diff --git a/urls.sh b/urls.sh index 1e26497..356713a 100755 --- a/urls.sh +++ b/urls.sh @@ -1,3 +1,9 @@ -wget --spider --force-html -r -l1 -H $@ 2>&1 \ - | grep '^--' | awk '{ print $3 }' \ - | grep -v '\.\(css\|js\|png\|gif\|jpg\|txt\|ico\|ttf\|svg\)$' +wget --spider --force-html -r -l1 -H -U 'searplbot/1.0' $@ 2>&1 | tee wg + +grep '^--' wg | awk '{ print $3 }' \ + | grep -v '\.\(css\|js\|png\|gif\|jpg\|txt\|ico\|ttf\|svg\)$' \ + | tee ur + +sleep 10 + +php crawl.php $(cat ur)