X-Git-Url: https://plomlompom.com/repos/%7B%7Bprefix%7D%7D/static/gitweb.js?a=blobdiff_plain;f=plomlombot.py;h=ffb4192ac8f2925b1d6a34684acfd585eec82c6f;hb=5b0bbd796a399d17311d0551ebd19fedaab717dd;hp=7fa9194af8236e27fd7d547843fc1a1e3577dc8b;hpb=19b5218056b02ed8454cdc095560856bdf48b8b9;p=plomlombot-irc.git diff --git a/plomlombot.py b/plomlombot.py index 7fa9194..ffb4192 100644 --- a/plomlombot.py +++ b/plomlombot.py @@ -71,6 +71,28 @@ class IO: line) return line +def url_check(msg): + matches = re.findall("(https?://[^\s]+)", msg) + for i in range(len(matches)): + url = matches[i] + try: + webpage = urllib.request.urlopen(url, timeout=15) + except urllib.error.HTTPError as error: + print("TROUBLE FOLLOWING URL: " + str(error)) + continue + charset = webpage.info().get_content_charset() + if not charset: + charset="utf-8" + content_type = webpage.info().get_content_type() + if not content_type in ('text/html', 'text/xml', + 'application/xhtml+xml'): + print("TROUBLE INTERPRETING URL: bad content_type " + content_type) + continue + content = webpage.read().decode(charset) + title = str(content).split('