jenkins-bot has submitted this change and it was merged. (
https://gerrit.wikimedia.org/r/345570 )
Change subject: T112043: Handle anchors without hrefs
......................................................................
T112043: Handle anchors without hrefs
Change-Id: Ia9accda1054d4ae41ec5d289a9e1ea1a77375aee
---
M lib/html2wt/LinkHandler.js
M tests/parserTests-blacklist.js
M tests/parserTests.txt
3 files changed, 54 insertions(+), 23 deletions(-)
Approvals:
Subramanya Sastry: Looks good to me, approved
jenkins-bot: Verified
diff --git a/lib/html2wt/LinkHandler.js b/lib/html2wt/LinkHandler.js
index fabe864..fa1cb3e 100644
--- a/lib/html2wt/LinkHandler.js
+++ b/lib/html2wt/LinkHandler.js
@@ -676,29 +676,46 @@
return false;
};
- var hrefStr;
+ var isFigure = false;
if (isComplexLink(node.attributes)) {
- env.log("error/html2wt/link", "Encountered",
- node.outerHTML,
+ env.log("error/html2wt/link", "Encountered",
node.outerHTML,
"-- serializing as extlink and dropping
<a> attributes unsupported in wikitext.");
- hrefStr = escapeExtLinkURL(getHref(env, node));
- return state.serializeLinkChildrenToString(node,
-
state.serializer.wteHandlers.aHandler).then(function(str) {
- state.emitChunk(new ExtLinkText('[' +
hrefStr + ' ' + str + ']',
- node, wiki, 'mw:ExtLink'),
node);
- });
- } else if (node.querySelector('IMG') &&
- node.querySelector('IMG').parentElement
=== node) {
+ } else {
+ var media = node.querySelector('img');
+ isFigure = !!(media && media.parentElement ===
node);
+ }
+
+ var hrefStr;
+ if (isFigure) {
// this is a basic html figure: <a><img></a>
return state.serializer.figureHandler(node);
} else {
// href is already percent-encoded, etc., but
it might contain
// spaces or other wikitext nasties. escape
the nasties.
hrefStr = escapeExtLinkURL(getHref(env, node));
- return state.serializeLinkChildrenToString(node,
-
state.serializer.wteHandlers.aHandler).then(function(str) {
- state.emitChunk(new ExtLinkText('[' +
hrefStr + ' ' + str + ']',
- node, wiki, 'mw:ExtLink'),
node);
+ var handler =
state.serializer.wteHandlers.aHandler;
+ return
state.serializeLinkChildrenToString(node, handler)
+ .then(function(str) {
+ var chunk;
+ if (!hrefStr) {
+ // Without an href, we just
emit the string as text.
+ // However, to preserve targets
for anchor links,
+ // serialize as a span with a
name.
+ var name =
node.getAttribute('name');
+ if (name) {
+ var doc =
node.ownerDocument;
+ var span =
doc.createElement('span');
+
span.setAttribute('name', name);
+
span.appendChild(doc.createTextNode(str));
+ chunk = span.outerHTML;
+ } else {
+ chunk = str;
+ }
+ } else {
+ chunk = new ExtLinkText('[' +
hrefStr + ' ' + str + ']',
+ node, wiki,
'mw:ExtLink');
+ }
+ state.emitChunk(chunk, node);
});
}
}
diff --git a/tests/parserTests-blacklist.js b/tests/parserTests-blacklist.js
index 8f2d8c0..18597f9 100644
--- a/tests/parserTests-blacklist.js
+++ b/tests/parserTests-blacklist.js
@@ -396,7 +396,7 @@
add("html2html", "Link containing double-single-quotes '' in text embedded in
italics (T6598 sanity check)", "<p data-parsoid='{\"dsr\":[0,84,0,0]}'><i
data-parsoid='{\"dsr\":[0,61,2,2]}'>Some
[/index.php?title=Link&action=edit&redlink=1 pretty </i>italics<i
data-parsoid='{\"dsr\":[68,84,2,2]}'> and stuff]!</i></p>\n");
add("html2html", "Plain link to page with question mark in title", "<p
data-parsoid='{\"dsr\":[0,16,0,0]}'><a rel=\"mw:WikiLink\"
href=\"./Wiki/A%3Fb\" title=\"Wiki/A?b\"
data-parsoid='{\"stx\":\"piped\",\"a\":{\"href\":\"./Wiki/A%3Fb\"},\"sa\":{\"href\":\"wiki/A?b\"},\"dsr\":[0,16,11,2]}'>A?b</a></p>\n\n<p
data-parsoid='{\"dsr\":[18,34,0,0]}'><a rel=\"mw:WikiLink\"
href=\"./Wiki/A%3Fb\" title=\"Wiki/A?b\"
data-parsoid='{\"stx\":\"piped\",\"a\":{\"href\":\"./Wiki/A%3Fb\"},\"sa\":{\"href\":\"wiki/A?b\"},\"dsr\":[18,34,11,2]}'>Baz</a></p>\n");
add("html2html", "Self-link to section should not be bold", "<p
data-parsoid='{\"dsr\":[0,44,0,0]}'><a rel=\"mw:WikiLink\"
href=\"./Wiki/Main_Page#section\" title=\"Wiki/Main Page\"
data-parsoid='{\"stx\":\"piped\",\"a\":{\"href\":\"./Wiki/Main_Page#section\"},\"sa\":{\"href\":\"wiki/Main
Page#section\"},\"dsr\":[0,44,25,2]}'>Main Page#section</a></p>\n");
-add("html2html", "Self-link to numeric title", "<p
data-parsoid='{\"dsr\":[0,4,0,0]}'>[ 0]</p>\n");
+add("html2html", "Self-link to numeric title", "<p
data-parsoid='{\"dsr\":[0,1,0,0]}'>0</p>\n");
add("html2html", "Link to numeric-equivalent title", "<p
data-parsoid='{\"dsr\":[0,14,0,0]}'><a rel=\"mw:WikiLink\" href=\"./Wiki/00\"
title=\"Wiki/00\"
data-parsoid='{\"stx\":\"piped\",\"a\":{\"href\":\"./Wiki/00\"},\"sa\":{\"href\":\"wiki/00\"},\"dsr\":[0,14,10,2]}'>00</a></p>\n");
add("html2html", "<nowiki> inside a link", "<p
data-parsoid='{\"dsr\":[0,84,0,0]}'><span typeof=\"mw:Nowiki\"
data-parsoid='{\"dsr\":[0,30,8,9]}'>[[Main Page]]</span> <a rel=\"mw:WikiLink\"
href=\"./Wiki/Main_Page\" title=\"Wiki/Main Page\"
data-parsoid='{\"stx\":\"piped\",\"a\":{\"href\":\"./Wiki/Main_Page\"},\"sa\":{\"href\":\"wiki/Main
Page\"},\"dsr\":[31,83,17,2]}'>the main page [it's not very good</a>]</p>\n");
add("html2html", "Non-breaking spaces in title", "<p
data-parsoid='{\"dsr\":[0,34,0,0]}'><a rel=\"mw:WikiLink\"
href=\"./Wiki/Main_Page\" title=\"Wiki/Main Page\"
data-parsoid='{\"stx\":\"piped\",\"a\":{\"href\":\"./Wiki/Main_Page\"},\"sa\":{\"href\":\"wiki/Main
Page\"},\"dsr\":[0,34,17,2]}'> Main Page </a></p>\n");
@@ -515,9 +515,9 @@
add("html2html", "ISBN with space-delimited number", "<p
data-parsoid='{\"dsr\":[0,58,0,0]}'><a rel=\"mw:WikiLink\"
href=\"./Wiki/Special:BookSources/9290170328\"
title=\"Wiki/Special:BookSources/9290170328\"
data-parsoid='{\"stx\":\"piped\",\"a\":{\"href\":\"./Wiki/Special:BookSources/9290170328\"},\"sa\":{\"href\":\"wiki/Special:BookSources/9290170328\"},\"dsr\":[0,58,38,2]}'>ISBN
92 9017 032 8</a></p>\n");
add("html2html", "ISBN length", "<p data-parsoid='{\"dsr\":[0,14,0,0]}'>ISBN
123456789</p>\n\n<p data-parsoid='{\"dsr\":[16,71,0,0]}'><a rel=\"mw:WikiLink\"
href=\"./Wiki/Special:BookSources/1234567890\"
title=\"Wiki/Special:BookSources/1234567890\"
data-parsoid='{\"stx\":\"piped\",\"a\":{\"href\":\"./Wiki/Special:BookSources/1234567890\"},\"sa\":{\"href\":\"wiki/Special:BookSources/1234567890\"},\"dsr\":[16,71,38,2]}'>ISBN
1234567890</a></p>\n\n<p data-parsoid='{\"dsr\":[73,89,0,0]}'>ISBN
12345678901</p>\n");
add("html2html", "ISBN with trailing year (T9110)", "<p
data-parsoid='{\"dsr\":[0,65,0,0]}'><a rel=\"mw:WikiLink\"
href=\"./Wiki/Special:BookSources/1234567890\"
title=\"Wiki/Special:BookSources/1234567890\"
data-parsoid='{\"stx\":\"piped\",\"a\":{\"href\":\"./Wiki/Special:BookSources/1234567890\"},\"sa\":{\"href\":\"wiki/Special:BookSources/1234567890\"},\"dsr\":[0,58,38,2]}'>ISBN
1-234-56789-0</a> - 2006</p>\n\n<p data-parsoid='{\"dsr\":[67,132,0,0]}'><a
rel=\"mw:WikiLink\" href=\"./Wiki/Special:BookSources/1234567890\"
title=\"Wiki/Special:BookSources/1234567890\"
data-parsoid='{\"stx\":\"piped\",\"a\":{\"href\":\"./Wiki/Special:BookSources/1234567890\"},\"sa\":{\"href\":\"wiki/Special:BookSources/1234567890\"},\"dsr\":[67,125,38,2]}'>ISBN
1 234 56789 0</a> - 2006</p>\n");
-add("html2html", "Self-link in language variants", "<p
data-parsoid='{\"dsr\":[0,52,0,0]}'>Both [ Dunav] and [ Дунав] are names for
this river.</p>\n");
-add("html2html", "Link to another existing title shouldn't be parsed as
self-link even if it's a variant of this title", "<p
data-parsoid='{\"dsr\":[0,85,0,0]}'><a rel=\"mw:WikiLink\" href=\"./Wiki/Дуна\"
title=\"Wiki/Дуна\"
data-parsoid='{\"stx\":\"piped\",\"a\":{\"href\":\"./Wiki/Дуна\"},\"sa\":{\"href\":\"wiki/Дуна\"},\"dsr\":[0,18,12,2]}'>Дуна</a>
is not a self-link while [ Duna] and [ Dуна] are still self-links.</p>\n");
-add("html2html", "Link to a section of a variant of this title shouldn't be
parsed as self-link", "<p data-parsoid='{\"dsr\":[0,98,0,0]}'>[ Dуна] is a
self-link while <a rel=\"mw:WikiLink\" href=\"./Wiki/Дуна\" title=\"Wiki/Дуна\"
data-parsoid='{\"stx\":\"piped\",\"a\":{\"href\":\"./Wiki/Дуна\"},\"sa\":{\"href\":\"wiki/Дуна\"},\"dsr\":[29,51,12,2]}'>Dunа#Foo</a>
and <a rel=\"mw:WikiLink\" href=\"./Wiki/Дуна\" title=\"Wiki/Дуна\"
data-parsoid='{\"stx\":\"piped\",\"a\":{\"href\":\"./Wiki/Дуна\"},\"sa\":{\"href\":\"wiki/Дуна\"},\"dsr\":[56,78,12,2]}'>Dуна#Foo</a>
are not self-links.</p>\n");
+add("html2html", "Self-link in language variants", "<p
data-parsoid='{\"dsr\":[0,46,0,0]}'>Both Dunav and Дунав are names for this
river.</p>\n");
+add("html2html", "Link to another existing title shouldn't be parsed as
self-link even if it's a variant of this title", "<p
data-parsoid='{\"dsr\":[0,79,0,0]}'><a rel=\"mw:WikiLink\" href=\"./Wiki/Дуна\"
title=\"Wiki/Дуна\"
data-parsoid='{\"stx\":\"piped\",\"a\":{\"href\":\"./Wiki/Дуна\"},\"sa\":{\"href\":\"wiki/Дуна\"},\"dsr\":[0,18,12,2]}'>Дуна</a>
is not a self-link while Duna and Dуна are still self-links.</p>\n");
+add("html2html", "Link to a section of a variant of this title shouldn't be
parsed as self-link", "<p data-parsoid='{\"dsr\":[0,95,0,0]}'>Dуна is a
self-link while <a rel=\"mw:WikiLink\" href=\"./Wiki/Дуна\" title=\"Wiki/Дуна\"
data-parsoid='{\"stx\":\"piped\",\"a\":{\"href\":\"./Wiki/Дуна\"},\"sa\":{\"href\":\"wiki/Дуна\"},\"dsr\":[26,48,12,2]}'>Dunа#Foo</a>
and <a rel=\"mw:WikiLink\" href=\"./Wiki/Дуна\" title=\"Wiki/Дуна\"
data-parsoid='{\"stx\":\"piped\",\"a\":{\"href\":\"./Wiki/Дуна\"},\"sa\":{\"href\":\"wiki/Дуна\"},\"dsr\":[53,75,12,2]}'>Dуна#Foo</a>
are not self-links.</p>\n");
add("html2html", "Link to pages in language variants", "<p
data-parsoid='{\"dsr\":[0,56,0,0]}'>Main Page can be written as <a
rel=\"mw:WikiLink\" href=\"./Wiki/Main_Page\" title=\"Wiki/Main Page\"
data-parsoid='{\"stx\":\"piped\",\"a\":{\"href\":\"./Wiki/Main_Page\"},\"sa\":{\"href\":\"wiki/Main
Page\"},\"dsr\":[28,56,17,2]}'>Маин Паге</a></p>\n");
add("html2html", "Multiple links to pages in language variants", "<p
data-parsoid='{\"dsr\":[0,113,0,0]}'><a rel=\"mw:WikiLink\"
href=\"./Wiki/Main_Page\" title=\"Wiki/Main Page\"
data-parsoid='{\"stx\":\"piped\",\"a\":{\"href\":\"./Wiki/Main_Page\"},\"sa\":{\"href\":\"wiki/Main
Page\"},\"dsr\":[0,28,17,2]}'>Main Page</a> can be written as <a
rel=\"mw:WikiLink\" href=\"./Wiki/Main_Page\" title=\"Wiki/Main Page\"
data-parsoid='{\"stx\":\"piped\",\"a\":{\"href\":\"./Wiki/Main_Page\"},\"sa\":{\"href\":\"wiki/Main
Page\"},\"dsr\":[47,75,17,2]}'>Маин Паге</a> same as <a rel=\"mw:WikiLink\"
href=\"./Wiki/Main_Page\" title=\"Wiki/Main Page\"
data-parsoid='{\"stx\":\"piped\",\"a\":{\"href\":\"./Wiki/Main_Page\"},\"sa\":{\"href\":\"wiki/Main
Page\"},\"dsr\":[84,112,17,2]}'>Маин Паге</a>.</p>\n");
add("html2html", "Prevent conversion of links with -{}- tags (language
variants)", "<p data-parsoid='{\"dsr\":[0,28,0,0]}'><a rel=\"mw:WikiLink\"
href=\"./Wiki/Main_Page\" title=\"Wiki/Main Page\"
data-parsoid='{\"stx\":\"piped\",\"a\":{\"href\":\"./Wiki/Main_Page\"},\"sa\":{\"href\":\"wiki/Main
Page\"},\"dsr\":[0,28,17,2]}'>Main Page</a></p>\n");
@@ -728,7 +728,7 @@
add("html2wt", "Plain link to page with question mark in title",
"[[wiki/A?b|A?b]]\n\n[[wiki/A?b|Baz]]\n");
add("html2wt", "T2337: Escaped self-links should be bold", "[[Bug462]]
[[Bug462]]\n");
add("html2wt", "Self-link to section should not be bold", "[[wiki/Main
Page#section|Main Page#section]]\n");
-add("html2wt", "Self-link to numeric title", "[ 0]\n");
+add("html2wt", "Self-link to numeric title", "0\n");
add("html2wt", "Link to numeric-equivalent title", "[[wiki/00|00]]\n");
add("html2wt", "<nowiki> inside a link", "<nowiki>[[Main Page]]</nowiki>
[[wiki/Main Page|the main page [it's not very good]]]\n");
add("html2wt", "Non-breaking spaces in title", "[[wiki/Main Page| Main Page
]]\n");
@@ -1042,9 +1042,9 @@
add("html2wt", "T8200: blockquotes and paragraph formatting",
"<blockquote>\nfoo\n\n</blockquote>\nbar\n\n baz\n");
add("html2wt", "T10293: Use of center tag ruins paragraph formatting",
"<center>\nfoo\n\n</center>\nbar\n\n baz\n");
add("html2wt", "Parsing of overlapping (improperly nested) inline html tags",
"<span><s>x</s></span>\n");
-add("html2wt", "Self-link in language variants", "Both [ Dunav] and [ Дунав]
are names for this river.\n");
-add("html2wt", "Link to another existing title shouldn't be parsed as
self-link even if it's a variant of this title", "[[wiki/Дуна|Дуна]] is not a
self-link while [ Duna] and [ Dуна] are still self-links.\n");
-add("html2wt", "Link to a section of a variant of this title shouldn't be
parsed as self-link", "[ Dуна] is a self-link while [[wiki/Дуна|Dunа#Foo]] and
[[wiki/Дуна|Dуна#Foo]] are not self-links.\n");
+add("html2wt", "Self-link in language variants", "Both Dunav and Дунав are
names for this river.\n");
+add("html2wt", "Link to another existing title shouldn't be parsed as
self-link even if it's a variant of this title", "[[wiki/Дуна|Дуна]] is not a
self-link while Duna and Dуна are still self-links.\n");
+add("html2wt", "Link to a section of a variant of this title shouldn't be
parsed as self-link", "Dуна is a self-link while [[wiki/Дуна|Dunа#Foo]] and
[[wiki/Дуна|Dуна#Foo]] are not self-links.\n");
add("html2wt", "Link to pages in language variants", "Main Page can be written
as [[wiki/Main Page|Маин Паге]]\n");
add("html2wt", "Multiple links to pages in language variants", "[[wiki/Main
Page|Main Page]] can be written as [[wiki/Main Page|Маин Паге]] same as
[[wiki/Main Page|Маин Паге]].\n");
add("html2wt", "Simple template in language variants", "This is a test
template\n");
diff --git a/tests/parserTests.txt b/tests/parserTests.txt
index 935c539..cb0001f 100644
--- a/tests/parserTests.txt
+++ b/tests/parserTests.txt
@@ -27449,6 +27449,20 @@
|}
!! end
+!! test
+Anchor without href scenarios
+!! options
+parsoid={ "modes": ["html2wt"], "suppressErrors": true }
+!! html/parsoid
+<a class="bc"></a>
+<a class="no">dice</a>
+<a name="foo"></a>
+!! wikitext
+
+dice
+<span name="foo"></span>
+!! end
+
# -----------------------------------------------------------------
# End of section for Parsoid-only html2wt tests for serialization
# of new content
--
To view, visit https://gerrit.wikimedia.org/r/345570
To unsubscribe, visit https://gerrit.wikimedia.org/r/settings
Gerrit-MessageType: merged
Gerrit-Change-Id: Ia9accda1054d4ae41ec5d289a9e1ea1a77375aee
Gerrit-PatchSet: 3
Gerrit-Project: mediawiki/services/parsoid
Gerrit-Branch: master
Gerrit-Owner: Arlolra <[email protected]>
Gerrit-Reviewer: Arlolra <[email protected]>
Gerrit-Reviewer: C. Scott Ananian <[email protected]>
Gerrit-Reviewer: Subramanya Sastry <[email protected]>
Gerrit-Reviewer: jenkins-bot <>
_______________________________________________
MediaWiki-commits mailing list
[email protected]
https://lists.wikimedia.org/mailman/listinfo/mediawiki-commits
