Commit75cf697

committed

Reintroduce the old sanitizer testsuite from html5lib-tests

This is imported into this repo as its expectations are very muchimplementation dependent, with expectations amended to match ouractual behaviour.

1 parenta2917e9 commit75cf697Copy full SHA for 75cf697

File tree

+494

-5

lines changed

+494

-5

lines changed

Lines changed: 6 additions & 5 deletions

Original file line number	Diff line number	Diff line change
`@@ -184,6 +184,12 @@ def serialize(self, treewalker, encoding=None):`
`184`	`184`	`ifencodingandself.inject_meta_charset:`
`185`	`185`	`from ..filters.inject_meta_charsetimportFilter`
`186`	`186`	`treewalker=Filter(treewalker,encoding)`
	`187`	`+# Alphabetical attributes is here under the assumption that none of`
	`188`	`+# the later filters add or change order of attributes; it needs to be`
	`189`	`+# before the sanitizer so escaped elements come out correctly`
	`190`	`+ifself.alphabetical_attributes:`
	`191`	`+from ..filters.alphabeticalattributesimportFilter`
	`192`	`+treewalker=Filter(treewalker)`
`187`	`193`	`# WhitespaceFilter should be used before OptionalTagFilter`
`188`	`194`	`# for maximum efficiently of this latter filter`
`189`	`195`	`ifself.strip_whitespace:`
`@@ -195,11 +201,6 @@ def serialize(self, treewalker, encoding=None):`
`195`	`201`	`ifself.omit_optional_tags:`
`196`	`202`	`from ..filters.optionaltagsimportFilter`
`197`	`203`	`treewalker=Filter(treewalker)`
`198`		`-# Alphabetical attributes must be last, as other filters`
`199`		`-# could add attributes and alter the order`
`200`		`-ifself.alphabetical_attributes:`
`201`		`-from ..filters.alphabeticalattributesimportFilter`
`202`		`-treewalker=Filter(treewalker)`
`203`	`204`
`204`	`205`	`fortokenintreewalker:`
`205`	`206`	`type=token["type"]`

Lines changed: 5 additions & 0 deletions

Original file line number	Diff line number	Diff line change
`@@ -2,11 +2,13 @@`
`2`	`2`
`3`	`3`	`from .tree_constructionimportTreeConstructionFile`
`4`	`4`	`from .tokenizerimportTokenizerFile`
	`5`	`+from .sanitizerimportSanitizerFile`
`5`	`6`
`6`	`7`	`_dir=os.path.abspath(os.path.dirname(__file__))`
`7`	`8`	`_testdata=os.path.join(_dir,"testdata")`
`8`	`9`	`_tree_construction=os.path.join(_testdata,"tree-construction")`
`9`	`10`	`_tokenizer=os.path.join(_testdata,"tokenizer")`
	`11`	`+_sanitizer_testdata=os.path.join(_dir,"sanitizer-testdata")`
`10`	`12`
`11`	`13`
`12`	`14`	`defpytest_collectstart():`
`@@ -24,3 +26,6 @@ def pytest_collect_file(path, parent):`
`24`	`26`	`elifdir==_tokenizer:`
`25`	`27`	`ifpath.ext==".test":`
`26`	`28`	`returnTokenizerFile(path,parent)`
	`29`	`+elifdir==_sanitizer_testdata:`
	`30`	`+ifpath.ext==".dat":`
	`31`	`+returnSanitizerFile(path,parent)`

Comments

(0)