeset	236 self._enqueue(Stream.START, (QName(tag), Attributes(fixed_attrib)))
1 821114ec4f69 Initial import. cmlenz parents: diff changeset	237 if tag in self._EMPTY_ELEMS:
26 039fc5b87405 * Split out the XPath tests into a separate `unittest`-based file. cmlenz parents: 21 diff changeset	238 self._enqueue(Stream.END, QName(tag))
1 821114ec4f69 Initial import. cmlenz parents: diff changeset	239 else:
821114ec4f69 Initial import. cmlenz parents: diff changeset	240 self._open_tags.append(tag)
821114ec4f69 Initial import. cmlenz parents: diff changeset	241
821114ec4f69 Initial import. cmlenz parents: diff changeset	242 def handle_endtag(self, tag):
821114ec4f69 Initial import. cmlenz parents: diff changeset	243 if tag not in self._EMPTY_ELEMS:
821114ec4f69 Initial import. cmlenz parents: diff changeset	244 while self._open_tags:
821114ec4f69 Initial import. cmlenz parents: diff changeset	245 open_tag = self._open_tags.pop()
821114ec4f69 Initial import. cmlenz parents: diff changeset	246 if open_tag.lower() == tag.lower():
821114ec4f69 Initial import. cmlenz parents: diff changeset	247 break
26 039fc5b87405 * Split out the XPath tests into a separate `unittest`-based file. cmlenz parents: 21 diff changeset	248 self._enqueue(Stream.END, QName(open_tag))
039fc5b87405 * Split out the XPath tests into a separate `unittest`-based file. cmlenz parents: 21 diff changeset	249 self._enqueue(Stream.END, QName(tag))
1 821114ec4f69 Initial import. cmlenz parents: diff changeset	250
821114ec4f69 Initial import. cmlenz parents: diff changeset	251 def handle_data(self, text):
26 039fc5b87405 * Split out the XPath tests into a separate `unittest`-based file. cmlenz parents: 21 diff changeset	252 self._enqueue(Stream.TEXT, text)
1 821114ec4f69 Initial import. cmlenz parents: diff changeset	253
821114ec4f69 Initial import. cmlenz parents: diff changeset	254 def handle_charref(self, name):
26 039fc5b87405 * Split out the XPath tests into a separate `unittest`-based file. cmlenz parents: 21 diff changeset	255 self._enqueue(Stream.TEXT, Markup('&#%s;' % name))
1 821114ec4f69 Initial import. cmlenz parents: diff changeset	256
821114ec4f69 Initial import. cmlenz parents: diff changeset	257 def handle_entityref(self, name):
26 039fc5b87405 * Split out the XPath tests into a separate `unittest`-based file. cmlenz parents: 21 diff changeset	258 self._enqueue(Stream.TEXT, Markup('&%s;' % name))
1 821114ec4f69 Initial import. cmlenz parents: diff changeset	259
821114ec4f69 Initial import. cmlenz parents: diff changeset	260 def handle_pi(self, data):
821114ec4f69 Initial import. cmlenz parents: diff changeset	261 target, data = data.split(maxsplit=1)
821114ec4f69 Initial import. cmlenz parents: diff changeset	262 data = data.rstrip('?')
26 039fc5b87405 * Split out the XPath tests into a separate `unittest`-based file. cmlenz parents: 21 diff changeset	263 self._enqueue(Stream.PI, (target.strip(), data.strip()))
1 821114ec4f69 Initial import. cmlenz parents: diff changeset	264
821114ec4f69 Initial import. cmlenz parents: diff changeset	265 def handle_comment(self, text):
26 039fc5b87405 * Split out the XPath tests into a separate `unittest`-based file. cmlenz parents: 21 diff changeset	266 self._enqueue(Stream.COMMENT, text)
1 821114ec4f69 Initial import. cmlenz parents: diff changeset	267
821114ec4f69 Initial import. cmlenz parents: diff changeset	268
821114ec4f69 Initial import. cmlenz parents: diff changeset	269 def HTML(text):
821114ec4f69 Initial import. cmlenz parents: diff changeset	270 return Stream(list(HTMLParser(StringIO(text))))

Mercurial > genshi > genshi-test

annotate markup/input.py @ 66:822089ae65ce