Holger Levsen wrote (22 May 2016 12:52:46 GMT) : > as I tried to say above: the problem is, an update arrives, and people > update their system. After this the update cannot be tested anymore (as > people are understandably not really willing to go back to a backed up > state and retest again).
I suspect that Georg had automated testing in mind, so what *people* are ready to do or not doesn't matter this much. E.g. an automated test could keep around copies of ~/.local/share/torbrowser/, restore the one corresponding to the previous Tor Browser release, and test the update to the last version as Georg described, right? (no, I didn't volunteer :) Cheers, -- intrigeri
