use -N instead of -n for useradd

[xsuchy]

-n is downstream patch in Fedora and RHEL now. It was used in past. Since RHEL6 the option -N is available.
Let use it.

But that means if you want to build for RHEL5 you have to alter config_opts['useradd'] and return back -n option.

Resolves: #1050

[Conan-Kudo]

Does this allow us to drop the useradd overrides we've been using for all the non-RH/Fedora chroots now?

[Conan-Kudo]

And in that case, we should put the override in the RHEL 5 configs as part of this PR and drop the overrides everywhere else.

[xsuchy]

I really hope that no one uses rhel5 config today.
I know of Brew people.

[praiskup]

Does this allow us to drop the useradd overrides we've been using for all the non-RH/Fedora chroots now?

I'm not sure. The difference for those chroots would be:

- '/usr/sbin/useradd -o -m -u {{chrootuid}} -g {{chrootgid}} -d {{chroothome}} {{chrootuser}}'
+ '/usr/sbin/useradd -o -m -u {{chrootuid}} -g {{chrootgid}} -d {{chroothome}} -N {{chrootuser}}

Is it safe to use -N (capital) in all of those chroots that currently
override? In the new RH distributions, -n is just an equivalent to -N (so
the change Mirek is doing is no-op).

I know of Brew people.

They maintain their own configs, and typically also run an older Mock version
anyway.

[Conan-Kudo]

Well, if -N has been the upstream one since RHEL 6 (2010), then everyone should have that option already. I didn't know what -n did when I wrote those stanzas and just dropped it. The correct thing to do is to put -N in and use it properly.

[praiskup]

@Conan-Kudo we discussed with @xsuchy today - and it seems that removing the useradd overrides from config would make the mock-core-configs code (released asynchronously, e.g. we plan to release ASAP because of the new Amazon config fiules) incompatible with the old mock releases. So when removing the overrides, we should at the same time bump the mock version requirement.

Therefore we should either merge as-is (and do the config change later on) or wait. We are not yet ready to wrap a new mock.rpm release.

[praiskup]

@Conan-Kudo what do you think?

[Aemn567]

n

[Conan-Kudo]

We should wait on this, then. It's not urgent to land, and I'd rather this be a complete change.

[praiskup]

I have no objections, @xsuchy would you mind updating this PR then so we have it prepared?

[xsuchy]

Updated.
I did not modified the eol/* configs. They will work even without modification. I hesitate to touch anything in eol/* (unless it is deleting operation).

[Conan-Kudo]

I would say you should probably fix the EL5 configs and add an override there, but I don't care that much.