Re: "Basic graphics mode" feature and criterion discussion

[Adam Williamson]

On Tue, 2019-03-26 at 14:30 -0600, Chris Murphy wrote:
> My two cents:
>
> If there's a fallback option, and if the user selects it, they
> shouldn't end up in an unambiguous state. Right now we're seeing
> systems hanging. I'd rather see a crash than a hang where the user
> can't get to a shell, and sees no useful information on the screen
> that tells them why they're hung up; or even generically that "by now
> you should see a login screen and if you don't we've faceplanted
> somewhere, sorry".
>
> Maybe split the criterion:
>
> Basic criterion: installation media must have a basic video boot entry
> that uses the accepted fallback boot parameter(s), e.g. nomodeset. The
> criterion just means the media must have the menu entry, and that it
> passes something intentional for this purpose as a boot parameter.
>
> Final criterion: installation media's basic video boot entry should
> either work (we get a successful graphical boot), or it should
> faceplant in some unambiguous way.
>
> If it's not possible to ensure basic video either works as intended or
> faceplants unambiguously; then I suggest dropping any beta or final
> criterion. And also I wonder if it's at all useful to include some
> kind of heads up description for the basic video boot entry? Like,
> "this may not work at all" or "wait 5 minutes for graphical boot,
> after 10 minutes assume it failed". Haha - I have no idea. Just
> something so they aren't waiting around going WTF? Now what?

So kinda aggregating all the response to this discussion, I propose we
go with a modified version of Chris' proposal.

1. We retain the 'entry must exist' part of the criterion at Basic (but
at that point it does not have to *work* - the idea is to ensure it's
testable so we catch bugs in it at that point)

2. We move the rest of the criterion to Final and tweak it a bit to
specify that it must be at least somewhat capable of reaching the
installer or desktop. We do not adopt Chris' "or faceplant
unambiguously" proposal, people seemed to prefer just requiring it to
work.

So, here's the specific text I propose. At Basic we would change this
requirement:

"The boot menu for all supported installer and live images should
include an entry which causes both installation and the installed
system to use a generic, highly compatible video driver (such as
'vesa'). This mechanism should work correctly, launching the installer
or desktop and attempting to use the generic driver."

To read only:

"The boot menu for all release-blocking installer and live images
should include an entry which causes both installation and the
installed system to use a generic, highly compatible video driver (such
as 'vesa')."

i.e. remove the second sentence (and change 'supported' to 'release-
blocking' - that is a better form of words that should have been used
all along).

At Final we would add this requirement:

"The generic video driver option ('basic graphics mode') on all
release-blocking installer and live images must function as intended
(launching the installer or desktop and attempting to use a generic
driver), and there must be no bugs that clearly prevent the installer
or desktop from being reached in this configuration on all systems or
on wide classes of hardware."

How does that sound to everyone? Thanks!

"Basic graphics mode" feature and criterion discussion

[Ben Cotton]

On Mon, Apr 1, 2019 at 3:04 PM Adam Williamson
 wrote:
> So, here's the specific text I propose. At Basic we would change this
> requirement:
> > <DELETIA>
>
> To read only:
>
> "The boot menu for all release-blocking installer and live images
> should include an entry which causes both installation and the
> installed system to use a generic, highly compatible video driver (such
> as 'vesa')."

This seems reasonable to me. +1
> At Final we would add this requirement:
>
> "The generic video driver option ('basic graphics mode') on all
> release-blocking installer and live images must function as intended
> (launching the installer or desktop and attempting to use a generic
> driver), and there must be no bugs that clearly prevent the installer
> or desktop from being reached in this configuration on all systems or
> on wide classes of hardware."

Am I correct in my interpretation that this failing on a narrow set of
hardware or configurations would NOT be considered a blocker?