From 1b00e01b85e834f361296e17382b5efbae4d4c1b Mon Sep 17 00:00:00 2001 From: "C. Scott Ananian" Date: Tue, 9 Oct 2018 11:57:03 -0400 Subject: [PATCH] Sync up with Parsoid parserTests.txt This now aligns with Parsoid commit f6c64caaf1becc5b5780c871d495a3d9f5aac7a8 Change-Id: I0d2722c100b522c68a8e35a8701d5bf7ef80dccb --- tests/parser/parserTests.txt | 360 ++++++++++++++++++++++++++--------- 1 file changed, 265 insertions(+), 95 deletions(-) diff --git a/tests/parser/parserTests.txt b/tests/parser/parserTests.txt index 83ed27040e..bbd9ecbe99 100644 --- a/tests/parser/parserTests.txt +++ b/tests/parser/parserTests.txt @@ -2046,8 +2046,6 @@ x !! test Block tag on one line (
) -!! options -parsoid=wt2html,html2html !! wikitext a
foo
@@ -2079,8 +2077,6 @@ b

foo
!! test Block tag on both lines (
) -!! options -parsoid=wt2html,html2html !! wikitext a
foo
@@ -2494,8 +2490,6 @@ Bar !! test T17491: / in blockquote -!! options -parsoid=wt2html !! wikitext
Foo bar baz quux @@ -2515,8 +2509,6 @@ Foo bar baz quux !! test T17491: / in blockquote (2) -!! options -parsoid=wt2html !! wikitext
Foo bar baz quux
@@ -2851,6 +2843,10 @@ Barack Obama of the United States !! end ## Remex doesn't account for fostered content. +## The difference between Parsoid and the PHP parser can be attributed to core +## commit 674e8388cba and 710618f89af in Parsoid's repo. Parsoid doesn't +## tokenize unmatched extension tags that shadow html tags as strings to ease +## an escaping mechanism. See the comment in `maybeExtensionTag`. !! test Handle broken pre-like tags (T66025) !! options @@ -2869,47 +2865,36 @@ parsoid=wt2html !! html/parsoid 
x
-

<pre

+

 !! end
 
+## Similar to the above, but shows the difference between extension and html tags
 !! test
-Parsoid: handle pre with space after attribute
-!! options
-parsoid=wt2html
+Less than in attribute position
 !! wikitext
-
{{echo|foo}}

-!! html/php
-
{{echo|foo}}

+
123

 
+
123

+!! html/php+tidy
+
123

+<div 
123

 !! html/parsoid
-
{{echo|foo}}

+
123

+
+<div 
123

 !! end
 
-# TODO / maybe: fix wt2wt for this
 !! test
-Parsoid: Don't paragraph-wrap fosterable content
+Parsoid: handle pre with space after attribute
 !! options
 parsoid=wt2html
 !! wikitext
-{|
-
-
-
-
-
-|}
-!! html
-
-
-
-
-
-
-
-
-
+
{{echo|foo}}

+!! html/php
+
{{echo|foo}}

 
-

+!! html/parsoid
+
{{echo|foo}}

 !! end
 
 !! test
@@ -2920,7 +2905,7 @@ Self-closed pre
 

 
 !! html/parsoid
-

+

 !! end
 
 !! test
@@ -3533,8 +3518,6 @@ parsoid=wt2html,html2html
 
 !! test
 3a. Indent-Pre and block tags (single-line html)
-!! options
-parsoid=wt2html,html2html
 !! wikitext
  a 
 foo 

  b 
 foo 

@@ -3654,7 +3637,7 @@ array (
 )
 
 !! html/parsoid
-
 a 

+
 a 

 !! end
 
 !! test
@@ -6170,6 +6153,15 @@ Normal text. '''''Bold italic text.''''' Normal text.
 

 !! end
 
+!! test
+Quotes wrapping HTML table
+!! wikitext
+'''
hi
'''
+!! html/php+tidy
+
hi

+!! html/parsoid
+
hi

+!! end
 
 # Parsoid inserts an empty bold tag pair at the end of the line, that the PHP
 # parser strips. The wikitext contains just the first half of the bold
@@ -7058,7 +7050,7 @@ T107652: s in templates that also generate table cell attributes should be
 
 !! html/parsoid
 
-
+
Foo[1]
Foo[1]

 

 
  1. ↑  foo

 !! end
@@ -8825,7 +8817,7 @@ Interwiki links that cannot be represented in wiki syntax
 !! html/parsoid
 
meatball:ok
 ok with fragment
-ok ending with ? mark
+ok ending with ? mark
 has query
 is just fragment

 !! end
@@ -8908,6 +8900,17 @@ parsoid=wt2html,wt2wt,html2html
 
local:local:local:local:mi:local:Foo

 !! end
 
+!! test
+Interwiki link with percent encoded target
+!! wikitext
+[[:es:Nueva Guip%C3%BAzcoa|Nueva Guipúzcoa]]
+!! html/php
+
Nueva Guipúzcoa
+

+!! html/parsoid
+
Nueva Guipúzcoa

+!! end
+
 ###
 ### Interlanguage links
 ### Language links (so that searching for '### language' matches..)
@@ -9566,10 +9569,32 @@ Handling html with a div self-closing tag
 !! html/parsoid
 

 

-

+

 

 

-

+

+!! end
+
+!! test
+Elements with solidus in various attribute positions
+!! options
+parsoid=wt2html,html2html
+!! wikitext
+
ha

+
+
ha

+
+
ha

+!! html/php+tidy
+
ha

+
ha

+
ha

+!! html/parsoid
+
ha

+
+
ha

+
+
ha

 !! end
 
 !! test
@@ -9979,9 +10004,9 @@ Multiple list tags generated by templates
 
  • c
 
    • 
 !! html/parsoid
-
  • a
-
  • b
-
  • c
    • 
+
  • a
    • 
+
  • b
    • 
+
  • c
    • 
 !! end
 
 !! test
@@ -10228,11 +10253,27 @@ parsoid=wt2html,wt2wt,html2html
 
    • a
      • 
 
    • b
    
 !! html/parsoid
-
    
+
    
 
    • a
      • 
 
    • b
    
 !! end
 
+# Output is ugly because of all the misnested tag fixups.
+!! test
+3. List embedded in a formatting tag in a misnested way
+!! wikitext
+
+
+** 123
+!! html/php+tidy
+
    
+
      • 123
    
+!! html/parsoid
+
    
+
+
      •  123
    
+!! end
+
 ###
 ### Magic Words
 ###
@@ -12042,7 +12083,7 @@ Un-closed 
 
 !! html/php
 !! html/parsoid
-
+
 !! end
 
 ## We used to, but no longer wt2wt this test since the default serializer
@@ -12061,6 +12102,10 @@ some
 *stuff
 *here
 
+
+some* stuff
+* here
+
 can have stuff===here===
 
 !! html/php
@@ -12069,6 +12114,10 @@ some
 
    
 
    • stuff
      • 
 
    • here
    
+

    
+some* stuff
+
    
+
    • here
    
 
    here
    
 
 !! html/parsoid
@@ -12079,7 +12128,11 @@ some
 
    • stuff
      • 
 
    • here
    
 
-
    here
    
+
+
    some* stuff
    
+
    • here
    
+
+
    here
    
 
 !! end
 
@@ -12734,7 +12787,7 @@ parsoid=wt2html
 |c
 |}
 !!html/parsoid
-
+
    
 
    c
    
 
    
 !!end
@@ -12885,6 +12938,18 @@ parsoid
 
    Bar
    
 !!end
 
+!! test
+Templates: Scopes should not be expanded unnecessarily
+!! options
+parsoid=wt2html,wt2wt
+!! wikitext
+{{echo|
    a
    }}b{{echo|
+
    c
    }}
+!! html
+
    a
    b
    
+
    c
    
+!! end
+
 !!test
 Templates: Inline Text: 1. Multiple template uses
 !! wikitext
@@ -13448,7 +13513,7 @@ Templates: Wiki Tables: 7. Fosterable s should get fostered
 
 
 !! html/parsoid
-
    [1]
    
+
    [1]
    
 
 
    
 
@@ -13561,8 +13626,8 @@ Templates: Ugly templates: 5. Template encapsulation test: Non-trivial overlap o
 {{echo|
    foo}}
 {{echo|}}
 !! html/parsoid
-
    foo
-
    
+
    foo
    
+
    
 
    
 !! end
 
@@ -13625,19 +13690,24 @@ Parser Functions: 2. Nested use (only outermost should be marked up)
 ## Note that the templates inside the references are not wrapped
 !! test
 Template nested in extension tag in template
+!! options
+language=zh
 !! wikitext
 {{echo|hi[[ho|{{echo|hi}}]]}}
 {{echo|hi[http://test.com?q={{echo|ho}}]}}
+{{echo|hi-{ho|{{echo|hi}}}-}}
 
 !! html/parsoid
-
    hi[1]
-hi[2]
    
-
    1. ↑  hi
    2. ↑  
    
+
    hi[1]
+hi[2]
+hi[3]
    
+
    1. ↑  hi
    2. ↑  
    3. ↑  
    
 !! end
 
 ###
 ### Pre-save transform tests
 ###
+
 !! test
 pre-save transform: subst:
 !! options
@@ -14194,11 +14264,20 @@ ab
 cd
 
 ef
-!! html
-
    <div>[[foo]]
    
-
    ab
    
-
    cd
    
-
    ef
    
+!! html/php+tidy
+
    <div>[[foo]]
+
    a<nowiki / >b
+
    cd
+
    e<nowiki/ >f
+
    
+!! html/parsoid
+
    <div>[[foo]]
    
+
+
    a<nowiki / >b
    
+
+
    cd
    
+
+
    e<nowiki/ >f
    
 !! end
 
 !! test
@@ -15857,7 +15936,7 @@ T93580: 2.  inside inline images
 
 
 !! html/parsoid
-
    
+
    
 
 
    1. ↑  foo
    
 !! end
@@ -17711,9 +17790,9 @@ Media link
 Media:Video.ogv
 
    
 !! html/parsoid
-
    Media:Foobar.jpg
-Media:Video.ogv
-Media:Video.ogv
    
+
    Media:Foobar.jpg
+Media:Video.ogv
+Media:Video.ogv
    
 !! end
 
 !! test
@@ -17724,7 +17803,24 @@ Media link with text
 
    A neat file to look at
 
    
 !! html/parsoid
-
    A neat file to look at
    
+
    A neat file to look at
    
+!! end
+
+!! test
+Localized media link
+!! options
+language=is
+!! wikitext
+[[Miðill:Foobar.jpg]]
+
+[[Miðill:Foobar.jpg|Icelandic]]
+!! html/php
+
    Miðill:Foobar.jpg
+
    Icelandic
+
    
+!! html/parsoid
+
    Miðill:Foobar.jpg
    
+
    Icelandic
    
 !! end
 
 # FIXME: this is still bad HTML tag nesting
@@ -17742,7 +17838,53 @@ parsoid=wt2html,html2html
 !! html/php+tidy
 
    Safe Link
    " onmouseover="alert(document.cookie)" onfoo="
    
 !! html/parsoid
-
    Safe Link
    " onmouseover="alert(document.cookie)" onfoo="
    
+
    Safe Link
    " onmouseover="alert(document.cookie)" onfoo="
    
+!! end
+
+!! test
+Media link from missing resource (parsoid-only)
+!! options
+parsoid=html2wt
+!! html/parsoid
+Bat!
+!! wikitext
+[[Media:Foobar.jpg|Bat!]]
+!! end
+
+!! test
+Media link from missing resource, localized (parsoid-only)
+!! options
+parsoid=html2wt
+language=is
+!! html/parsoid
+Bat!
+!! wikitext
+[[Miðill:Foobar.jpg|Bat!]]
+!! end
+
+# This is a sanity test to ensure we don't crash or choke when given
+# old cached 1.7.0 HTML for media links; we don't require that the
+# result be "valid" because wt2wt was completely broken in 1.7.0
+!! test
+Media link from old 1.7.0 DOM spec (crash test, parsoid-only) T198511
+!! options
+parsoid=html2wt
+!! html/parsoid
+
    This is what Parsoid would emit (and we still handle this HTML fine):
    
+
    A neat file to look at
    
+
    But VisualEditor would mangle the rel attribute, and give us back this:
    
+
    A neat file to look at
    
+
    But at least we don't crash when trying to handle the mangled HTML!
    
+!! wikitext
+This is what Parsoid would emit (and we still handle this HTML fine):
+
+[[Media:Foobar.jpg|A neat file to look at]]
+
+But VisualEditor would mangle the rel attribute, and give us back this:
+
+[//example.com/images/3/3a/Foobar.jpg A neat file to look at]
+
+But at least we don't crash when trying to handle the mangled HTML!
 !! end
 
 !! test
@@ -17755,8 +17897,8 @@ Media link to nonexistent file (T3702)
 Media:No_such file.jpg
 
    
 !! html/parsoid
-
    Media:No such.jpg
-Media:No_such file.jpg
    
+
    Media:No such.jpg
+Media:No_such file.jpg
    
 !! end
 
 !! test
@@ -18385,7 +18527,7 @@ array (
 
 
 !! html/parsoid
-
    
+
    
 !! end
 
 !! test
@@ -18400,7 +18542,7 @@ array (
 
 
 !! html/parsoid
-
    
+
    
 !! end
 
 !! test
@@ -18550,7 +18692,7 @@ array (
 text
 
 !! html/parsoid
-
    text
    
+
    text
    
 !! end
 
 ##  should be output literally since there is no matching tag that begins it
@@ -18577,7 +18719,7 @@ array (
 </tag>
 
    
 !! html/parsoid
-
    other stuff
+
    other stuff
 </tag>
    
 !! end
 
@@ -18637,7 +18779,7 @@ hello, world
 
    
 !! html/parsoid
 
    
-
    hello, world
    
+
    hello, world
    
 !! end
 
 !! test
@@ -18650,7 +18792,7 @@ Parser hook: static parser hook inside a comment
 
    
 !! html/parsoid
 
-
    
+
    
 !! end
 
 # Nested template calls; this case was broken by Parser.php rev 1.506,
@@ -23504,8 +23646,6 @@ B
 # T8200: 
     should behave like 
     with respect to line breaks
 !! test
 T8200: paragraphs inside blockquotes (no extra line breaks)
-!! options
-parsoid=wt2html,html2html
 !! wikitext
 
    Line one
 
@@ -23521,8 +23661,6 @@ Line two
    
 
 !! test
 T8200: paragraphs inside blockquotes (extra line break on open)
-!! options
-parsoid=wt2html,html2html
 !! wikitext
 
    
 Line one
@@ -23543,8 +23681,6 @@ Line two
    
 
 !! test
 T8200: paragraphs inside blockquotes (extra line break on close)
-!! options
-parsoid=wt2html,html2html
 !! wikitext
 
    Line one
 
@@ -25180,9 +25316,9 @@ LST Sections: Newfangled approach
 
    
 
    
 !! html/parsoid
-
    
+
    
 
-
+
 
    
 !! end
 
@@ -27326,28 +27462,29 @@ Indented table with an empty td
 
 !!end
 
-## We have some newline diffs RT-ing this edge case
-## and it is not important enough -- we seem to be emitting
-## at most 2 newlines after a  and this is unrelated to
-## the issue from T85627 that this is testing.
-!!test
+!! test
 Indented table with blank lines in between (T85627)
-!! options
-parsoid=wt2html
 !! wikitext
  {|
  |foo
  
  
  |}
-!! html
+!! html/php
 
-
 
    
 foo
+

     
+
    
 
    
 
-!!end
+!! html/parsoid
+ 
+ 
+ 
    foo
+ 
    
+ 
    
+!! end
 
 !!test
 Indented block & table
@@ -27449,7 +27586,7 @@ New element inserted (without intervening newlines) after an old sol-transparent
 !! options
 parsoid=html2wt
 !! html/parsoid
-
    new para
    
+
    new para
    
 
 
    new heading
    
 !! wikitext
@@ -27738,9 +27875,11 @@ Properly encapsulate empty-content transclusions in fosterable positions
 
 !! end
 
+## No longer captures its original intention
 !! test
 Always encapsulate foster box when template range is expanded to table
 !! options
+disabled
 parsoid=wt2wt
 !! wikitext
 {|
@@ -28798,6 +28937,23 @@ parsoid={
 ''[[Foo]]''
 !! end
 
+!! test
+7. T194083 Regression test: Manual edit test to ensure nested diff markers aren't inserted
+!! options
+parsoid={
+  "modes": ["selser"],
+  "scrubWikitext": true,
+  "changes": [
+    ["div", "after", "\n
    Foo , placeholder
    "]
+  ]
+}
+!! wikitext
+
    placeholder
    
+!! wikitext/edited
+
    placeholder
    
+[[Foo]] , placeholder
+!! end
+
 #------------------------------
 # End of tag minimization tests
 #------------------------------
@@ -30867,6 +31023,20 @@ T72875: Test for brackets in attributes of elements in internal link texts
 link span
    
 !! end
 
+!! test
+T199926: html arrow wt: Parsoid sometimes trips up on verticalbar chars in hrefs
+!! options
+parsoid={
+  "modes": ["html2wt"]
+}
+!! html/parsoid
+9
+10
+!! wikitext
+[https://stats.wikimedia.org/v2/#/fr.wikipedia.org/reading/page-views-by-country/normal|map|2-Year~2016060100~2018071100|~total 9]
+[[stats:v2/#/fr.wikipedia.org/reading/page-views-by-country/normal%7Cmap%7C2-Year~2016060100~2018071100%7C~total|10]]
+!! end
+
 !! test
 T179544: {{anchorencode:}} output should be always usable in links
 !! config
-- 
2.20.1