yum clean bug

[Nicolas Mailhot]

Jeff Spaleta wrote:
> On 12/9/05, Nicolas Mailhot <nicolas.mailhot at laposte.net> wrote:
>> So this is not-disposable data needing protection. I translate this into
>> data that must be preserved. If it's not preserved it's not really
>> protected, right ?
> 
> More evidence that communicating in English is inherently prone to
> miscommunication.
> 
> Disposable and re-generatable are not perfectly equivalent concepts. 
> While by definition data in the cache filesystem can be re-generated
> when needed for tool operation, nothing says it preferable to
> re-generate if it can be avoided.  Yum's clean policy is coded to
> avoid unnecessary re-generation.
> 
> In fact, because cached items are defined to be resource intensive to
> generate in the FHS definition for /var/cache ... i think that
> automatically implies that well coded tools treat the cache
> conservatively when removals are done to avoid wasteful calculation or
> i/o when re-generating cache later that could have been avoided.  Its
> the difference between doing something just because the action is
> strictly allowed and refraining from doing something to avoid doing
> more unncessary work later to re-generate time or i/o intensive cache
> when it can be avoided.

Ok. So we're back to the cache aspects, and you agree these is no
mysterious data needing to be preserved in /var/cache/yum, only standard
cache files.

1. shouldn't yum remove every part of /var/cache/yum not covered by a
repo (enabled or not) since obviously it can not use it for optimization
purposes ? (this would correspond to standard home-keeping after repo
removal/renaming). Other tools routinely remove this kind of leftovers
or at least flag them (yum leftovers can be huge, a repo can consume
hundreds of MiBs in old downloaded packages, and people seem to like
renaming repositories). Unless there is some other reason to keep it ?
(but then it does not belong in /var/cache/yum anymore)

2. couldn't the user be allowed to easily purge the cache (and no
enablerepo=* does not count as easily, especially considering Seth finds
rpm flags "cryptic"¹) ? Even Mozilla/Firefox/etc allow this, and they
have to manage users that would never approach the command line let
alone yum

Remember, this is yum's private repository, only yum writes in it (the
FHS allows nuking the cache of someone else not messing in any other way
with it), and the data inside can be re-generated when needed (your words)

And this was my whole argument, apologies to all the people who got it
twenty messages before.

¹ Or whatever the exact term he used last time, I don't remember it exactly

-- 
Nicolas Mailhot