| 1 |
Gordon Malm <gengor@g.o> posted |
| 2 |
200903091617.48682.gengor@g.o, excerpted below, on Mon, 09 Mar |
| 3 |
2009 16:17:48 -0700: |
| 4 |
|
| 5 |
> There is an important security aspect to retiring folks - commit |
| 6 |
> abilities. Perhaps in the case a dev wants to contribute but cannot in |
| 7 |
> the near future their commit privs can just be revoked until such time |
| 8 |
> they ask for them to be turned back on? I guess that would be an |
| 9 |
> 'extended devaway' ? |
| 10 |
|
| 11 |
This is my concern, and one that infra has expressed on occasions when |
| 12 |
the topic has come up before. On principle, a stale yet still active |
| 13 |
authorization is an authorization just begging to have some cracker |
| 14 |
stumble on it. We don't want some still active authorization and key |
| 15 |
from two years ago getting stolen and used to try to slip a bad commit |
| 16 |
under the radar, where the dev won't have a clue as he's not been active |
| 17 |
for years and has forgotten the key was even stashed somewhere /to/ get |
| 18 |
stolen. |
| 19 |
|
| 20 |
The six-month retirement thing is a backstop to prevent things like that |
| 21 |
from occurring. Maybe an extended "dev-away" mechanism, whereby the dev |
| 22 |
just needs to notify devrel that he's going active again (preferably in- |
| 23 |
person or with some inside joke or other verifiable mechanism so it's |
| 24 |
demonstrably /not/ some cracker simply finding the key), could be |
| 25 |
preferred to the whole inboarding, mentoring, etc, process all over again. |
| 26 |
|
| 27 |
Then again, a lot can change in two years. The former dev may not know |
| 28 |
about the new EAPIs and other policy and etc. changes, and having taken |
| 29 |
the quiz and gone thru the process before, it should be easier the second |
| 30 |
time and could be expedited, but there's an argument for still going thru |
| 31 |
it, just to start off again on the right foot, as they say. Also, that |
| 32 |
time would serve as an intro period between the returning dev and new |
| 33 |
ones that have come in since his retirement. So yeah, maybe a somewhat |
| 34 |
abridged inboarding is appropriate for a returning dev, but eliminating |
| 35 |
it entirely, as an extended dev-away, might be going too far. |
| 36 |
|
| 37 |
-- |
| 38 |
Duncan - List replies preferred. No HTML msgs. |
| 39 |
"Every nonfree program has a lord, a master -- |
| 40 |
and if you use the program, he is your master." Richard Stallman |
Archives