dnf update vs Software Udpates

[Andreas M. Kirchwitz]

Suvayu Ali <fatkasuvayu+linux at gmail.com> wrote:

>> I hope this will be done *fast*, because I have to "clean all"
>> *everytime* checking for updates. Otherwise, no updates are shown, even
>> though they exist. This is a major bug.
>
> I'm sorry but "clean all" is not necessary at all!  "clean metadata" or
> "clean expire-cache" should be sufficient.  

In practice, there's not much of a difference between "clean all"
or just "clean metadata". Because both require the update/upgrade
command to download all stuff from the network and build to whole
meta database from scratch, even if that wouldn't be necessary.

For the typical desktop PC, "clean metadata" is not a time-saver
compared to "clean all". Just more letters to type.

"clean metadata" may make sense if you add "--disablerepo=fedora",
because this really saves bandwidth and CPU for the update/upgrade.

Btw, "clean expire-cache" (or --refresh) won't do the trick to
get the latest updates available. (However, it's still better
than just update/upgrade without anything else.)

The design of yum/dnf is somewhat flawed. I totally understand that
some caching makes sense so that metadata isn't checked, downloaded
or built for every single call to yum/dnf.

However, if somebody runs "dnf upgrade" on the command shell then
he clearly wants the latest updates. Right now! No caching or other
magic involved. That's the whole point of running "dnf upgrade"
manually, otherwise the user would have left the whole updating
business to some automated background task.

> That said, I sometimes do not understand what's the harm in getting
> updates few hours later.

Caching might be cool for automated tasks (eg, cron jobs or background
processes) and also for some actions that do not require up-to-date
metadata.

But if the user wants to download an update he should get the update.

This issue comes up for years. And there's always this debate
about "clean all" vs. "clean metadata" which in fact isn't a big
difference in the real world, and - what is much more important -
both won't solve the underlying problem, so the very same issue
comes up again and again.

	Greetings, Andreas