non-standard URLs are not correctly interpreted and cannot be corrected

Furthermore, it becomes a crasher when you restart specto, so I'm marking this as critical:

Traceback (most recent call last):
  File "/usr/bin/specto", line 38, in ?
    specto = Specto()
  File "/usr/lib/python2.4/site-packages/specto/main.py", line 75, in __init__
    watch_value_db = self.watch_io.read_options()
  File "/usr/lib/python2.4/site-packages/specto/watch.py", line 163, in read_options
    values['type'] = int(watch_options['type'])
KeyError: 'type'

The content created in watches.list was this:

[perdu test]
updated = False

This information is too little for specto to handle ;) Actually, it completely bypassed collecting other information as soon as the URL did not work. This is something that should be fixed: even if the URL does not work, save the rest of the information.

AttributeError: web_watch instance has no attribute 'timer_id': that's because the watch hasn't been started.

KeyError: 'type': as you can see on your last comment, the watch is missing a lot of options

I think this has to be fixed using exceptions to make sure all options are correct.
Else the watch shouldn't be created at all and a warning should be displayed that the user has to input the right values.

if the first 7 characters aren't "http://", it will be added to the url

I just had a bug for the second time with this feature, and this time I figured out what happened. See the screenshot: specto did not correctly interpret the URL and added http:// even if I had that already. Uh oh!

Here is the output when I try pasting http://gnome-scanning.blogspot.com/ into the URL field and creating the watch:

  File "/usr/lib/python2.4/site-packages/specto/add_watch.py", line 202, in add_clicked
    self.create_watch()
  File "/usr/lib/python2.4/site-packages/specto/add_watch.py", line 294, in create_watch
    self.specto.add_watch(values) #write the options in the configuration file
  File "/usr/lib/python2.4/site-packages/specto/main.py", line 237, in add_watch
    self.start_watch(id)
  File "/usr/lib/python2.4/site-packages/specto/main.py", line 211, in start_watch
    self.watch_db[id].start_watch()
  File "/usr/lib/python2.4/site-packages/specto/watch_web_static.py", line 62, in start_watch
    self.update()
  File "/usr/lib/python2.4/site-packages/specto/watch_web_static.py", line 95, in update
    response = urllib2.urlopen(request)
  File "/usr/lib/python2.4/urllib2.py", line 130, in urlopen
    return _opener.open(url, data)
  File "/usr/lib/python2.4/urllib2.py", line 358, in open
    response = self._open(req, data)
  File "/usr/lib/python2.4/urllib2.py", line 376, in _open
    '_open', req)
  File "/usr/lib/python2.4/urllib2.py", line 337, in _call_chain
    result = func(*args)
  File "/usr/lib/python2.4/urllib2.py", line 1021, in http_open
    return self.do_open(httplib.HTTPConnection, req)
  File "/usr/lib/python2.4/urllib2.py", line 980, in do_open
    h = http_class(host) # will parse host:port
  File "/usr/lib/python2.4/httplib.py", line 586, in __init__
    self._set_hostport(host, port)
  File "/usr/lib/python2.4/httplib.py", line 598, in _set_hostport
    raise InvalidURL("nonnumeric port: '%s'" % host[i+1:])
httplib.InvalidURL: nonnumeric port: ''

Ok, seems like the behavior disappeared for me. Could you confirm everything works on your side, Wout? If not, reopen the bug.