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

Commita137d14

Browse files
kovidgoyalgsnedders
authored andcommitted
Preserve attribute order when parsing
1 parent0a885c6 commita137d14

File tree

7 files changed

+100
-92
lines changed

7 files changed

+100
-92
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: 21 additions & 83 deletions
Original file line numberDiff line numberDiff line change
@@ -1,8 +1,13 @@
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

@@ -17,6 +22,7 @@
1722
namespaces,
1823
htmlIntegrationPointElements,mathmlTextIntegrationPointElements,
1924
adjustForeignAttributesasadjustForeignAttributesMap,
25+
adjustMathMLAttributes,adjustSVGAttributes,
2026
E,
2127
ReparseException
2228
)
@@ -273,96 +279,18 @@ def normalizeToken(self, token):
273279
""" HTML5 specific normalizations to the token stream """
274280

275281
iftoken["type"]==tokenTypes["StartTag"]:
276-
token["data"]=dict(token["data"][::-1])
282+
token["data"]=OrderedDict(token['data'][::-1])
277283

278284
returntoken
279285

280286
defadjustMathMLAttributes(self,token):
281-
replacements= {"definitionurl":"definitionURL"}
282-
fork,vinreplacements.items():
283-
ifkintoken["data"]:
284-
token["data"][v]=token["data"][k]
285-
deltoken["data"][k]
287+
adjust_attributes(token,adjustMathMLAttributes)
286288

287289
defadjustSVGAttributes(self,token):
288-
replacements= {
289-
"attributename":"attributeName",
290-
"attributetype":"attributeType",
291-
"basefrequency":"baseFrequency",
292-
"baseprofile":"baseProfile",
293-
"calcmode":"calcMode",
294-
"clippathunits":"clipPathUnits",
295-
"contentscripttype":"contentScriptType",
296-
"contentstyletype":"contentStyleType",
297-
"diffuseconstant":"diffuseConstant",
298-
"edgemode":"edgeMode",
299-
"externalresourcesrequired":"externalResourcesRequired",
300-
"filterres":"filterRes",
301-
"filterunits":"filterUnits",
302-
"glyphref":"glyphRef",
303-
"gradienttransform":"gradientTransform",
304-
"gradientunits":"gradientUnits",
305-
"kernelmatrix":"kernelMatrix",
306-
"kernelunitlength":"kernelUnitLength",
307-
"keypoints":"keyPoints",
308-
"keysplines":"keySplines",
309-
"keytimes":"keyTimes",
310-
"lengthadjust":"lengthAdjust",
311-
"limitingconeangle":"limitingConeAngle",
312-
"markerheight":"markerHeight",
313-
"markerunits":"markerUnits",
314-
"markerwidth":"markerWidth",
315-
"maskcontentunits":"maskContentUnits",
316-
"maskunits":"maskUnits",
317-
"numoctaves":"numOctaves",
318-
"pathlength":"pathLength",
319-
"patterncontentunits":"patternContentUnits",
320-
"patterntransform":"patternTransform",
321-
"patternunits":"patternUnits",
322-
"pointsatx":"pointsAtX",
323-
"pointsaty":"pointsAtY",
324-
"pointsatz":"pointsAtZ",
325-
"preservealpha":"preserveAlpha",
326-
"preserveaspectratio":"preserveAspectRatio",
327-
"primitiveunits":"primitiveUnits",
328-
"refx":"refX",
329-
"refy":"refY",
330-
"repeatcount":"repeatCount",
331-
"repeatdur":"repeatDur",
332-
"requiredextensions":"requiredExtensions",
333-
"requiredfeatures":"requiredFeatures",
334-
"specularconstant":"specularConstant",
335-
"specularexponent":"specularExponent",
336-
"spreadmethod":"spreadMethod",
337-
"startoffset":"startOffset",
338-
"stddeviation":"stdDeviation",
339-
"stitchtiles":"stitchTiles",
340-
"surfacescale":"surfaceScale",
341-
"systemlanguage":"systemLanguage",
342-
"tablevalues":"tableValues",
343-
"targetx":"targetX",
344-
"targety":"targetY",
345-
"textlength":"textLength",
346-
"viewbox":"viewBox",
347-
"viewtarget":"viewTarget",
348-
"xchannelselector":"xChannelSelector",
349-
"ychannelselector":"yChannelSelector",
350-
"zoomandpan":"zoomAndPan"
351-
}
352-
fororiginalNameinlist(token["data"].keys()):
353-
iforiginalNameinreplacements:
354-
svgName=replacements[originalName]
355-
token["data"][svgName]=token["data"][originalName]
356-
deltoken["data"][originalName]
290+
adjust_attributes(token,adjustSVGAttributes)
357291

358292
defadjustForeignAttributes(self,token):
359-
replacements=adjustForeignAttributesMap
360-
361-
fororiginalNameintoken["data"].keys():
362-
iforiginalNameinreplacements:
363-
foreignName=replacements[originalName]
364-
token["data"][foreignName]=token["data"][originalName]
365-
deltoken["data"][originalName]
293+
adjust_attributes(token,adjustForeignAttributesMap)
366294

367295
defreparseTokenNormal(self,token):
368296
# pylint:disable=unused-argument
@@ -2791,6 +2719,16 @@ def processEndTag(self, token):
27912719
}
27922720

27932721

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+
27942732
defimpliedTagToken(name,type="EndTag",attributes=None,
27952733
selfClosing=False):
27962734
ifattributesisNone:

‎html5lib/tests/test_parser2.py‎

Lines changed: 7 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -49,3 +49,10 @@ def test_namespace_html_elements_1_etree():
4949

5050
deftest_unicode_file():
5151
assertparse(io.StringIO("a"))isnotNone
52+
53+
54+
deftest_duplicate_attribute():
55+
# This is here because we impl it in parser and not tokenizer
56+
doc=parse('<p class=a class=b>')
57+
el=doc[1][0]
58+
assertel.get("class")=="a"

‎html5lib/utils.py‎

Lines changed: 4 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -1,5 +1,6 @@
11
from __future__importabsolute_import,division,unicode_literals
22

3+
importsys
34
fromtypesimportModuleType
45

56
fromsiximporttext_type
@@ -12,9 +13,11 @@
1213

1314
__all__= ["default_etree","MethodDispatcher","isSurrogatePair",
1415
"surrogatePairToCodepoint","moduleFactoryFactory",
15-
"supports_lone_surrogates"]
16+
"supports_lone_surrogates","PY27"]
1617

1718

19+
PY27=sys.version_info[0]==2andsys.version_info[1]>=7
20+
1821
# Platforms not supporting lone surrogates (\uD800-\uDFFF) should be
1922
# caught by the below test. In general this would be any platform
2023
# using UTF-16 as its encoding of unicode strings, such as

‎requirements-optional.txt‎

Lines changed: 0 additions & 4 deletions
Original file line numberDiff line numberDiff line change
@@ -15,7 +15,3 @@ lxml ; platform_python_implementation == 'CPython'
1515
# DATrie can be used in place of our Python trie implementation for
1616
# slightly better parsing performance.
1717
datrie ; platform_python_implementation == 'CPython'
18-
19-
# Can be used to force attributes to be serialized in alphabetical
20-
# order.
21-
ordereddict ; python_version < '2.7'

‎requirements.txt‎

Lines changed: 1 addition & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -1,2 +1,3 @@
11
six
22
webencodings
3+
ordereddict ;python_version<'2.7'

0 commit comments

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