[or-cvs] r22113: {weather} Clean up TODO file for weather (weather/trunk)

[Christian Fromme <kaner@xxxxxxxxxx>]

Author: kaner
Date: 2010-04-02 14:03:14 +0000 (Fri, 02 Apr 2010)
New Revision: 22113

Modified:
   weather/trunk/TODO
Log:
Clean up TODO file for weather


Modified: weather/trunk/TODO
===================================================================
--- weather/trunk/TODO	2010-04-02 12:14:45 UTC (rev 22112)
+++ weather/trunk/TODO	2010-04-02 14:03:14 UTC (rev 22113)
@@ -2,41 +2,11 @@
 
 * What happens if the openssl command gets a go-slow DOS attack in response?
 
-* figure out how to disable HTTP logging through web.py, and should decide what
-  diagnostic logging makes sense (currently, there is a debug variable in each
-  of weather.py and poll.py; turning these on currently produces a lot of
-  semi-informative chatter on stdout and turns the web server into an oracle
-  for whether address x has already subscribed to alerts about node y, which is
-  of course unsatisfactory) 
-
-* either disable the reason parameter in send_failure_email() or replace it
-  with something that's less likely to leak anything sensitive (either about
-  Tor's config, through control port errors, or about a third party's response
-  to attempted openssl connections)
-
-* replace the openssl s_client call with a lookup to tor's v3 status Running?
-  consensus?
-
 * add /usr/sbin/sendmail support for sending email 
 
-* Should we send an alert if the host isn't up at the time of sign up?
-
-* Let the user set a useful name of the node being monitored
-
 * Tolerate a string preceding the identity fingerprint, as long as the
 rest of the line is a valid fingerprint. So people can paste in lines like
   moria1 FFCB 46DB 1339 DA84 674C 70D7 CB58 6434 C437 0441
 
-* Let the user add a reminder string when they sign up, so the string
-gets sent to them along with the 'your node is down' message.
-
-* Let the user pick a timeframe for how long the node needs to be down
-before the reminder gets sent. Rather than actually specifying a time
-interval, I was thinking of a pull-down box with "immediate", "a day or
-so", and "looks like it's gone". Then people who want high-availability
-servers can pick the first choice, normal people can pick the middle one,
-and people who convinced an org to run a node can pick the last one to
-learn if they need a nudge.
-
 * Let's also look at server versions and alert admins when they should upgrade.