<html>
<head>
<meta content="text/html; charset=utf-8" http-equiv="Content-Type">
</head>
<body text="#000000" bgcolor="#FFFFFF">
<div class="moz-cite-prefix"><br>
In the context of a snap running on Ubuntu Core, the all-snap
version of Ubuntu, the gadget snap (which defines behaviors
specific to that particular device) would describe additional
auto-connection rules.<br>
<br>
So in your case, a switch that had certified your OpenSwitch NOS
would specify that, when installed, OpenSwitch plugs and slots be
connected appropriately.<br>
<br>
Think if it as that "the user OR the device manufacturer can take
this decision" and the gadget snap is how the device manufacturer
would do so.<br>
<br>
Would that be useful?<br>
<br>
Mark<br>
<br>
On 13/07/16 10:54, MikeB wrote:<br>
</div>
<blockquote
cite="mid:CAAuRgKAm57aJ5GQO9Jw5Un23U52150RkrHvW0RAjWn8J73bR0A@mail.gmail.com"
type="cite">
<div dir="ltr">
<div class="gmail_default"
style="font-family:tahoma,sans-serif;font-size:small">Yes, I
see your point.</div>
<div class="gmail_default"
style="font-family:tahoma,sans-serif;font-size:small"><br>
</div>
<div class="gmail_default"
style="font-family:tahoma,sans-serif;font-size:small">Perhaps
a concept of optional vs required is needed. </div>
<div class="gmail_default"
style="font-family:tahoma,sans-serif;font-size:small"><br>
</div>
<div class="gmail_default"
style="font-family:tahoma,sans-serif;font-size:small">Mike</div>
</div>
<div class="gmail_extra"><br>
<div class="gmail_quote">On Wed, Jul 13, 2016 at 12:37 PM, Kyle
Fazzari <span dir="ltr"><<a moz-do-not-send="true"
href="mailto:kyle.fazzari@canonical.com" target="_blank">kyle.fazzari@canonical.com</a>></span>
wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0
.8ex;border-left:1px #ccc solid;padding-left:1ex">
<p dir="ltr">Hey Mike,</p>
<span class="">
<p dir="ltr">On Jul 13, 2016 3:46 AM, "MikeB" <<a
moz-do-not-send="true"
href="mailto:mabnhdev@gmail.com" target="_blank"><a class="moz-txt-link-abbreviated" href="mailto:mabnhdev@gmail.com">mabnhdev@gmail.com</a></a>>
wrote:<br>
> I understand why you want manual connections for
some "critcal" interfaces. However, if you're going to
require manual intervention, the snapd should be smart
enough to wait for the required connections before
starting up any daemons that have plugs for those
"critical" interfaces. Perhaps even nagging about
unconnected plugs in the snap.</p>
</span>
<p dir="ltr">I understand where you're coming from here.
However, I can imagine use-cases where such a daemon can
actually run successfully without one of its plugs
(particularly when seccomp starts using ERRNO), and have
extra functionality if it's connected. If snapd didn't
start services until all their plugs were connected, such
a scenario would be impossible. Also, keep in mind that
the user can disconnect interfaces at any time, so ideally
daemons would be able to handle such things (though I
realize that's not always the case).</p>
<p dir="ltr">Just my two cents.</p>
<span class="HOEnZb"><font color="#888888">
<p dir="ltr">Kyle</p>
</font></span></blockquote>
</div>
<br>
</div>
<br>
<fieldset class="mimeAttachmentHeader"></fieldset>
<br>
</blockquote>
<br>
</body>
</html>