Some request can be missed in racy calls
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Ubuntu Single Sign On Client |
Triaged
|
High
|
Ubuntu One Client Engineering team | ||
ubuntu-sso-client (Ubuntu) |
Triaged
|
Medium
|
Ubuntu One Client Engineering team |
Bug Description
If you request a dbus call when the service is going to shutdown, you can miss some incoming request in ubuntu sso.
Here is a test file:
1. bzr branch lp:oneconf
2. cd oneconf/
3. python ssohandler.py
The included test simulate a first request, and then a second request in a 10 seconds timeout (as if your network disconnected/
What I get 50% of the time is:
$ python ssohandler.py
DEBUG:__main__:look for credential
DEBUG:_
dbus.Dictionary
DEBUG:__main__:look for credential
and no signal callback the second time (it's clearly racy, sometimes I can have it). Seems to be because my timeout is exactly the same amount of time of the sso service one, however it accepted the request, which it shouldn't have done.
Server logs are:
2011-06-23 15:41:43,
2011-06-23 15:41:43,
2011-06-23 15:41:43,
2011-06-23 15:41:43,
2011-06-23 15:41:53,
No input of the second request.
Changed in ubuntu-sso-client: | |
status: | New → Confirmed |
assignee: | nobody → Ubuntu One Desktop+ team (ubuntuone-desktop+) |
Changed in ubuntu-sso-client (Ubuntu): | |
assignee: | nobody → Ubuntu One Desktop+ team (ubuntuone-desktop+) |
status: | New → Confirmed |
Changed in ubuntu-sso-client: | |
importance: | Undecided → Medium |
Changed in ubuntu-sso-client (Ubuntu): | |
importance: | Undecided → Medium |
Changed in ubuntu-sso-client: | |
importance: | Medium → High |
status: | Confirmed → Triaged |
Changed in ubuntu-sso-client (Ubuntu): | |
status: | Confirmed → Triaged |