summaryrefslogtreecommitdiff
path: root/TODO
diff options
context:
space:
mode:
authorNick Mathewson <nickm@torproject.org>2008-11-22 17:39:57 +0000
committerNick Mathewson <nickm@torproject.org>2008-11-22 17:39:57 +0000
commit6c78fd8ff6b3d39fbae85103bda0651ca2f764d9 (patch)
treef392cef3d3e0b3810964d8f8a3f594229f8bf90d /TODO
parent06b83a522f6f05aed23b6dca4bfd4eeb6e6b057a (diff)
more docs and tests
git-svn-id: file:///home/or/svnrepo/updater/trunk@17364 55e972cd-5a19-0410-ae62-a4d7a52db4cd
Diffstat (limited to 'TODO')
-rw-r--r--TODO33
1 files changed, 20 insertions, 13 deletions
diff --git a/TODO b/TODO
index 758361f..53e8e88 100644
--- a/TODO
+++ b/TODO
@@ -1,18 +1,19 @@
-03:07 < nickm> ok. tomorrow morning I try to get messaging sorted, and try to
- write up a registry-based exe version checker.
+- get messaging sorted
+o try to write up a registry-based exe version checker.,
+- Decouple install from check: they are not necessarily related.
-- Download improvements.
- - Back off on download failure.
- - Handle full stalled file in download.
+. Download improvements.
+ o Back off on download failure.
+ o Handle full stalled file in download.
- Use if-modified-since on timestamp
-- Write client-side code
- . Decide early if a python implementation will do for v1.
+o Write client-side code
+ o Decide early if a python implementation will do for v1.
o Adjust httplib, urllib2 to use socks4a.
o Check SOCKS package for suitability as basis for socks4a support?
- - Look into best packageing practices
+ o Look into best packaging practices
- . Write code to run, telling another process about status,
+ o Write code to run, telling another process about status,
eventually coming up with a list of packages to install or an
"A-OK" signal.
@@ -21,7 +22,7 @@
o DL-via-Tor
o Install-when-done
o Verbose output
- - quiet output.
+ o quiet output.
X Rendezvous-back with Tor when done.
- Better configurability: let users override mirrors, keys, etc.
@@ -41,9 +42,10 @@
- Testing
- Much bigger unit tests.
-- Bugs Roger keeps noticing
- - you can add the same role to a key twice.
- - wishlist item: thandy-pk dumpkey-all
+o Bugs Roger keeps noticing
+ o you can add the same role to a key twice.
+ o wishlist item: thandy-pk dumpkey-all
+ o Already there: Run "thandy-pk dumpkey" with no arguments
- Confusing tracebacks that could use a check and error message:
@@ -66,6 +68,10 @@ line 343, in decryptSecret
raise thandy.BadPassword()
thandy.BadPassword
+ [ I maintain that BadPassword _is_ an error message. I'll make it say
+ "Password Incorrect", though, so it looks less like a bug.]
+
+
when your ~/.thandy/timestamp_key didn't get the secret key dumped too,
thandy@moria:~$ thandy-server timestamp
Traceback (most recent call last):
@@ -90,3 +96,4 @@ __getattr__
return getattr(self.key, attr)
AttributeError: rsaKey instance has no attribute 'd'
+ [ Gives a more descriptive error now. ]