How should one update from the commandline?

[rokejulianlockhart]

@hcvv,

@hcvv

As already said above, you really do not need that zypper ref every time you sneeze. It is done automatic when needed. There can be a reason to do an extra one, but in such cases it is very obvious and logical to do so.

If I didn’t do it, couldn’t I potentially update to the non-latest packages? I learnt this after testing a dup without a refresh and a dup with, to which I believe the result was a slightly different package choice.

@hcvv

updating a Tumbleweed system only needs zypper dup. Nothing less, nothing more.

That’s why I’m asking – why do you think so? Also, aren’t flatpak and snap rather excluded from that statement? (zypper isn’t going to touch them!)

@hcvv

I have no idea why you make such a huge statement with all sorts of actions without any possibility to interfere. This will roll on and you probably at one day will stare at a lot of lines scrolling before your eyes and you not knowing what program is telling what to you. Doing a ^C then will obviously be much too late to stop disaster.

This is meant to be entirely automatic. The script -a should allow me to investigate any problems, right? It’s not always easily readable, but I don’t know of a better logger (pwsh’s Start-Transcript screws stuff up by routing it through another Out-* instead of spawning a new shell).

@hcvv

I am not sure which user is supposed to submit this command

Myself. This is for a personal machine.

@hcvv

when [the user] is not root already (in which case I do not understand the use of su)

This seems contradictory to me. If the user isn’t authenticated as the superuser, surely I do need to elevate permissions. With zypper and flatpak, I don’t get PolKit appearing if I don’t provide credentials beforehand like snap does, and since this needs to be automatic, that wouldn’t be ideal anyway.