Re: [gentoo-user] portage-2.2.0_alpha38 & --depclean

[Alan McKinnon]

Apparently, though unproven, at 23:39 on Tuesday 07 June 2011, Mark Knecht did 
opine thusly:

>    I have no problem with saying someone needs to understand what less
> does. less isn't important. It's just the example at hand today. The
> 'problem' that I'm trying to get closer to answering is how does
> anyone other than a Gentoo dev, assuming some reasonable amount of
> effort, know that less isn't called by some script somewhere during
> the init process? How does one come to understand that maybe less is
> just as import as python is to the emerge process?  (and I know it
> isn't...)
>
>    What I didn't like about this issue popping up yesterday is that it
> altered the idea that average users never touch anything in @system.
> Iin fact, TTBOMK I've never in 11 or 12 years of running Gentoo ever
> done an emerge -C on a @system package until this morning when I
> removed nano.

OK, now we're tracking.

In the specific case of less, the answer is self-evident - it isn't needed. A 
dev would just know that. More likely, he would assume he knows that.

In the general case, they suck their thumbs and guess. Some think more than 
others before they guess, they should all do some basic tests to catch severe 
errors before committing changes and additions, and all of them rely on 
unstable users finding other oddities and bugs.

flameeyes gave some hints and clues into how this works on his blog recently:

http://blog.flameeyes.eu/2011/05/25/psa-packages-failing-to-install-with-new-
openrc-based-stages-missing-users-and-groups

It's specific to openrc, but if you follow his blog it's easy to read between 
the lines to see what he's getting at usually.

I don't think I've ever met a dev that releases code any other way :-)

None of the above is fact and all of it is my opinion but I do think I'm close 
to the mark.


-- 
alan dot mckinnon at gmail dot com