Fix content not getting parsed by linkedom properly without <html> tag by replacing innerHtml with outerHtml

This commit is contained in:
Hongbo Wu 2022-05-18 15:52:16 +08:00
parent 8f0447ed3f
commit 0cc7e84a82
2 changed files with 4 additions and 4 deletions

View file

@ -497,7 +497,7 @@ async function retrievePage(url) {
logRecord.finalUrl = response.url();
logRecord.contentType = response.headers()['content-type'];
return { context, page, response, finalUrl: finalUrl, contentType: contentType };
return { context, page, response, finalUrl, contentType };
} catch (error) {
if (lastPdfUrl) {
return { context, page, finalUrl: lastPdfUrl, contentType: 'application/pdf' };
@ -609,7 +609,7 @@ async function retrieveHtml(page) {
}
}
});
return document.documentElement.innerHTML;
return document.documentElement.outerHTML;
}, iframes);
logRecord.puppeteerSuccess = true;
logRecord.timing = {

View file

@ -678,7 +678,7 @@ async function retrievePage(url) {
logRecord.finalUrl = response.url();
logRecord.contentType = response.headers()['content-type'];
return { context, page, response, finalUrl: finalUrl, contentType: contentType };
return { context, page, response, finalUrl, contentType };
} catch (error) {
if (lastPdfUrl) {
return { context, page, finalUrl: lastPdfUrl, contentType: 'application/pdf' };
@ -790,7 +790,7 @@ async function retrieveHtml(page) {
}
}
});
return document.documentElement.innerHTML;
return document.documentElement.outerHTML;
}, iframes);
logRecord.puppeteerSuccess = true;
logRecord.timing = {