jenkins-bot has submitted this change and it was merged.
Change subject: Fix parserTests to serialize-to-html + reparse-to-DOM exactly
once.
......................................................................
Fix parserTests to serialize-to-html + reparse-to-DOM exactly once.
* The DOM was being serialized to HTML and parsed back multiple
times on the wt2wt and selser paths which would change DOM
semantics for HTML-pre (by stripping one newline each time).
* This patches fixes it to always do it exactly once.
* Fixes the HTML-pre test wt2wt failure.
* Speeds up parser tests run by over 20%.
Change-Id: Ic7e9a4a8fc92603cc96e25ea3ea9915f5abb37d8
---
M tests/parserTests-blacklist.js
M tests/parserTests.js
2 files changed, 20 insertions(+), 13 deletions(-)
Approvals:
GWicke: Looks good to me, approved
jenkins-bot: Verified
diff --git a/tests/parserTests-blacklist.js b/tests/parserTests-blacklist.js
index 39237d2..4d1d733 100644
--- a/tests/parserTests-blacklist.js
+++ b/tests/parserTests-blacklist.js
@@ -488,7 +488,6 @@
add("wt2wt", "4. Indent-Pre and extension tags", " a <gallery>\n
File:foobar.jpg\n </gallery>\n");
add("wt2wt", "Leading pipes outside of tables 2", "a\n| foo\n\nb\n");
add("wt2wt", "Leading pipes outside of tables 3", "a\n| class=\"foo bar\" |
baz\n\nb\n");
-add("wt2wt", "HTML-pre: 1. embedded newlines",
"<pre>foo</pre>\n\n<pre>\nfoo\n</pre>\n\n<pre>\nfoo\n</pre>\n\n<pre>\n\nfoo\n</pre>");
add("wt2wt", "Definition lists: self-closed tag", ";one<br />two : two-line
fun");
add("wt2wt", "BUG 289: \">\"-token in bracketed URL",
"[http://www.example.com/ <hello> stuff]\n");
add("wt2wt", "BUG 289: literal \">\"-token in bracketed URL",
"[http://www.example.com/ <b>html</b> stuff]\n");
diff --git a/tests/parserTests.js b/tests/parserTests.js
index c83970a..9d32d9f 100755
--- a/tests/parserTests.js
+++ b/tests/parserTests.js
@@ -428,7 +428,7 @@
self = this,
startsAtWikitext = mode === 'wt2wt' || mode === 'wt2html' ||
mode === 'selser';
try {
- this.env.page.dom = item.cachedHTMLStr ?
DU.parseHTML(item.cachedHTMLStr).body : null;
+ this.env.page.dom = item.cachedDOM ? item.cachedDOM.body : null;
if ( mode === 'selser' ) {
// console.warn("--> selsering: " + content.outerHTML);
this.env.setPageSrcInfo( item.wikitext );
@@ -937,17 +937,29 @@
// First conversion stage
if ( startsAtWikitext ) {
- if ( item.cachedHTMLStr === null ) {
+ // Always serialize DOM to string and reparse before passing to
wt2wt
+ if ( item.cachedDOM === null ) {
testTasks.push( this.convertWt2Html.bind( this, mode,
item.wikitext ) );
- // Caching stage 1 - save the result of the first two
stages so we can maybe skip them later
+ // Caching stage 1 - save the result of the first two
stages
+ // so we can maybe skip them later
testTasks.push( function ( result, cb ) {
// Cache parsed HTML
- item.cachedHTMLStr = DU.serializeNode(result);
- cb( null, result );
+ item.cachedDOM =
DU.parseHTML(DU.serializeNode(result));
+
+ // - In wt2html mode, pass through original DOM
+ // so that it is serialized just once.
+ // - In wt2wt and selser modes, pass through
serialized and
+ // reparsed DOM so that
fostering/normalization effects
+ // are reproduced.
+ if (mode === "wt2html") {
+ cb(null, result);
+ } else {
+ cb(null,
item.cachedDOM.body.cloneNode(true));
+ }
} );
} else {
testTasks.push( function ( cb ) {
- cb( null, DU.parseHTML(item.cachedHTMLStr) );
+ cb(null, item.cachedDOM.body.cloneNode(true));
} );
}
} else if ( startsAtHtml ) {
@@ -978,11 +990,7 @@
} );
}
- // Always serialize DOM to string and reparse before passing to wt2wt
if (mode === 'wt2wt') {
- testTasks.push( function ( doc, cb ) {
- cb( null, DU.parseHTML(DU.serializeNode(doc)).body);
- } );
// handle a 'changes' option if present.
if (item.options.parsoid && item.options.parsoid.changes) {
testTasks.push( function( doc, cb ) {
@@ -1729,7 +1737,7 @@
}
// Push the caches
forward!
- item.cachedHTMLStr =
newitem.cachedHTMLStr;
+ item.cachedDOM =
newitem.cachedDOM;
item.cachedNormalizedHTML = newitem.cachedNormalizedHTML;
setImmediate( cb );
@@ -1778,7 +1786,7 @@
// Reset the cached results for the new case.
// All test modes happen in a single run of processCase.
- item.cachedHTMLStr = null;
+ item.cachedDOM = null;
item.cachedNormalizedHTML = null;
//console.log( 'processCase ' + i + JSON.stringify( item ) );
--
To view, visit https://gerrit.wikimedia.org/r/134787
To unsubscribe, visit https://gerrit.wikimedia.org/r/settings
Gerrit-MessageType: merged
Gerrit-Change-Id: Ic7e9a4a8fc92603cc96e25ea3ea9915f5abb37d8
Gerrit-PatchSet: 4
Gerrit-Project: mediawiki/services/parsoid
Gerrit-Branch: master
Gerrit-Owner: Subramanya Sastry <ssas...@wikimedia.org>
Gerrit-Reviewer: Arlolra <abrea...@wikimedia.org>
Gerrit-Reviewer: Cscott <canan...@wikimedia.org>
Gerrit-Reviewer: GWicke <gwi...@wikimedia.org>
Gerrit-Reviewer: Marcoil <marc...@wikimedia.org>
Gerrit-Reviewer: Subramanya Sastry <ssas...@wikimedia.org>
Gerrit-Reviewer: jenkins-bot <>
_______________________________________________
MediaWiki-commits mailing list
MediaWiki-commits@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/mediawiki-commits
