Re: [Ayatana] Restart Required

[Matthew Paul Thomas <mpt@xxxxxxxxxxxxx>]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Frederik Nnaji wrote on 04/09/10 09:03:
>
> Hello MPT ;)
>
> On Tue, Aug 3, 2010 at 21:02, Matthew Paul Thomas <mpt@xxxxxxxxxxxxx
>...
>> In the update handling specification, I've specified that it should.
>> <https://wiki.ubuntu.com/SoftwareUpdateHandling#alert> (For the
>> implementation, see
>> <https://launchpad.net/ubuntu/+spec/foundations-lucid-distiguishing-updates-that-require-restart>.
>> I don't know how far that ever went.)
>
> The blueprint you linked here was superseded by another one.. here:
> https://blueprints.launchpad.net/ubuntu/+spec/upgrading-running-software/
> The original blueprint holds implementation ideas (tagging) and
> questions (will tagging apps require changes to Soyuz?), while the one
> superseding it suggests a brainstorming session.. did such a session
> happen? Was it documented by anyone?

Not that I know of. Perhaps we could have another go at UDS Natty.

> Another more "Restart Required" specific spec lies here, created by
> Joel Ebel:
> https://wiki.ubuntu.com/UpgradingRunningPrograms

I've never seen that page before.

> does Joel's spec somehow interfere with the changes you propose in your
> spec?

No -- as far as I can tell, neither of them come to an implementable
conclusion. :-)

>...
>>>> so then red is the wrong color
>>>
>>> Depends on what fix isn't in place. Leaving a remote root exploit in
>>> place seems pretty red to me.
>>>...
>>
>> But you're no less secure, at that moment, than you were when Ubuntu
>> knew the update was available but you hadn't installed it yet. Should
>> the icon be red then too? And if not, why should it be red now?
>
> what is your personal shot at this?

Mark designed the current presentation, so it wouldn't be appropriate
for me to answer that question, sorry.

>                                     I think apt doesn't warn about this,
> keeps its interaction dialog lean and purposeful. Our discussion here
> leads me to the conclusion that we want updates to be a seamless
> service: interruptions of human workflow are bad app behaviour and
> deserve to be flagged, so USER can avoid them timely.

I was supposing that there would be good reasons to delay updates that
require a restart until a more convenient time -- for example, until
you're intending to shut down the computer anyway. (Shutting down and
later starting up is just as good as restarting, for this purpose.)

If there aren't good reasons to delay updates that require a restart,
then maybe we don't need that text at all. But we'd still need an
equivalent system for applications that will require relaunching.

- -- 
Matthew Paul Thomas
http://mpt.net.nz/
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.10 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/

iEYEARECAAYFAkyGKfMACgkQ6PUxNfU6ecpeJQCfY8037934XdM6k4abj4lidtpD
VhQAoLKw9+hcEZsR5ymJoHVFVBKAp7zW
=V5P2
-----END PGP SIGNATURE-----