Johannes Schauer Marin Rodrigues <jo...@debian.org> writes:

> I recently (with version 1.3.2) extended the documentation for unshare mode in
> the mmdebstrap  manual page to also cover these two files:
>
> https://gitlab.mister-muffin.de/josch/mmdebstrap/commit/46fc269b549abe89d99e63addba0813bcbc938ac
>
> Does this answer some of the questions you had or do you think I should add
> more?

I like the docs. When debugging problems it's helpful to

- have a clear error message that says what the problem is

- have a clear connection between the error message, and a chunk of the
  docs that talks about that failure

Here I had:

    W: no entry in /etc/subuid for dima
    E: invalid idmap

and with the older mmdebstrap:

    newuidmap: uid range [1-2) -> [100000-100001) not allowed
    E: newuidmap 2086656  0 60017 1 1 100000 1 failed: 
    E: child had a non-zero exit status: 1
    E: chown failed

Can we change "W: no entry in /etc/subuid for dima" to something like
"W: no entry in /etc/subuid for dima: mode=unshare will fail; see THIS
section of the docs", or maybe make it an error? If the docs contained
the exact error message we would see with this issue, that would be
super helpful too. Do you know why the older mmdebstrap has a different
error message? Is it something you changed in the code, or is there
something about that machine that's causing it?


> There are two problems:
>
>  2) whatever method you use to create new users does not create these
>  entries

I don't know why they're missing. It's an old install of sid,
continually being updated: /etc goes back to 2006! I don't think I ever
did anything funky with the users, but who knows. It's not an mmdebstrap
problem, in any case.


> I have a patch for you that should fix this problem in the sense that
> mmdebstrap should not choose the unshare mode anymore. If you like, apply the
> following to mmdebstrap from unstable:
>
> https://mister-muffin.de/p/ZwXV.diff

Neither of your patches apply to the current mmdebstrap from unstable
(I'm at 5d24b65 in the git tree). If you want me to test, you should
give me another patch. But I trust you to fix it, and I don't NEED the
patch, since I now know to fix the /etc/subuid and /etc/subgid. So you
can just apply the patch to the tree and close this bug.

Thank you very much for your help!

Reply via email to