mirrors/zotero
2
0
Fork 0
Code Issues Pull requests Projects Releases Packages 1 Wiki Activity Actions

SAXXMLReader: Handle non-UTF-8 encodings (#3846)

Browse source
This commit is contained in:
Abe Jellinek 2024-03-13 12:20:59 -04:00 committed by Dan Stillman
parent c9fc68658b
commit 8b13f717b4
3 changed files with 59 additions and 1 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

10
resource/feeds/SAXXMLReader.js
View file

@ -62,7 +62,15 @@ class SAXXMLReader {
if (!response.ok) {
throw new Error("Unable to fetch data");
}
this._data = await response.text();
let buf = await response.arrayBuffer();
// We should use NetUtil.parseResponseContentType, but we don't have access to it here
let charset = response.headers.get("Content-Type")
?.match(/charset=([^;]+)/)
?.[1];
if (!charset) {
charset = 'utf-8';
}
this._data = new TextDecoder(charset).decode(buf);
this._parseAndNotify();
}

15
test/tests/data/feedWindows1252.rss Normal file
View file

@ -0,0 +1,15 @@
<?xml version="1.0" encoding="windows-1252"?>
<rss xmlns:dc="http://purl.org/dc/elements/1.1/" version="2.0">
<channel>
<title>Stortinget: Helse- og omsorgsministeren</title>
<link>https://www.stortinget.no</link>
<description>https://www.stortinget.no</description>
<language>no-NO</language>
<item>
<title>Skriftlig spørsmål fra Tage Pettersen (H) til helse- og omsorgsministeren. Til behandling</title>
<link>https://www.stortinget.no/no/Saker-og-publikasjoner/Sporsmal/Skriftlige-sporsmal-og-svar/Skriftlig-sporsmal/?qid=98117&amp;utm_medium=rss&amp;utm_source=www.stortinget.no&amp;utm_campaign=Helse- og omsorgsministeren</link>
<description>Hva vil helseministeren foreta seg på kort sikt for å sikre rekruttering av helsepersonell og at stortingets vedtak om å tilby en tverrfaglig helsekartlegging av barn som flyttes ut av hjemmet etterleves?</description>
<dc:date>2024-03-13</dc:date>
</item>
</channel>
</rss>

35
test/tests/feedReaderTest.js
View file

@ -251,4 +251,39 @@ describe("Zotero.FeedReader", function () {
assert.equal(item.enclosedItems[0].url, "https://static01.nyt.com/images/2021/06/16/world/16biden-photos1/16biden-photos1-moth.jpg");
});
});
describe("Legacy text encodings", function () {
var httpd;
var port = 16213;
var baseURL = `http://127.0.0.1:${port}/`;
before(function () {
Cu.import("resource://zotero-unit/httpd.js");
httpd = new HttpServer();
httpd.start(port);
httpd._handler._mimeMappings.rss = "text/xml; charset=ISO-8859-1";
httpd.registerPathHandler("/feedWindows1252.rss", {
handle(request, response) {
response.setStatusLine(null, 200, 'OK');
let file = getTestDataDirectory();
file.append("feedWindows1252.rss");
httpd._handler._writeFileResponse(request, file, response, 0, file.fileSize);
}
});
});
after(async function () {
await new Promise(resolve => httpd.stop(resolve));
});
it("should handle an ISO-8859-1 (windows-1252) feed", async function () {
let fr = new Zotero.FeedReader(baseURL + "feedWindows1252.rss");
await fr.process();
let itemIterator = new fr.ItemIterator();
let item = await itemIterator.next().value;
assert.equal(item.title, "Skriftlig spørsmål fra Tage Pettersen (H) til helse- og omsorgsministeren. Til behandling");
});
});
})