Git-Mirrors/brozzler
1
0
Fork 0
mirror of https://github.com/internetarchive/brozzler.git synced 2025-08-09 06:52:46 -04:00
Code Issues Projects Releases Packages Wiki Activity Actions

better regex, outlinks.push(m[2])

Browse source
This commit is contained in:
Barbara Miller 2019-09-24 16:40:45 -07:00
parent 2ec284c88b
commit ac9950a1ea
1 changed files with 4 additions and 2 deletions
Download patch file Download diff file Expand all files Collapse all files

6
brozzler/js-templates/extract-outlinks.js
View file

@ -1,7 +1,7 @@
// we have problems if the page has changed the definition of Set or Array // we have problems if the page has changed the definition of Set or Array
// http://www.polyvore.com/ does this for example // http://www.polyvore.com/ does this for example
var __brzl_framesDone = new Set(); var __brzl_framesDone = new Set();
var __brzl_popup_re = /window.open\(.*?(['"])(http.*?)\1,/ var __brzl_popup_re = /window.open\(.*?(['"])((?:https?)?\/\/.*?)\1,/
var __brzl_compileOutlinks = function(frame) { var __brzl_compileOutlinks = function(frame) {
__brzl_framesDone.add(frame); __brzl_framesDone.add(frame);
var outlinks = []; var outlinks = [];
@ -14,7 +14,9 @@ var __brzl_compileOutlinks = function(frame) {
if (popups && popups.length > 0) { if (popups && popups.length > 0) {
for (var p=0; p < popups.length; p++) { for (var p=0; p < popups.length; p++) {
m = popups[p].onclick.toString().match(__brzl_popup_re); m = popups[p].onclick.toString().match(__brzl_popup_re);
popups[p] = m[2]; if (m) {
outlinks.push(m[2]);
}
} }
outlinks = outlinks.concat(popups); outlinks = outlinks.concat(popups);
} }