jenkins-bot has submitted this change and it was merged.
Change subject: html2wt should not need access to original source
......................................................................
html2wt should not need access to original source
Bug: T96923
Change-Id: I104982e061abf20c63a6342caa5600a090c9ced7
---
M lib/dom.cleanup.js
M lib/mediawiki.WikitextSerializer.js
M lib/wts.SerializerState.js
M lib/wts.TagHandlers.js
4 files changed, 25 insertions(+), 32 deletions(-)
Approvals:
Subramanya Sastry: Looks good to me, approved
jenkins-bot: Verified
diff --git a/lib/dom.cleanup.js b/lib/dom.cleanup.js
index e888aab..1af70c0 100644
--- a/lib/dom.cleanup.js
+++ b/lib/dom.cleanup.js
@@ -91,11 +91,14 @@
DU.setDataMw( node, undefined );
}
- // Remove dp.src from elements that have valid data-mw and dsr. This
- // should reduce data-parsoid bloat.
- if ( validDataMW && Util.isValidDSR(dp.dsr) ) {
+ // Remove dp.src from elements that have valid data-mw and dsr.
+ // This should reduce data-parsoid bloat.
+ var validDSR = validDataMW && Util.isValidDSR(dp.dsr);
+ var isPageProp = (node.nodeName === 'META' &&
+
/^mw\:PageProp\/(.*)$/.test(node.getAttribute('property')));
+ if (validDSR && !isPageProp) {
dp.src = undefined;
- } else if ( tplInfo && tplInfo.first === node && (!atTopLevel ||
!dp.tsr) ) {
+ } else if (tplInfo && tplInfo.first === node && (!atTopLevel ||
!dp.tsr)) {
// Transcluded nodes will not have dp.tsr set
// and don't need dp.src either.
dp.src = undefined;
diff --git a/lib/mediawiki.WikitextSerializer.js
b/lib/mediawiki.WikitextSerializer.js
index 7e3d796..d51f4c5 100644
--- a/lib/mediawiki.WikitextSerializer.js
+++ b/lib/mediawiki.WikitextSerializer.js
@@ -650,8 +650,7 @@
if
(/(?:^|\s)mw:Transclusion(?=$|\s)/.test(typeOf)) {
dataMW = DU.getDataMw(node);
if (dataMW.parts) {
- src =
state.serializer._buildTemplateWT(node,
- state,
dataMW.parts );
+ src =
state.serializer._buildTemplateWT(node, state, dataMW.parts);
} else if (dp.src) {
self.env.log("error", "data-mw
missing in: " + node.outerHTML);
src = dp.src;
@@ -659,7 +658,11 @@
throw new Error("Cannot
serialize transclusion without data-mw.parts or data-parsoid.src.");
}
} else if
(/(?:^|\s)mw:Param(?=$|\s)/.test(typeOf)) {
- src = dp.src;
+ if (dp.src) {
+ src = dp.src;
+ } else {
+ throw new Error("No source for
params.");
+ }
} else if
(/(?:^|\s)mw:Extension\/LabeledSectionTransclusion/.test(typeOf)) {
// FIXME: Special case for <section>
until LST is implemented
// natively in Parsoid
@@ -698,24 +701,10 @@
throw new Error("Should never reach
here");
}
- // FIXME: Just adding this here temporarily
till we go in and
- // clean this up and strip this out if we can
verify that data-mw
- // is going to be present always when necessary
and indicate that
- // a missing data-mw is either a parser bug or
a client error.
- //
- // Fallback: should be exercised only in
exceptional situations.
- if (src === undefined && state.env.page.src &&
Util.isValidDSR(dp.dsr) && !dp.fostered) {
- src = state.getOrigSrc(dp.dsr[0],
dp.dsr[1]);
- }
- if (src !== undefined) {
- state.singleLineContext.disable();
- self.emitWikitext(src, state, cb, node);
- state.singleLineContext.pop();
- return
DU.skipOverEncapsulatedContent(node);
- } else {
- self.env.log( "error", "Serializing as
HTML. No handler for: " + node.outerHTML );
- return self._htmlElementHandler(node,
state, cb);
- }
+ state.singleLineContext.disable();
+ self.emitWikitext(src, state, cb, node);
+ state.singleLineContext.pop();
+ return DU.skipOverEncapsulatedContent(node);
},
sepnls: {
// XXX: This is questionable, as the template
can expand
diff --git a/lib/wts.SerializerState.js b/lib/wts.SerializerState.js
index a16c9b7..257677f 100644
--- a/lib/wts.SerializerState.js
+++ b/lib/wts.SerializerState.js
@@ -181,6 +181,7 @@
};
SSP.getOrigSrc = function(start, end) {
+ console.assert(this.selserMode);
return this.env.page.src.substring(start, end);
};
diff --git a/lib/wts.TagHandlers.js b/lib/wts.TagHandlers.js
index 9af1be9..6388d1e 100644
--- a/lib/wts.TagHandlers.js
+++ b/lib/wts.TagHandlers.js
@@ -931,17 +931,15 @@
// Check for property before type so that page
properties with templated attrs
// roundtrip properly. Ex: {{DEFAULTSORT:{{echo|foo}}
}}
- if ( property ) {
- var switchType = property.match(
/^mw\:PageProp\/(.*)$/ );
- if ( switchType ) {
+ if (property) {
+ var switchType =
property.match(/^mw\:PageProp\/(.*)$/);
+ if (switchType) {
var out = switchType[1];
var cat =
out.match(/^(?:category)?(.*)/);
if ( cat && Util.magicMasqs.has(cat[1])
) {
- var wtSource =
DU.getWTSource(state.env, node);
- // Use content so that VE
modifications are preserved
var contentInfo =
state.serializer.serializedAttrVal(node, "content", {});
- if (wtSource) {
- out =
wtSource.replace(/^([^:]+:)(.*)$/, "$1" + contentInfo.value + "}}");
+ if (dp.src) {
+ out =
dp.src.replace(/^([^:]+:)(.*)$/, "$1" + contentInfo.value + "}}");
} else {
var magicWord =
cat[1].toUpperCase();
state.env.log("warning", cat[1] + ' is missing source. Rendering as ' +
magicWord + ' magicword');
@@ -951,6 +949,8 @@
out =
state.env.conf.wiki.getMagicWordWT( switchType[1], dp.magicSrc ) || '';
}
cb(out, node);
+ } else {
+
state.serializer._htmlElementHandler(node, state, cb);
}
} else if ( type ) {
switch ( type ) {
--
To view, visit https://gerrit.wikimedia.org/r/210821
To unsubscribe, visit https://gerrit.wikimedia.org/r/settings
Gerrit-MessageType: merged
Gerrit-Change-Id: I104982e061abf20c63a6342caa5600a090c9ced7
Gerrit-PatchSet: 2
Gerrit-Project: mediawiki/services/parsoid
Gerrit-Branch: master
Gerrit-Owner: Arlolra <abrea...@wikimedia.org>
Gerrit-Reviewer: Cscott <canan...@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
