Movatterモバイル変換

[0]ホーム
URL:

Skip to content

Navigation Menu

Sign in
Appearance settings

Search code, repositories, users, issues, pull requests...

Provide feedback

We read every piece of feedback, and take your input very seriously.

Saved searches

Use saved searches to filter your results more quickly

 Sign up
Appearance settings

Commit5288737

Browse files
committed
Merge pull request#245 from gsnedders/calibre-selected-1
Selected patches from Calibre; r=nobody!
2 parents143b0d4 +761f3ab commit5288737

File tree

8 files changed

+167
-126
lines changed

8 files changed

+167
-126
lines changed

‎README.rst‎

Lines changed: 0 additions & 4 deletions
Original file line numberDiff line numberDiff line change
@@ -116,10 +116,6 @@ functionality:
116116
- ``chardet`` can be used as a fallback when character encoding cannot
117117
be determined.
118118

119-
- ``ordereddict`` can be used under Python 2.6
120-
(``collections.OrderedDict`` is used instead on later versions) to
121-
serialize attributes in alphabetical order.
122-
123119

124120
Bugs
125121
----

‎html5lib/constants.py‎

Lines changed: 67 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -437,6 +437,73 @@
437437
(namespaces["mathml"],"mtext")
438438
])
439439

440+
adjustSVGAttributes= {
441+
"attributename":"attributeName",
442+
"attributetype":"attributeType",
443+
"basefrequency":"baseFrequency",
444+
"baseprofile":"baseProfile",
445+
"calcmode":"calcMode",
446+
"clippathunits":"clipPathUnits",
447+
"contentscripttype":"contentScriptType",
448+
"contentstyletype":"contentStyleType",
449+
"diffuseconstant":"diffuseConstant",
450+
"edgemode":"edgeMode",
451+
"externalresourcesrequired":"externalResourcesRequired",
452+
"filterres":"filterRes",
453+
"filterunits":"filterUnits",
454+
"glyphref":"glyphRef",
455+
"gradienttransform":"gradientTransform",
456+
"gradientunits":"gradientUnits",
457+
"kernelmatrix":"kernelMatrix",
458+
"kernelunitlength":"kernelUnitLength",
459+
"keypoints":"keyPoints",
460+
"keysplines":"keySplines",
461+
"keytimes":"keyTimes",
462+
"lengthadjust":"lengthAdjust",
463+
"limitingconeangle":"limitingConeAngle",
464+
"markerheight":"markerHeight",
465+
"markerunits":"markerUnits",
466+
"markerwidth":"markerWidth",
467+
"maskcontentunits":"maskContentUnits",
468+
"maskunits":"maskUnits",
469+
"numoctaves":"numOctaves",
470+
"pathlength":"pathLength",
471+
"patterncontentunits":"patternContentUnits",
472+
"patterntransform":"patternTransform",
473+
"patternunits":"patternUnits",
474+
"pointsatx":"pointsAtX",
475+
"pointsaty":"pointsAtY",
476+
"pointsatz":"pointsAtZ",
477+
"preservealpha":"preserveAlpha",
478+
"preserveaspectratio":"preserveAspectRatio",
479+
"primitiveunits":"primitiveUnits",
480+
"refx":"refX",
481+
"refy":"refY",
482+
"repeatcount":"repeatCount",
483+
"repeatdur":"repeatDur",
484+
"requiredextensions":"requiredExtensions",
485+
"requiredfeatures":"requiredFeatures",
486+
"specularconstant":"specularConstant",
487+
"specularexponent":"specularExponent",
488+
"spreadmethod":"spreadMethod",
489+
"startoffset":"startOffset",
490+
"stddeviation":"stdDeviation",
491+
"stitchtiles":"stitchTiles",
492+
"surfacescale":"surfaceScale",
493+
"systemlanguage":"systemLanguage",
494+
"tablevalues":"tableValues",
495+
"targetx":"targetX",
496+
"targety":"targetY",
497+
"textlength":"textLength",
498+
"viewbox":"viewBox",
499+
"viewtarget":"viewTarget",
500+
"xchannelselector":"xChannelSelector",
501+
"ychannelselector":"yChannelSelector",
502+
"zoomandpan":"zoomAndPan"
503+
}
504+
505+
adjustMathMLAttributes= {"definitionurl":"definitionURL"}
506+
440507
adjustForeignAttributes= {
441508
"xlink:actuate": ("xlink","actuate",namespaces["xlink"]),
442509
"xlink:arcrole": ("xlink","arcrole",namespaces["xlink"]),

‎html5lib/html5parser.py‎

Lines changed: 36 additions & 105 deletions
Original file line numberDiff line numberDiff line change
@@ -1,24 +1,31 @@
11
from __future__importabsolute_import,division,unicode_literals
2-
fromsiximportwith_metaclass
2+
fromsiximportwith_metaclass,viewkeys,PY3
33

44
importtypes
55

6+
try:
7+
fromcollectionsimportOrderedDict
8+
exceptImportError:
9+
fromordereddictimportOrderedDict
10+
611
from .importinputstream
712
from .importtokenizer
813

914
from .importtreebuilders
1015
from .treebuilders._baseimportMarker
1116

1217
from .importutils
13-
from .importconstants
14-
from .constantsimportspaceCharacters,asciiUpper2Lower
15-
from .constantsimportspecialElements
16-
from .constantsimportheadingElements
17-
from .constantsimportcdataElements,rcdataElements
18-
from .constantsimporttokenTypes,ReparseException,namespaces
19-
from .constantsimporthtmlIntegrationPointElements,mathmlTextIntegrationPointElements
20-
from .constantsimportadjustForeignAttributesasadjustForeignAttributesMap
21-
from .constantsimportE
18+
from .constantsimport (
19+
spaceCharacters,asciiUpper2Lower,
20+
specialElements,headingElements,cdataElements,rcdataElements,
21+
tokenTypes,tagTokenTypes,
22+
namespaces,
23+
htmlIntegrationPointElements,mathmlTextIntegrationPointElements,
24+
adjustForeignAttributesasadjustForeignAttributesMap,
25+
adjustMathMLAttributes,adjustSVGAttributes,
26+
E,
27+
ReparseException
28+
)
2229

2330

2431
defparse(doc,treebuilder="etree",encoding=None,
@@ -272,96 +279,18 @@ def normalizeToken(self, token):
272279
""" HTML5 specific normalizations to the token stream """
273280

274281
iftoken["type"]==tokenTypes["StartTag"]:
275-
token["data"]=dict(token["data"][::-1])
282+
token["data"]=OrderedDict(token['data'][::-1])
276283

277284
returntoken
278285

279286
defadjustMathMLAttributes(self,token):
280-
replacements= {"definitionurl":"definitionURL"}
281-
fork,vinreplacements.items():
282-
ifkintoken["data"]:
283-
token["data"][v]=token["data"][k]
284-
deltoken["data"][k]
287+
adjust_attributes(token,adjustMathMLAttributes)
285288

286289
defadjustSVGAttributes(self,token):
287-
replacements= {
288-
"attributename":"attributeName",
289-
"attributetype":"attributeType",
290-
"basefrequency":"baseFrequency",
291-
"baseprofile":"baseProfile",
292-
"calcmode":"calcMode",
293-
"clippathunits":"clipPathUnits",
294-
"contentscripttype":"contentScriptType",
295-
"contentstyletype":"contentStyleType",
296-
"diffuseconstant":"diffuseConstant",
297-
"edgemode":"edgeMode",
298-
"externalresourcesrequired":"externalResourcesRequired",
299-
"filterres":"filterRes",
300-
"filterunits":"filterUnits",
301-
"glyphref":"glyphRef",
302-
"gradienttransform":"gradientTransform",
303-
"gradientunits":"gradientUnits",
304-
"kernelmatrix":"kernelMatrix",
305-
"kernelunitlength":"kernelUnitLength",
306-
"keypoints":"keyPoints",
307-
"keysplines":"keySplines",
308-
"keytimes":"keyTimes",
309-
"lengthadjust":"lengthAdjust",
310-
"limitingconeangle":"limitingConeAngle",
311-
"markerheight":"markerHeight",
312-
"markerunits":"markerUnits",
313-
"markerwidth":"markerWidth",
314-
"maskcontentunits":"maskContentUnits",
315-
"maskunits":"maskUnits",
316-
"numoctaves":"numOctaves",
317-
"pathlength":"pathLength",
318-
"patterncontentunits":"patternContentUnits",
319-
"patterntransform":"patternTransform",
320-
"patternunits":"patternUnits",
321-
"pointsatx":"pointsAtX",
322-
"pointsaty":"pointsAtY",
323-
"pointsatz":"pointsAtZ",
324-
"preservealpha":"preserveAlpha",
325-
"preserveaspectratio":"preserveAspectRatio",
326-
"primitiveunits":"primitiveUnits",
327-
"refx":"refX",
328-
"refy":"refY",
329-
"repeatcount":"repeatCount",
330-
"repeatdur":"repeatDur",
331-
"requiredextensions":"requiredExtensions",
332-
"requiredfeatures":"requiredFeatures",
333-
"specularconstant":"specularConstant",
334-
"specularexponent":"specularExponent",
335-
"spreadmethod":"spreadMethod",
336-
"startoffset":"startOffset",
337-
"stddeviation":"stdDeviation",
338-
"stitchtiles":"stitchTiles",
339-
"surfacescale":"surfaceScale",
340-
"systemlanguage":"systemLanguage",
341-
"tablevalues":"tableValues",
342-
"targetx":"targetX",
343-
"targety":"targetY",
344-
"textlength":"textLength",
345-
"viewbox":"viewBox",
346-
"viewtarget":"viewTarget",
347-
"xchannelselector":"xChannelSelector",
348-
"ychannelselector":"yChannelSelector",
349-
"zoomandpan":"zoomAndPan"
350-
}
351-
fororiginalNameinlist(token["data"].keys()):
352-
iforiginalNameinreplacements:
353-
svgName=replacements[originalName]
354-
token["data"][svgName]=token["data"][originalName]
355-
deltoken["data"][originalName]
290+
adjust_attributes(token,adjustSVGAttributes)
356291

357292
defadjustForeignAttributes(self,token):
358-
replacements=adjustForeignAttributesMap
359-
360-
fororiginalNameintoken["data"].keys():
361-
iforiginalNameinreplacements:
362-
foreignName=replacements[originalName]
363-
token["data"][foreignName]=token["data"][originalName]
364-
deltoken["data"][originalName]
293+
adjust_attributes(token,adjustForeignAttributesMap)
365294

366295
defreparseTokenNormal(self,token):
367296
# pylint:disable=unused-argument
@@ -434,7 +363,7 @@ def getPhases(debug):
434363
deflog(function):
435364
"""Logger that records which phase processes each token"""
436365
type_names=dict((value,key)forkey,valuein
437-
constants.tokenTypes.items())
366+
tokenTypes.items())
438367

439368
defwrapped(self,*args,**kwargs):
440369
iffunction.__name__.startswith("process")andlen(args)>0:
@@ -443,7 +372,7 @@ def wrapped(self, *args, **kwargs):
443372
info= {"type":type_names[token['type']]}
444373
except:
445374
raise
446-
iftoken['type']inconstants.tagTokenTypes:
375+
iftoken['type']intagTokenTypes:
447376
info["name"]=token['name']
448377

449378
self.parser.log.append((self.parser.tokenizer.state.__name__,
@@ -1022,17 +951,9 @@ def __init__(self, parser, tree):
1022951
self.endTagHandler.default=self.endTagOther
1023952

1024953
defisMatchingFormattingElement(self,node1,node2):
1025-
ifnode1.name!=node2.nameornode1.namespace!=node2.namespace:
1026-
returnFalse
1027-
eliflen(node1.attributes)!=len(node2.attributes):
1028-
returnFalse
1029-
else:
1030-
attributes1=sorted(node1.attributes.items())
1031-
attributes2=sorted(node2.attributes.items())
1032-
forattr1,attr2inzip(attributes1,attributes2):
1033-
ifattr1!=attr2:
1034-
returnFalse
1035-
returnTrue
954+
return (node1.name==node2.nameand
955+
node1.namespace==node2.namespaceand
956+
node1.attributes==node2.attributes)
1036957

1037958
# helper
1038959
defaddFormattingElement(self,token):
@@ -2798,6 +2719,16 @@ def processEndTag(self, token):
27982719
}
27992720

28002721

2722+
defadjust_attributes(token,replacements):
2723+
ifPY3orutils.PY27:
2724+
needs_adjustment=viewkeys(token['data'])&viewkeys(replacements)
2725+
else:
2726+
needs_adjustment=frozenset(token['data'])&frozenset(replacements)
2727+
ifneeds_adjustment:
2728+
token['data']=OrderedDict((replacements.get(k,k),v)
2729+
fork,vintoken['data'].items())
2730+
2731+
28012732
defimpliedTagToken(name,type="EndTag",attributes=None,
28022733
selfClosing=False):
28032734
ifattributesisNone:

‎html5lib/tests/test_parser2.py‎

Lines changed: 40 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -1,11 +1,13 @@
11
from __future__importabsolute_import,division,unicode_literals
22

3+
fromsiximportPY2,text_type
4+
35
importio
46

57
from .importsupport# noqa
68

79
fromhtml5lib.constantsimportnamespaces
8-
fromhtml5libimportparse
10+
fromhtml5libimportparse,HTMLParser
911

1012

1113
# tests that aren't autogenerated from text files
@@ -49,3 +51,40 @@ def test_namespace_html_elements_1_etree():
4951

5052
deftest_unicode_file():
5153
assertparse(io.StringIO("a"))isnotNone
54+
55+
56+
deftest_duplicate_attribute():
57+
# This is here because we impl it in parser and not tokenizer
58+
doc=parse('<p class=a class=b>')
59+
el=doc[1][0]
60+
assertel.get("class")=="a"
61+
62+
63+
deftest_debug_log():
64+
parser=HTMLParser(debug=True)
65+
parser.parse("<!doctype html><title>a</title><p>b<script>c</script>d</p>e")
66+
67+
expected= [('dataState','InitialPhase','InitialPhase','processDoctype', {'type':'Doctype'}),
68+
('dataState','BeforeHtmlPhase','BeforeHtmlPhase','processStartTag', {'name':'title','type':'StartTag'}),
69+
('dataState','BeforeHeadPhase','BeforeHeadPhase','processStartTag', {'name':'title','type':'StartTag'}),
70+
('dataState','InHeadPhase','InHeadPhase','processStartTag', {'name':'title','type':'StartTag'}),
71+
('rcdataState','TextPhase','TextPhase','processCharacters', {'type':'Characters'}),
72+
('dataState','TextPhase','TextPhase','processEndTag', {'name':'title','type':'EndTag'}),
73+
('dataState','InHeadPhase','InHeadPhase','processStartTag', {'name':'p','type':'StartTag'}),
74+
('dataState','AfterHeadPhase','AfterHeadPhase','processStartTag', {'name':'p','type':'StartTag'}),
75+
('dataState','InBodyPhase','InBodyPhase','processStartTag', {'name':'p','type':'StartTag'}),
76+
('dataState','InBodyPhase','InBodyPhase','processCharacters', {'type':'Characters'}),
77+
('dataState','InBodyPhase','InBodyPhase','processStartTag', {'name':'script','type':'StartTag'}),
78+
('dataState','InBodyPhase','InHeadPhase','processStartTag', {'name':'script','type':'StartTag'}),
79+
('scriptDataState','TextPhase','TextPhase','processCharacters', {'type':'Characters'}),
80+
('dataState','TextPhase','TextPhase','processEndTag', {'name':'script','type':'EndTag'}),
81+
('dataState','InBodyPhase','InBodyPhase','processCharacters', {'type':'Characters'}),
82+
('dataState','InBodyPhase','InBodyPhase','processEndTag', {'name':'p','type':'EndTag'}),
83+
('dataState','InBodyPhase','InBodyPhase','processCharacters', {'type':'Characters'})]
84+
85+
ifPY2:
86+
fori,loginenumerate(expected):
87+
log= [x.encode("ascii")ifisinstance(x,text_type)elsexforxinlog]
88+
expected[i]=tuple(log)
89+
90+
assertparser.log==expected

0 commit comments

Comments
 (0)
[8]ページ先頭
©2009-2025 Movatter.jp