Git-Mirrors/brozzler
1
0
Fork 0
mirror of https://github.com/internetarchive/brozzler.git synced 2025-04-20 23:56:34 -04:00
Code Issues Actions 1 Packages Projects Releases Wiki Activity

add pop urls using regex for better match

Browse Source
This commit is contained in:
Barbara Miller 2019-09-03 14:47:48 -07:00
parent 16f886259d
commit 7bb52faca9
1 changed files with 10 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

10
brozzler/js-templates/extract-outlinks.js
View File

@ -1,6 +1,7 @@
// we have problems if the page has changed the definition of Set or Array
// http://www.polyvore.com/ does this for example
var __brzl_framesDone = new Set();
var __brzl_popup_re = /window.open\(.*?(['"])(http.*?)\1,/
var __brzl_compileOutlinks = function(frame) {
__brzl_framesDone.add(frame);
var outlinks = [];
@ -8,6 +9,15 @@ var __brzl_compileOutlinks = function(frame) {
if (frame && frame.document) {
outlinks = Array.prototype.slice.call(
frame.document.querySelectorAll('a[href], area[href]'));
popups = Array.prototype.slice.call(
frame.document.querySelectorAll('a[onclick]'));
if (popups && popups.length > 0) {
for (var p=0; p < popups.length; p++) {
m = popups[p].onclick.toString().match(__brzl_popup_re);
popups[p] = m[2];
}
outlinks = outlinks.concat(popups);
}
for (var i = 0; i < frame.frames.length; i++) {
if (frame.frames[i] && !__brzl_framesDone.has(frame.frames[i])) {
outlinks = outlinks.concat(