<html>
  <head>
    <meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
  </head>
  <body>
    <p><br>
    </p>
    <div class="moz-cite-prefix">On 24/10/2022 17:34, Christopher Obbard
      wrote:<br>
    </div>
    <blockquote type="cite"
      cite="mid:3a428f220e78ceaa737e9d88d5f66dce409774ef.camel@collabora.com">
      <pre class="moz-quote-pre" wrap="">I have proposed[1] to check if systemd-resolved is available at
runtime, to at least let users know *why* name resolution doesn't work
inside their fakemachine over letting the user debugging it themselves.</pre>
    </blockquote>
    <blockquote type="cite">
      <pre class="moz-quote-pre" wrap="">[1]: <a class="moz-txt-link-freetext" href="https://github.com/go-debos/fakemachine/pull/115">https://github.com/go-debos/fakemachine/pull/115</a></pre>
    </blockquote>
    <p>That's nice! I thought of it as well, and I was wondering if
      systemd-resolved (and possibly other services) should be listed
      under Requires= instead of Wants= (talking about
      <a class="moz-txt-link-freetext" href="https://github.com/go-debos/fakemachine/blob/main/machine.go#L288">https://github.com/go-debos/fakemachine/blob/main/machine.go#L288</a>).
      But then, I noticed commit
      4c60b85a8302f0fa544adae73f0649726034711c, and why using Wants= is
      the intention. So your approach works better.</p>
    <p><br>
    </p>
    <blockquote type="cite"
      cite="mid:3a428f220e78ceaa737e9d88d5f66dce409774ef.camel@collabora.com">
      <pre class="moz-quote-pre" wrap="">Perhaps we should add systemd-resolved to Suggests in
debos/fakemachine? Adding it as a Depends/Recommends would break users
who have some other package on their machine handling name resolution.

@Arnaud, how does that sound?</pre>
    </blockquote>
    <p>Well, I'm not sure for the packaging part. If fakemachine needs
      systemd-resolved to be functional, then it should be a Depends.
      That's what Depends is for.<br>
    </p>
    <p>At a quick glance, there's no reverse dependencies of fakemachine
      / debos. So the only users that would be surprised by the change
      are the ones who installed it explicitly, so we can assume those
      are technical users and they'll find their way? But then, what if
      there are some installations on servers (think builders), and the
      upgrade breaks the name resolution? Not a nice surprise.</p>
    <p>OTOH, an error message saying that "<span class="blob-code-inner
        blob-code-marker js-code-nav-pass" data-code-marker="+"><span
          class="pl-s">/lib/systemd/systemd-resolved is missing", plus a
          Suggests: systemd-resolved, both together gives a strong hint
          regarding what should be done. It sounds sensible as well.<br>
        </span></span></p>
    <p><span class="blob-code-inner blob-code-marker js-code-nav-pass"
        data-code-marker="+"><span class="pl-s">Sorry, that's not really
          a clear answer :) <br>
        </span></span></p>
    <p><span class="blob-code-inner blob-code-marker js-code-nav-pass"
        data-code-marker="+"><span class="pl-s">Best,</span></span></p>
    <pre class="moz-signature" cols="72">-- 
Arnaud Rebillout / Offensive Security / Kali Linux Developer</pre>
  </body>
</html>