Snapshotting for rollback after updates was[ Re: Drawing lessons from fatal SELinux bug #1054350]

[Chris Murphy]

On Jan 26, 2014, at 4:47 PM, Reindl Harald <h.reindl at thelounge.net> wrote:

> 
> Am 27.01.2014 00:41, schrieb Chris Murphy:
>> Great, well I'll tell you what. I will just keep living dangerously, and when I find a real world case of this, I'll file a bug. How about that?
> 
> do that, your problem
> 
>>> because nobody *can* know what exactly packages, versions are installed
>>> in which combination or which *user specific* data may exist on exactly
>>> the FS which is restored *additionally* to what the system sofware knows
>>> 
>>> frankly you can have your kwallet or the files your browser stores
>>> passwords you recently created and thought they are safe on exactly
>>> that FS, and they *maybe* saved *between* upgrade, realize a problem
>>> and restore the snapshot
>> 
>> Oh for f's sake. And *maybe* the next time I take a shower I'll slip and fall
> 
> off-topic
> 
> you do not understand anything this theard is about so why not leaves us in peace?

I'll add rampant hyperbole to your list of personal attributes. I'm the one who doesn't understand anything even though you don't use or have any use for snapshots, and you also want no one to have them or it'll make developers careless:

On Jan 25, 2014, at 6:10 PM, Reindl Harald <h.reindl at thelounge.net> wrote:

> the short version of ahwat you said could have been "forget snapshots at all to solve
> such problems" to not lead dvelopers into temptation of "i can be less caeful because
> we have snapshots"
> 
> in other words: don't work around problems by create new ones	


And then you propose a ridonkulous snapshot-rollback strategy that would for certain cause major problems if the rollback were actually done, and then use that as fait accompli for why the entire concept of fs rollbacks are stupid. Your arguments are asinine. Your emails belong in a kill file.


Chris Murphy