# This file has tests to verify TOC section metadata output for wikitext snippets

!! options
version=2
!! end

!! article
Template:1x
!! text
{{{1}}}
!! endarticle

!! article
Template:Test
!! text
==th2==
===th2.1===
!! endarticle

!! test
Basic test
!! options
nohtml
showtocdata
parsoid={
	"modes": [ "wt2html" ],
	"wrapSections": true
}
!! wikitext
==h1==
===h1.1===
==h2==
===h2.1===
====h2.1.1====
===h2.2===
====h2.2.1====
====h2.2.2====
==h3==
!! html
{"toclevel":1,"level":"2","line":"h1","number":"1","index":"1","fromtitle":"Parser_test","byteoffset":0,"anchor":"h1","linkAnchor":"h1"}
{"toclevel":2,"level":"3","line":"h1.1","number":"1.1","index":"2","fromtitle":"Parser_test","byteoffset":7,"anchor":"h1.1","linkAnchor":"h1.1"}
{"toclevel":1,"level":"2","line":"h2","number":"2","index":"3","fromtitle":"Parser_test","byteoffset":18,"anchor":"h2","linkAnchor":"h2"}
{"toclevel":2,"level":"3","line":"h2.1","number":"2.1","index":"4","fromtitle":"Parser_test","byteoffset":25,"anchor":"h2.1","linkAnchor":"h2.1"}
{"toclevel":3,"level":"4","line":"h2.1.1","number":"2.1.1","index":"5","fromtitle":"Parser_test","byteoffset":36,"anchor":"h2.1.1","linkAnchor":"h2.1.1"}
{"toclevel":2,"level":"3","line":"h2.2","number":"2.2","index":"6","fromtitle":"Parser_test","byteoffset":51,"anchor":"h2.2","linkAnchor":"h2.2"}
{"toclevel":3,"level":"4","line":"h2.2.1","number":"2.2.1","index":"7","fromtitle":"Parser_test","byteoffset":62,"anchor":"h2.2.1","linkAnchor":"h2.2.1"}
{"toclevel":3,"level":"4","line":"h2.2.2","number":"2.2.2","index":"8","fromtitle":"Parser_test","byteoffset":77,"anchor":"h2.2.2","linkAnchor":"h2.2.2"}
{"toclevel":1,"level":"2","line":"h3","number":"3","index":"9","fromtitle":"Parser_test","byteoffset":92,"anchor":"h3","linkAnchor":"h3"}
!! end

!! test
HTML headings should get TOC entries with some empty properties
!! options
nohtml
showtocdata
parsoid={
	"modes": [ "wt2html" ],
	"wrapSections": true
}
!! wikitext
==a==
<h2>b</h2>
<h3>c</h3>
===d===
<h2>e</h2>
!! html
{"toclevel":1,"level":"2","line":"a","number":"1","index":"1","fromtitle":"Parser_test","byteoffset":0,"anchor":"a","linkAnchor":"a"}
{"toclevel":1,"level":"2","line":"b","number":"2","index":"","fromtitle":false,"byteoffset":null,"anchor":"b","linkAnchor":"b"}
{"toclevel":2,"level":"3","line":"c","number":"2.1","index":"","fromtitle":false,"byteoffset":null,"anchor":"c","linkAnchor":"c"}
{"toclevel":2,"level":"3","line":"d","number":"2.2","index":"2","fromtitle":"Parser_test","byteoffset":28,"anchor":"d","linkAnchor":"d"}
{"toclevel":1,"level":"2","line":"e","number":"3","index":"","fromtitle":false,"byteoffset":null,"anchor":"e","linkAnchor":"e"}
!! end

!! test
Duplicate headings should get unique anchors
!! options
nohtml
showtocdata
parsoid={
	"modes": [ "wt2html" ],
	"wrapSections": true
}
!! wikitext
==a==
==a==
==b==
!! html
{"toclevel":1,"level":"2","line":"a","number":"1","index":"1","fromtitle":"Parser_test","byteoffset":0,"anchor":"a","linkAnchor":"a"}
{"toclevel":1,"level":"2","line":"a","number":"2","index":"2","fromtitle":"Parser_test","byteoffset":6,"anchor":"a_2","linkAnchor":"a_2"}
{"toclevel":1,"level":"2","line":"b","number":"3","index":"3","fromtitle":"Parser_test","byteoffset":12,"anchor":"b","linkAnchor":"b"}
!! end

!! test
Templated sections (heading from template content)
!! options
nohtml
showtocdata
parsoid={
	"modes": [ "wt2html" ],
	"wrapSections": true
}
!! wikitext
==a==
{{Test}}
!! html
{"toclevel":1,"level":"2","line":"a","number":"1","index":"1","fromtitle":"Parser_test","byteoffset":0,"anchor":"a","linkAnchor":"a"}
{"toclevel":1,"level":"2","line":"th2","number":"2","index":"T-1","fromtitle":"Template:Test","byteoffset":null,"anchor":"th2","linkAnchor":"th2"}
{"toclevel":2,"level":"3","line":"th2.1","number":"2.1","index":"T-2","fromtitle":"Template:Test","byteoffset":null,"anchor":"th2.1","linkAnchor":"th2.1"}
!! end

!! test
Ensure headings with html tags get the right anchors
!! options
nohtml
showtocdata
parsoid={
	"modes": [ "wt2html" ],
	"wrapSections": true
}
!! wikitext
==<span>x</span>==
==<strike>y</strike>==
!! html
{"toclevel":1,"level":"2","line":"<span>x<\/span>","number":"1","index":"1","fromtitle":"Parser_test","byteoffset":0,"anchor":"x","linkAnchor":"x"}
{"toclevel":1,"level":"2","line":"<strike>y<\/strike>","number":"2","index":"2","fromtitle":"Parser_test","byteoffset":19,"anchor":"y","linkAnchor":"y"}
!! end

!! test
Ensure unallowed tag wrappers in headings get stripped
!! options
nohtml
showtocdata
parsoid={
	"modes": [ "wt2html" ],
	"wrapSections": true
}
!! wikitext
==<div>b</div>==
==<font>c</font>==
!! html
{"toclevel":1,"level":"2","line":"b","number":"1","index":"1","fromtitle":"Parser_test","byteoffset":0,"anchor":"b","linkAnchor":"b"}
{"toclevel":1,"level":"2","line":"c","number":"2","index":"2","fromtitle":"Parser_test","byteoffset":17,"anchor":"c","linkAnchor":"c"}
!! end

!! test
Ensure disallowed attributes are stripped, but dir attribute in a span is left behind
!! options
nohtml
showtocdata
parsoid={
	"modes": [ "wt2html" ],
	"wrapSections": true
}
!! wikitext
==<span dir='ltr' title='x'>a</span>==
==<span dir='rtl' title='x'>b</span>==
==<i dir='rtl' title='ha'>c</i>==
!! html
{"toclevel":1,"level":"2","line":"<span dir=\"ltr\">a<\/span>","number":"1","index":"1","fromtitle":"Parser_test","byteoffset":0,"anchor":"a","linkAnchor":"a"}
{"toclevel":1,"level":"2","line":"<span dir=\"rtl\">b<\/span>","number":"2","index":"2","fromtitle":"Parser_test","byteoffset":39,"anchor":"b","linkAnchor":"b"}
{"toclevel":1,"level":"2","line":"<i>c<\/i>","number":"3","index":"3","fromtitle":"Parser_test","byteoffset":78,"anchor":"c","linkAnchor":"c"}
!! end

# Legacy parser behavior is broken. Parsoid's output is correct.
!! test
Ensure empty tags are stripped
!! options
nohtml
showtocdata
parsoid={
	"modes": [ "wt2html" ],
	"wrapSections": true
}
!! wikitext
==<span><div></div></span>x==
==<span dir='ltr'><i dir='ltr'></i></span>y==
!! html/php
{"toclevel":1,"level":"2","line":"x","number":"1","index":"1","fromtitle":"Parser_test","byteoffset":0,"anchor":"x","linkAnchor":"x"}
{"toclevel":1,"level":"2","line":"<span dir=\"ltr\"><i><\/i><\/span>y","number":"2","index":"2","fromtitle":"Parser_test","byteoffset":30,"anchor":"y","linkAnchor":"y"}
!! html/parsoid
{"toclevel":1,"level":"2","line":"x","number":"1","index":"1","fromtitle":"Parser_test","byteoffset":0,"anchor":"x","linkAnchor":"x"}
{"toclevel":1,"level":"2","line":"y","number":"2","index":"2","fromtitle":"Parser_test","byteoffset":30,"anchor":"y","linkAnchor":"y"}
!! end

!! test
Handle links in heading content
!! options
nohtml
showtocdata
parsoid={
	"modes": [ "wt2html" ],
	"wrapSections": true
}
!! wikitext
==[[Cat]]==
==[[Dog]]s==
==[[Cat|I love my ''cat'']]==
!! html
{"toclevel":1,"level":"2","line":"Cat","number":"1","index":"1","fromtitle":"Parser_test","byteoffset":0,"anchor":"Cat","linkAnchor":"Cat"}
{"toclevel":1,"level":"2","line":"Dogs","number":"2","index":"2","fromtitle":"Parser_test","byteoffset":12,"anchor":"Dogs","linkAnchor":"Dogs"}
{"toclevel":1,"level":"2","line":"I love my <i>cat<\/i>","number":"3","index":"3","fromtitle":"Parser_test","byteoffset":25,"anchor":"I_love_my_cat","linkAnchor":"I_love_my_cat"}
!! end

# Parsoid uses HTML5 semantics for anchors as the default.
# Legacy parser tests require wgFragmentMode=[ "html5" ] for the same
# behavior, which isn't (yet) the default in ParserTestRunner
!! test
Ensure headings with special chars get the right anchors
!! options
nohtml
showtocdata
parsoid={
	"modes": [ "wt2html" ],
	"wrapSections": true
}
!! config
wgFragmentMode=[ "html5", "legacy" ]
!! wikitext
===a=
=''x''=
!! html
{"toclevel":1,"level":"1","line":"==a","number":"1","index":"1","fromtitle":"Parser_test","byteoffset":0,"anchor":"==a","linkAnchor":"==a"}
{"toclevel":1,"level":"1","line":"<i>x<\/i>","number":"2","index":"2","fromtitle":"Parser_test","byteoffset":6,"anchor":"x","linkAnchor":"x"}
!! end

# For heading that come from template args, core, Parsoid-standalone, and
# Parsoid-integrate different in ther "index" output.
#
# - For templated sections, Parsoid and core differ in some of the metadata.
#   Parsoid doesn't distinguish between sections that come from template args
#   and from templated content - it treats them all as template-generated headings.
#   But core does not consider headings from template-args as "inTemplate", but
#   'fromtitle' is set to false even so.
#
# - Parsoid's behavior is more consistent wrt templates. But, this is potentially
#   a breaking change for users of TOC data. The index for templated
#   content is used to generate section edit links for content coming
#   from a template, and in updating the appropriate article when that
#   section is saved.  So this is potentially a breaking change when
#   Parsoid starts to generate section edit links, and possible one
#   which could cause db corruption.
#
# - Additionally, heading-index is reset in integrated mode because templates are
#   processed by core preprocessor whereas in native preprocessing mode (used by
#   parsertests), the heading index is not reset.
# - See also T222419.
!! test
Templated sections (heading from template arg)
!! options
nohtml
showtocdata
parsoid={
	"modes": [ "wt2html" ],
	"wrapSections": true
}
!! wikitext
==a==
{{1x|1=
==b==
}}
!! html/php
{"toclevel":1,"level":"2","line":"a","number":"1","index":"1","fromtitle":"Parser_test","byteoffset":0,"anchor":"a","linkAnchor":"a"}
{"toclevel":1,"level":"2","line":"b","number":"2","index":"","fromtitle":false,"byteoffset":null,"anchor":"b","linkAnchor":"b"}
!! html/parsoid
{"toclevel":1,"level":"2","line":"a","number":"1","index":"1","fromtitle":"Parser_test","byteoffset":0,"anchor":"a","linkAnchor":"a"}
{"toclevel":1,"level":"2","line":"b","number":"2","index":"T-2","fromtitle":"Template:1x","byteoffset":null,"anchor":"b","linkAnchor":"b"}
!! html/parsoid+integrated
{"toclevel":1,"level":"2","line":"a","number":"1","index":"1","fromtitle":"Parser_test","byteoffset":0,"anchor":"a","linkAnchor":"a"}
{"toclevel":1,"level":"2","line":"b","number":"2","index":"T-1","fromtitle":"Template:1x","byteoffset":null,"anchor":"b","linkAnchor":"b"}
!! end

# FIXME:
# 1. This potentially exposes a subtle issue in the interaction between TOC, Cite, and
#    localization. Parsoid doesn't localize numbers in Cite (at all!) since it relies on CSS.
#    But, with the legacy parser, a wiki with localized numbers will have an anchor
#    based off a localized number if a <ref> is used in a heading. Parsoid won't!.
# 2. We probably should be a bit smarter about what to strip for tags when that output
#    comes from extensions. Below, we have stripped all extension typeofs and info.
#    But, depends on who uses this 'line' property and how.

# FIXME: test disabled because CI doesn't have the Cite extension installed
!! test
Handle extension content in section headers properly
!! config
wgFragmentMode=[ "html5", "legacy" ]
!! options
nohtml
showtocdata
parsoid={
	"modes": [ "wt2html" ],
	"wrapSections": true
}
!! wikitext
==c<ref>d</ref>==
!! html+disabled
{"toclevel":1,"level":"2","line":"c<sup>&#91;1&#93;<\/sup>","number":"1","index":"1","fromtitle":"Parser_test","byteoffset":0,"anchor":"c[1]","linkAnchor":"c[1]"}
!! html/parsoid
{"toclevel":1,"level":"2","line":"c<sup><span>[1]<\/span><\/sup>","number":"1","index":"1","fromtitle":"Parser_test","byteoffset":0,"anchor":"c[1]","linkAnchor":"c[1]"}
!! end

# FIXME: This test shows a similar difference between legacy and
# Parsoid output, but it may be due to a subtle difference in the
# whitespace emitted by the parser test runner's <tag> implementation
!! test
Handle extension content in section headers properly (non-Cite version)
!! config
wgFragmentMode=[ "html5", "legacy" ]
!! options
nohtml
showtocdata
parsoid={
	"modes": [ "wt2html" ],
	"wrapSections": true
}
!! wikitext
==c<tag>d</tag>==
!! html
{"toclevel":1,"level":"2","line":"c\n'd'\narray (\n)","number":"1","index":"1","fromtitle":"Parser_test","byteoffset":0,"anchor":"c_'d'_array_(_)","linkAnchor":"c_'d'_array_(_)"}
!! html/parsoid
{"toclevel":1,"level":"2","line":"c'd'\narray (\n)","number":"1","index":"1","fromtitle":"Parser_test","byteoffset":0,"anchor":"c'd'_array_(_)","linkAnchor":"c'd'_array_(_)"}
!! end