Subramanya Sastry has uploaded a new change for review.
https://gerrit.wikimedia.org/r/253050
Change subject: T53444: Strip <br>s from headings via new HTML normalization
routine
......................................................................
T53444: Strip <br>s from headings via new HTML normalization routine
* Added parser tests that spec behavior.
Change-Id: Ia80e07be71cb2d10beb440bb075edc4e93f6399a
---
M lib/html2wt/normalizeDOM.js
M tests/parserTests.txt
2 files changed, 43 insertions(+), 0 deletions(-)
git pull ssh://gerrit.wikimedia.org:29418/mediawiki/services/parsoid
refs/changes/50/253050/1
diff --git a/lib/html2wt/normalizeDOM.js b/lib/html2wt/normalizeDOM.js
index 1363bad..309e92e 100644
--- a/lib/html2wt/normalizeDOM.js
+++ b/lib/html2wt/normalizeDOM.js
@@ -231,6 +231,21 @@
}
};
+Normalizer.prototype.stripBRs = function(node) {
+ var child = node.firstChild;
+ while (child) {
+ var next = child.nextSibling;
+ if (child.nodeName === 'BR') {
+ // replace <br/> with a single space
+ node.removeChild(child);
+ node.insertBefore(node.ownerDocument.createTextNode('
'), next);
+ } else if (DU.isElt(child)) {
+ stripBRs(node);
+ }
+ child = next;
+ }
+};
+
/**
* Normalizations implemented right now:
* -------------------------------------
@@ -239,6 +254,7 @@
* 3. Force SOL transparent links to serialize before/after heading
* 4. Trailing spaces are migrated out of links
* 5. Space is added before escapable prefixes in table cells
+ * 6. Strip <br/> from headings
*/
Normalizer.prototype.normalizeNode = function(node) {
// Only if scrubWikitext flag is enabled
@@ -261,6 +277,7 @@
if (/^H[1-6]$/.test(node.nodeName)) {
this.hoistLinks(node, false);
this.hoistLinks(node, true);
+ this.stripBRs(node);
return this.stripIfEmpty(node);
// Quote tags
diff --git a/tests/parserTests.txt b/tests/parserTests.txt
index 324a4e0..9c4e58b 100644
--- a/tests/parserTests.txt
+++ b/tests/parserTests.txt
@@ -24910,6 +24910,32 @@
!! end
!! test
+Headings: Replace <br/> with a single whitespace char (when scrubWikitext =
true)
+!! options
+parsoid={
+ "modes": ["html2wt"],
+ "scrubWikitext": true
+}
+!! html/parsoid
+<h2>foo<br/>bar</h2>
+!! wikitext
+== foo bar ==
+!! end
+
+!! test
+Headings: Replace <br/> with a single whitespace char (when scrubWikitext =
false)
+!! options
+parsoid={
+ "modes": ["html2wt"],
+ "scrubWikitext": false
+}
+!! html/parsoid
+<h2>foo<br/>bar</h2>
+!! wikitext
+== foo<br> bar ==
+!! end
+
+!! test
Parsoid: Serialize positional parameters with = in them as named parameter
!! options
parsoid=html2wt
--
To view, visit https://gerrit.wikimedia.org/r/253050
To unsubscribe, visit https://gerrit.wikimedia.org/r/settings
Gerrit-MessageType: newchange
Gerrit-Change-Id: Ia80e07be71cb2d10beb440bb075edc4e93f6399a
Gerrit-PatchSet: 1
Gerrit-Project: mediawiki/services/parsoid
Gerrit-Branch: master
Gerrit-Owner: Subramanya Sastry <[email protected]>
_______________________________________________
MediaWiki-commits mailing list
[email protected]
https://lists.wikimedia.org/mailman/listinfo/mediawiki-commits
