jyapayne/worm-scraper
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Take care of another broken Next Chapter pathology

Browse source
This commit is contained in:
Domenic Denicola 2015-05-06 01:30:55 +02:00
commit 972fcaa294
1 changed files with 5 additions and 3 deletions
Download patch file Download diff file Expand all files Collapse all files

8
lib/worm-scraper.js
View file

@ -136,12 +136,14 @@ function cleanContentEl(el) {
} }
function getNextChapterUrl(rawChapterDoc) { function getNextChapterUrl(rawChapterDoc) {
// a[title="Next Chapter"] doesn't always work (e.g. https://parahumans.wordpress.com/2011/09/27/shell-4-2/) // a[title="Next Chapter"] doesn't always work. Two different pathologies:
// So instead search for the first <a> within the main content area starting with "Next". // - https://parahumans.wordpress.com/2011/09/27/shell-4-2/
// - https://parahumans.wordpress.com/2012/04/21/sentinel-9-6/
// So instead search for the first <a> within the main content area starting with "Next", trimmed.
const aEls = rawChapterDoc.querySelectorAll(".entry-content a"); const aEls = rawChapterDoc.querySelectorAll(".entry-content a");
for (let i = 0; i < aEls.length; ++i) { for (let i = 0; i < aEls.length; ++i) {
if (aEls[i].textContent.startsWith("Next")) { if (aEls[i].textContent.trim().startsWith("Next")) {
return aEls[i].href; return aEls[i].href;
} }
} }