From f52f3db6539ddde6f8c74165efc292a1121486b7 Mon Sep 17 00:00:00 2001 From: Subramanya Sastry Date: Thu, 2 Jan 2014 15:01:32 -0600 Subject: [PATCH] Sync up with Parsoid parserTests. This now aligns with Parsoid commit 97bb35b73a4e4779b9be0bf2aedc351f539b0a0d Change-Id: Id9e18cd6f33c6f71d0dfb57bd854a98d81b5ad84 --- tests/parser/parserTests.txt | 779 +++++++++++++++++++++++++++++++---- 1 file changed, 689 insertions(+), 90 deletions(-) diff --git a/tests/parser/parserTests.txt b/tests/parser/parserTests.txt index cc935da8f0..ce2c73406a 100644 --- a/tests/parser/parserTests.txt +++ b/tests/parser/parserTests.txt @@ -146,6 +146,12 @@ Template:Bullet * Bar !!endarticle +!!article +Template:OpenTable +!!text +{| +!!endarticle + ### ### Basic tests ### @@ -916,27 +922,36 @@ Non-html5 tags should be accepted !! test Non-word characters don't terminate tag names (bug 17663, 40670, 52022) !! input - doesn't work! - - doesn't work! + doesn't work! - works fine + doesn't work! -foo + works fine s.foo !! result -

<b→> doesn't work! </b> -

<bä> doesn't work! </b> -

<boo> works fine </b> -

<s.foo>foo</s> +

<b→> doesn't work! </b→> +

<bä> doesn't work! </bä> +

<boo> works fine </boo>

<s.foo>s.foo</s.foo>

<sub-ID#1>

!! end +!! test +Isolated close tags should be treated as literal text (bug 52760) +!! input + + +s +!! result +

</b> +

<s.foo>s</s> +

+!! end + ### ### Special characters ### @@ -1336,6 +1351,22 @@ b !! end +## PHP parser emits output which is broken +!! test +Unclosed HTML p-tags should be handled properly +!! options +parsoid +!! input +

foo

+a + +b +!! result +

foo

+

a

+

b

+!! end + ### ### Preformatted text ### @@ -1589,6 +1620,46 @@ Entities inside 
 

 !! end
 
+!! test
+Empty pre; pre inside other HTML tags (bug 54946)
+!! input
+a
+
+
+foo
+

+

+!! result
+
a
+

+
+foo
+

+

+
+!! end
+
+!! test
+HTML pre followed by indent-pre
+!! input
+
foo

+ bar
+!! result
+
foo

+
bar
+

+!! end
+
+!!test
+Block tag pre
+!!options
+parsoid
+!!input
+
foo

+!!result
+
foo

+!!end
+
 !!test
 Templates: Indent-Pre: 1a. Templates that break a line should suppress 
 !!input
@@ -1708,6 +1779,25 @@ Templates: Indent-Pre: 1f: Wrapping should be based on expanded content
 

 !!end
 
+!! test
+Things that look like 
 tags aren't treated as such
+!! input
+Barack Obama  of the United States
+!! result
+
Barack Obama <President> of the United States
+

+!! end
+
+!! test
+Parsoid: handle pre with space after attribute
+!! options
+parsoid=wt2html
+!! input
+
{{echo|foo}}

+!! result
+
{{echo|foo}}

+!! end
+
 # TODO / maybe: fix wt2wt for this
 !! test
 Parsoid: Don't paragraph-wrap fosterable content
@@ -1928,6 +2018,17 @@ parsoid=html2wt,wt2wt
 
bar

 !! end
 
+## Bug 56733
+!! test
+Templates parameters with special tokenizing behavior dont get modified because of arg escaping
+!! options
+parsoid
+!! input
+{{echo|a : b}}
+!! result
+
a : b

+!! end
+
 ###
 ### Parsoid-centric tests for testing RT edge cases for pre
 ###
@@ -2042,18 +2143,30 @@ c
 !!test
 3a. Indent-Pre and block tags (single-line html)
 !!input
- 
 foo 

- 
 foo 

- 
 foo 

+ a 
 foo 

+ b 
 foo 

+ c 
 foo 

   foo 
 !!result
- 
 foo 

- 
 foo 

- 
 foo 

+ a 
 foo 

+ b 
 foo 

+ c 
 foo 

 
 foo 
 

 !!end
 
+!!test
+3c. Indent-Pre and block tags (multi-line html)
+!!input
+ a foo
+ b 
 foo 

+!!result
+
a foo
+

+ b 
 foo 

+
+!!end
+
 !!test
 3b. Indent-Pre and block tags (pre-content on separate line)
 !!input
@@ -2137,11 +2250,11 @@ foo
 
 !!end
 
+## NOTE: the leading white-space chars on empty line are significant
 !! test
-5. White-space in indent-pre
-NOTE: the white-space char on 2nd line is significant
+5a. White-space in indent-pre
 !! input
- a

+ a

  
  b
 !! result
@@ -2151,6 +2264,39 @@ b
 

 !! end
 
+## NOTE: the leading white-space chars on empty line are significant
+!! test
+5b. White-space in indent-pre
+!! input
+ a
+ 
+ b
+ 
+ 
+ c
+!! result
+
a
+
+b
+
+
+c
+

+!! end
+
+!! test
+5c. White-space in indent-pre
+!! input
+ ''a''
+  ''b''
+   ''c''
+!! result
+
a
+ b
+  c
+

+!! end
+
 !! test
 6. Pre-blocks should extend across lines with leading WS even when there is no wrappable content
 !! input
@@ -3935,6 +4081,83 @@ http://example.com
 
http://example.com

 !! end
 
+!! test
+Parenthesis in external links, especially URL links
+!! options
+php
+!! input 
+http://example.com)
+
+http://example.com/test)
+
+http://example.com/(test)
+
+http://example.com/((test)
+
+(http://example.com/(test))
+
+(http://example.com/(test)))))
+
+http://example.com/a)b
+
+[http://example.com) foo]
+!! result
+
http://example.com)
+
http://example.com/test)
+
http://example.com/(test)
+
http://example.com/((test)
+
(http://example.com/(test))
+
(http://example.com/(test)))))
+
http://example.com/a)b
+
foo
+

+!! end
+
+!! test
+Parenthesis in external links, especially URL links (Parsoid)
+!! options
+parsoid
+!! input 
+http://example.com)
+
+http://example.com/test)
+
+http://example.com/(test)
+
+http://example.com/((test)
+
+(http://example.com/(test))
+
+(http://example.com/(test)))))
+
+http://example.com/a)b
+
+[http://example.com) foo]
+!! result
+
http://example.com)

+
http://example.com/test)

+
http://example.com/(test)

+
http://example.com/((test)

+
(http://example.com/(test))

+
(http://example.com/(test)))))

+
http://example.com/a)b

+
foo

+!! end
+
+!! test
+Parenthesis in external links, w/ transclusion or comment
+!! options
+parsoid
+!! input
+(http://example.com/{{echo|hi}})
+
+(http://example.com)
+!! result
+
(http://example.com/hi)

+
+
(http://example.com)

+!! end
+
 ###
 ### Quotes
 ###
@@ -4048,16 +4271,71 @@ php
 !! input
 {||}
 !! result
+
 !! end
+
 # Parsoid team replies: empty table tags are legal in HTML5
 !! test
 A table with no data. (parsoid)
 !! options
-parsoid
+parsoid=wt2html
 !! input
 {||}
 !! result
 

+
+!! end
+
+!! test
+A table with stray table end tags on start tag line (wt2html)
+!! options
+parsoid=wt2html
+!! input
+{|style="color: red;"|}
+
+{|style="color: red;" |}
+|foo
+|}
+
+{|style="color: red;"|} id="foo"
+|foo
+|}
+
+{|style="color: red;" |} id="foo"
+|foo
+|}
+!! result
+

+
+
+
+
+
+
foo

+
+
+
+
+
+
foo

+
+
+
+
+
+
foo

+
+!! end
+
+!! test
+A table with no data (take 2) (parsoid)
+!! options
+parsoid
+!! input
+{|
+|}
+!! result
+

 !! end
 
 # A table with nothing but a caption is invalid XHTML, we might want to render
@@ -4757,6 +5035,18 @@ parsoid=html2wt
 bazquux
 !! end
 
+!! test
+Parsoid: newline inducing block nodes don't suppress 
+!! options
+parsoid=html2wt
+!! input
+ a
+
+= foo =
+!! result
+ a
foo

+!! end
+
 ###
 ### Internal links
 ###
@@ -4899,6 +5189,17 @@ Piped link with 3 brackets
 

 !! end
 
+!! test
+Piped link with extlink-like text
+!! input
+[[Main Page|[bar]]]
+[[Main Page|This is a [bar]]]
+!! result
+
[bar]
+This is a [bar]
+

+!! end
+
 !! test
 Link with multiple pipes
 !! input
@@ -5451,6 +5752,25 @@ parsoid
 
en:Foo

 !! end
 
+!! test
+Interwiki links that cannot be represented in wiki syntax
+!! options
+parsoid
+!! input
+[[meatball:ok]]
+[[meatball:ok#foo|ok with fragment]]
+[[meatball:ok_as_well?|ok ending with ? mark]]
+[http://de.wikipedia.org/wiki/Foo?action=history has query]
+[http://de.wikipedia.org/wiki/#foo is just fragment]
+
+!! result
+
meatball:ok
+ok with fragment
+ok ending with ? mark
+has query
+is just fragment

+!! end
+
 
 ###
 ### Interlanguage links
@@ -7972,6 +8292,30 @@ parsoid
 

 !!end
 
+!!test
+4. includeonly in table attributes
+!!input
+{|
+|- 
+|-
+|a
+
+|- 
+|-
+|b
+
+|}
+!!result
+
+
+
+
+
+
a
+

+
+!!end
+
 ###
 ### Testing parsing of templates where a template arg
 ### has the same name as the template itself.
@@ -9393,6 +9737,36 @@ parsoid
 
Caption text

 !! end
 
+!! test
+Image with caption, bug 53312 #1 (parsoid)
+!! options
+parsoid
+!! input
+[[File:Foobar.jpg|right|Caption page stuff]]
+!! result
+
Caption page stuff

+!! end
+
+!! test
+Image with caption, bug 53312 #2 (parsoid)
+!! options
+parsoid
+!! input
+[[File:Foobar.jpg|right|Caption page=]]
+!! result
+
Caption page=

+!! end
+
+!! test
+Image with caption, bug 53312 #3 (parsoid)
+!! options
+parsoid
+!! input
+[[File:Foobar.jpg|right|Caption page=stuff]]
+!! result
+
Caption page=stuff

+!! end
+
 !! test
 Image with empty attribute (php)
 !! options
@@ -9840,9 +10214,9 @@ Thumbnail image caption with a free URL and explicit alt
 SVG thumbnails with no language set
 !! options
 !! input
-[[File:Foobar.svg|thumb|width=200]]
+[[File:Foobar.svg|thumb|caption]]
 !! result
-
  
width=200

+
  
caption

 
 !! end
 
@@ -9850,9 +10224,9 @@ SVG thumbnails with no language set
 SVG thumbnails with language de
 !! options
 !! input
-[[File:Foobar.svg|thumb|width=200|lang=de]]
+[[File:Foobar.svg|thumb|caption|lang=de]]
 !! result
-
  
width=200

+
  
caption

 
 !! end
 
@@ -9860,7 +10234,7 @@ SVG thumbnails with language de
 SVG thumbnails with invalid language code
 !! options
 !! input
-[[File:Foobar.svg|thumb|width=200|lang=invalid.language.code]]
+[[File:Foobar.svg|thumb|caption|lang=invalid.language.code]]
 !! result
 
  
lang=invalid.language.code

 
@@ -10208,22 +10582,22 @@ parsoid
 #

 
 !! test
-Caption with a template in it
+Parsoid-specific image handling - caption with a template in it
 !! options
 parsoid
 !! input
-[[Image:Foobar.jpg|thumb|200px|This caption has a {{echo|transclusion}} in it.]]
+[[File:Foobar.jpg|thumb|200x200px|This caption has a {{echo|transclusion}} in it.]]
 !! result
 
This caption has a transclusion in it.

 !! end
 
 !! test
-Caption with unbalanced tags in it
+Parsoid-specific image handling - caption with unbalanced tags in it
 !! options
-parsoid
+parsoid=wt2html,wt2wt,html2html
 !! input
 foo
-[[Image:Foobar.jpg|thumb|200px|This caption has a 
unbalanced tag in it.]]
+[[File:Foobar.jpg|thumb|200x200px|This caption has a 
unbalanced tag in it.]]
 bar
 !! result
 
foo

@@ -10531,6 +10905,16 @@ parsoid
 
 !! end
 
+!! test
+Parsoid: Defaultsort
+!! options
+parsoid
+!! input
+{{DEFAULTSORT:Foo}}
+!! result
+
+!! end
+
 ###
 ### Inter-language links
 ###
@@ -11241,6 +11625,29 @@ I always thought &xacute; was a cute letter.
 

 !! end
 
+# TODO: generalize to PHP parser?
+!! test
+HTML5 tags
+!! options
+parsoid
+!! input
+five
+
+This highlighted text
+!! result
+
five
+
+This highlighted text

+!! end
+
+!! test
+HTML tag with leading space is parsed as text
+!! input
+< div>foo< /div>
+!! result
+
< div>foo< /div>
+

+!! end
 
 ###
 ### Nesting tests (see bug 41545, 50604, 51081)
@@ -16799,6 +17206,22 @@ B1  and  tags
+!!options
+parsoid=wt2html
+!!input
+A foo
+
+
+!!result
+
A [1]

+
+
    
+
  1. ↑ foo

+!!end
+
 !!test
 References: 1. references tag without any refs should be handled properly
 !!options
@@ -16817,7 +17240,7 @@ parsoid
 A foo
 B bar
 
-
+
 !!result
 
A [a 1]
 B [b 1]

@@ -16858,7 +17281,7 @@ parsoid
 A afoo
 B bfoo
 
-
+
 
 C cfoo
 
@@ -16904,9 +17327,9 @@ References: 6.  from a transclusion
 !!options
 parsoid
 !!input
-{{echo|}}
+Foo {{echo|}}
 !!result
-
    
+[1] 
    1. ↑ Foo
    
 !!end
 
 !! test
@@ -16983,41 +17406,48 @@ parsoid
 
 !! test
 Headings: 1. Nested inside html
+(New headings and existing headings are handled differently)
 !! options
-parsoid
+parsoid=html2wt
 !! input
-==foo==
+= =foo= =
 
-===foo===
+== =foo= ==
 
-====foo====
+=== =foo= ===
 
+==foo==
+===foo===
+====foo====
 =====foo=====
-
 ======foo======
-
 =======foo=======
+
 !! result
-
    =foo=
    
-
    =foo=
    
-
    =foo=
    
-
    =foo=
    
-
    =foo=
    
-
    =foo=
    
+
    =foo=
    
+
    =foo=
    
+
    =foo=
    
+
+
    =foo=
    
+
    =foo=
    
+
    =foo=
    
+
    =foo=
    
+
    =foo=
    
+
    =foo=
    
 !!end
 
 !! test
 Headings: 2. Outside heading nest on a single line 
    foo
    *bar
 !! options
-parsoid
+parsoid=html2wt
 !! input
-=foo=
+= foo =
 *bar
 
-=foo=
+= foo =
 =bar
 
-=foo=
+= foo =
 =bar=
 !! result
 
    foo
    *bar
@@ -17028,9 +17458,9 @@ parsoid
 !! test
 Headings: 3. Nested inside html with wikitext split by html tags
 !! options
-parsoid
+parsoid=html2wt
 !! input
-=='''bold'''foo==
+= ='''bold'''foo= =
 !! result
 
    =boldfoo=
    
 !!end
@@ -17038,23 +17468,23 @@ parsoid
 !! test
 Headings: 4a. No escaping needed (testing just h1 and h2)
 !! options
-parsoid
+parsoid=html2wt
 !! input
-==foo=
+= =foo =
 
-=foo==
+= foo= =
 
 = =foo= =
 
-==foo= bar=
+= =foo= bar =
 
-===foo==
+== =foo ==
 
-==foo===
+== foo= ==
 
-=''=''foo==
+= ''=''foo= =
 
-===
+= = =
 !! result
 
    =foo
    
 
    foo=
    
@@ -17069,7 +17499,7 @@ parsoid
 !! test
 Headings: 4b. No escaping needed (inside p-tags)
 !! options
-parsoid
+parsoid=html2wt
 !! input
 ===
 =foo= x
@@ -17171,8 +17601,8 @@ Headings: 6d. Heading chars in SOL context (with interspersed comments)
 parsoid
 !! input
 =a=
-
-=a= 	 
+
+=a= 	 
 !! result
 
    =a=
    
 
    =a= 	 
    
@@ -17326,6 +17756,8 @@ parsoid
 *''foo''*bar
 
 *[[Foo]]: bar
+
+*[[Foo]]*bar
 !! result
 
      
 
    • foo*bar
@@ -17339,6 +17771,10 @@ parsoid
 
    • Foo: bar
 
      • 
 
    
+
 !!end
 
 !! test
@@ -17482,14 +17918,18 @@ Tables: 1d. No escaping needed
 !! test
 Tables: 2a. Nested in td
 !! options
-parsoid
+parsoid=html2wt
 !! input
 {|
 |foo|bar
+|-
+|x
    a|b
    
 |}
 !! result
 
-
    foo|bar
    
+foo|bar
+x
    a|b
    
+
 !! end
 
 !! test
@@ -17612,6 +18052,9 @@ parsoid
 bar|baz
 +bar
 -bar
+|-
+|x
+
    a|b
    
 |}
 !! result
 
@@ -17621,32 +18064,27 @@ bar|baz
 
    bar|baz
 +bar
 -bar
    
+
    x
+
    a|b
    		
 
    
 !! end
 
-### SSS FIXME: Disabled right now because accurate html2wt
-### on this snippet requires data-parsoid flags that we've
-### stripped out of these tests.  We should scheme how we
-### we want to handle these kind of tests that require
-### data-parsoid flags for accurate html2wt serialization
-
 !! test
 Tables: 4d. No escaping needed
 !! options
-disabled
+parsoid
 !! input
 {|
+|[[Foo]]-bar
 ||+1
 ||-2
 |}
 !! result
 
-
-
-
    +1
--2
-
    
-
+Foo-bar
++1
+-2
+
 !! end
 
 #### --------------- Links ----------------
@@ -17713,9 +18151,16 @@ parsoid
 !! input
 [http://google.com [google]]
 [http://google.com google]]
+
+[http://google.com]
+
+[http://google.com google]
+
 !! result
-[google]
-google]
+
    [google]
+google]
    
+
    [http://google.com]
    
+
    [http://google.com google]
    
 !! end
 
 !! test
@@ -17754,6 +18199,8 @@ parsoid=html2wt,wt2wt
 ''''foo'''
 '''foo''''
 ''''foo''''
+''fools' errand''
+''fool's errand''
 !! result
 
    'foo'
 ''foo''
@@ -17770,6 +18217,8 @@ parsoid=html2wt,wt2wt
 'foo
 foo'
 'foo'
    
+fools' errand
+fool's errand
 !! end
 
 !! test
@@ -17901,6 +18350,7 @@ __|__
 #### 1. a tags
 #### 2. other tags
 #### 3. multi-line html tag
+#### 4. extension tags
 #### -----------------------------------------
 !! test
 1. a tags
@@ -17940,8 +18390,14 @@ parsoid
 4. extension tags
 !! input
 foo
+
+bar
+
+baz
 !! result
 
    <ref>foo</ref>
+
    <ref>bar
+
    baz</ref>
 
    
 !! end
 
@@ -18531,6 +18987,26 @@ parsoid=wt2wt,wt2html
 b
 !!end
 
+!!test
+Table in fosterable position
+!!options
+parsoid=wt2html,wt2wt
+!!input
+{{OpenTable}}
+
    
+{|
+|}
+
    
+|}
+!!result
+
    
+
+
    
+
+
+
    
+!!end
+
 # -----------------------------------------------------------------
 # The following section of tests are primarily to spec requirements
 # around serialization of new/edited content.
@@ -18560,16 +19036,11 @@ parsoid=html2wt
 
    123
    456
    
 !!end
 
-# Wacky -- the leading newline in input is required because
-# that is what the serializer emits. To be fixed. Not fixing
-# the test because this test is required to test serialization of
-# new content and preferred whitespace style.
 !! test
 Lists: Serialize correctly even when list content is wrapped in p-tags (like VE does)
 !! options
 parsoid=html2wt
 !! input
-
 * foo
 !! result
 
      
@@ -18577,6 +19048,16 @@ parsoid=html2wt
 
    
 !! end
 
+!! test
+Lists: Serialize correctly even when list tags has unneeded whitespace between tags
+!! options
+parsoid=html2wt
+!! input
+* foo
+!! result
+
       
    • foo
    
+!! end
+
 !! test
 Don't strip leading whitespace when handling indent-pre suppressing tags
 !! options
@@ -18623,6 +19104,10 @@ foo
 
 
    foo
    
 bar
+
+
    
+foo
+
    
 !! result
 
    foo
    
  bar
@@ -18632,18 +19117,17 @@ foo
 
 
    foo
    
  bar
+
+
    
+ foo
+
    
 !! end
 
-# Wacky -- the leading newline in input is required because
-# that is what the serializer emits. To be fixed. Not fixing
-# the test because this test is required to test serialization of
-# new content and preferred whitespace style.
 !! test
 Lists: Add space after bullets
 !! options
 parsoid=html2wt
 !! input
-
 * foo
 * bar
 *  baz
@@ -18655,6 +19139,25 @@ parsoid=html2wt
 
 !! end
 
+!! test
+Headings: Add space before/after == (Bug 51744)
+!! options
+parsoid=html2wt
+!! input
+== foo ==
+
+== bar ==
+
+== baz ==
+
+==  baz ==
+!! result
+
    foo
    
+
     bar
    
+
    baz 
    
+
     baz
    
+!! end
+
 !! test
 Parsoid: Serialize positional parameters with = in them as named parameter
 !! options
@@ -18781,8 +19284,12 @@ parsoid=html2wt
 
    AB
    
 !! end
 
+#------------------------------------
+# End of I/B quote minimization tests
+#------------------------------------
+
 !!test
-5. Bug 54262: New entities
+Bug 54262: New entities
 !! options
 parsoid=html2wt
 !! input
@@ -18791,6 +19298,98 @@ foo
 foo
 !! end
 
+## Note that there is no wikitext output for 'unknownproperty' ##
+## Unknown magic words are silently dropped ##
+
+!! test
+Magic words
+!! options
+parsoid=html2wt
+!! input
+__TOC__
+__NOTOC__
+__FORCETOC__
+__INDEX__
+__NOINDEX__
+__NOGALLERY__
+__NOEDITSECTION__
+__NOTITLECONVERT__
+__NOCONTENTCONVERT__
+!! result
+
+
+
+
+
+
+
+
+
+
+!! end
+
+!! test
+Consecutive 
    s should not get merged
+!! options
+parsoid=html2wt,html2html
+!! input
+ a
+
+ b
+
+ c
+
+ d
+
+ e
+ 
+
+ 
+ f
+!! result
+
    a
    b
    
+
+
    c
+
    +d
    
+
+
    e
+
+
    +
+f
    
+!! end
+
+!! test
+Edited ISBN links not serializable as ISBN links should serialize as wikilinks
+!! options
+parsoid=html2wt
+!! input
+[[Special:BookSources/1234567890|ISBN 1234567895]]
+!! result
+ISBN 1234567895
+!! end
+
+!! test
+Edited RFC links not serializable as RFC links should serialize as extlinks
+!! options
+parsoid=html2wt
+!! input
+[//tools.ietf.org/html/rfc123 New RFC]
+!! result
+New RFC
+!! end
+
+!! test
+Edited PMID links not serializable as PMID links should serialize as extlinks
+!! options
+parsoid=html2wt
+!! input
+[//www.ncbi.nlm.nih.gov/pubmed/123?dopt=Abstract New PMID]
+!! result
+New PMID
+!! end
+
 # -----------------------------------------------------------------
 # End of section for Parsoid-only html2wt tests for serialization
 # of new content
-- 
2.20.1