Re: [gentoo-user] why does revdep-rebuild object to mounting /var on /mnt/var ?

[Alan McKinnon]

On 11/09/2013 04:02, gottlieb@×××.edu wrote:
> On Tue, Sep 10 2013, Alan McKinnon wrote:
>
>> On 10/09/2013 18:57, gottlieb@×××.edu wrote:
>>
>>> Alan McKinnon wrote:
>
>>>>> There's rules of thumb about this that will always work:
>>>>>
>>>>> No object in /tmp can be expected to survive successive invocations of
>>>>> the program that created the object, and never survive a reboot;
>>>>> No object in /var/tmp can be expected to survive a reboot
>>>>>
>>>>> The best place for temp files, ironically, is ~
>>> I set tmpwatch and wipe_tmp so that files survive in /tmp and /var/tmp
>>> for a month.
>>>
>>> I don't like ~ for temp files since on some, admittedly rare, occasions
>>> I actually use the gnome gui file manager and don't want a huge ~.  I
>>> have long ago created ~/tmp (also cleaned after a month by tmpwatch) so
>>> the only problem is breaking the habit of placing short-term files in
>>> /tmp instead of ~/tmp.
>>
>> OK, I get it. I'd write all that temp stuff to /var/tmp so it doesn't
>> get nuked by something cleverly trying to manage /tmp.
>
> But A. McKinnon says (above) that an always valid rule of thumb is
> "No object in /var/tmp can be expected to survive a reboot".


There's another rule of thumb that's even more applicable:

"We always do it this way, except when we don't"

I use this to great effect all the time, usually when I'm determined to
get my own way at work. I highly recommend this approach, when used with
a good healthy dose of intelligence you can get awesome things done :-)


>
>>> I realize that habit is bad for my (system's) health, but still find it
>>> hard to break.  I shall try again. Perhaps this is very mild form of
>>> what intelligent smokers feel :-).
>>
>> There is no such thing as an intelligent smoker; there are only stupid
>> smokers :-)
>>
>> I'm a two-packs-a-day man myself, I speak from many years experience!
>
> I promise not to mention it again, but you really should quit.
>
> I consider one of my contributions to computer architecture is being at
> least a little influential in getting Per Stenstrom to quit.  At the
> time he was a rising star, who I felt would contribute greatly if he
> didn't get sick from the cigs.  He did quit and has certainly
> contributed.

You are correct of course, and I have no sensible answer for that :-)

The one answer I do have is "see earlier comment 8 paragraphs above"




-- 
Alan McKinnon
alan.mckinnon@×××××.com