Xubuntu Saucy on Mir/XMir

Mon Aug 5 10:48:17 UTC 2013

Hey Elizabeth,

first of all thanks for the very structured and valueable feedback. I
will comment on the individual issues inline:

On Mon, Aug 5, 2013 at 12:29 AM, Elizabeth Krumbach Joseph
<lyz at ubuntu.com> wrote:
> On Thu, Jul 18, 2013 at 11:14 AM, Thomas Voß <thomas.voss at canonical.com> wrote:
>> Hey Elizabeth,
>>
>> sorry for the long delay, we were quite busy fixing a lot of issues
>> encountered by the first wave of testers :-)
>>
>> However, most of the issues you were reporting should be fixed by now
>> and giving Mir a spin according to:
>>
>>   http://unity.ubuntu.com/mir/installing_prebuilt_on_pc.html
>>
>> definitely makes sense. Please let me know if you encounter any issues.
>
> When Mir hit 0.0.8 we decided to launch a broader round of testing,
> today we worked to collate all the major results that are currently
> blockers for us.
>
> Rather than a meeting at this point, we're looking for remarks
> addressing some of the following issues. In general we understand that
> the code needs to land soon for archive inclusion and that Beta Freeze
> is at the end of the month. We are coming up on a decision-making
> point on August 22nd where we need to decide whether or not this
> technology is ready for our users.
>

Thanks for clarifying your timelines here. The 22nd of August is
feature freeze iirc, and I think you want to make sure that you can
move forward with reasonable confidence in XMir. Is that correct?

> And a few questions about the following results: Are these results
> atypical?  What, in your view, is going to be the minimum hardware
> platform that can be supported by XMir at this point in time?  Antique
> hardware is generally not in use as far as we understand it though if
> our definition of such differs from your own we need to resolve that
> quickly. Is there anything we are we doing that is perhaps suboptimal?
>  Do you have any recommendations to improve the ability to boot
> Xubuntu using XMir?  Have you reviewed the ISO image one of our
> community members has created to make things easier for our testers
> (http://vanir.unit193.tk/mir/) and have any remarks as to changes you
> would make in producing an image on your own?
>

Issue 1 is by no means atypical, it is a known issue at this point and
we are busy fixing it, please see:

  * lp:~vanvugt/mir/switch
  * lp:~vanvugt/mir/fix-1199450

The first branch is a pre-requisite for composity bypass, too.

For minimum HW requirements: We consider common netbook setups as a
baseline for XMir, and whatever used to work with X (under the
assumptions of using free drivers) should work with XMir as well.

I don't think you are doing anything suboptimal, but I would like to
cross-check on the bootable ISO:

  * Which version did you include with the ISO? Is it 0.0.8? If so,
did you compile manually or did you use the system-compositor-testing
PPA? Depending on the version you have pulled in, I would propose to
remix a newer bootable ISO.

We are busy landing (X)Mir to the archive, and that should
significantly ease testing on your side, too. Please let us know if
you need any further help. I think we should iterate quickly to
resolve any outstanding boot/startup/login issue such that your users
can really test XMir.

At any rate: I'm downloading the image right now and will test it on
the HW I have available here.

> We look forward to engaging in dialogue via e-mail as this is a
> complex matter that most likely cannot be resolved in a single IRC
> meeting or Google Hangout.
>

Very good point, happy to proceed via mail.

> tl;dr list of bugs/issues:
>
> https://bugs.launchpad.net/mir/+bug/1208242

This is a duplicate of:

  https://bugs.launchpad.net/mir/+bug/1199450

and we are busy reviewing and landing the fix linked to that bug.

> https://bugs.launchpad.net/ubuntu/+source/plymouth/+bug/1198741

This is a duplicate of:

  https://bugs.launchpad.net/mir/+bug/1102756

and we are busy fixing it. Obviously, it's critical and marked as such.

> https://bugs.launchpad.net/mir/+bug/1208250

We are looking into the issue and are trying to reproduce. I have set
the importance to critical.

> Multi-monitor support is poor
>

Agreed, multi-monitor is the remaining feature and multiple MPs are
ligned up against Mir to add the required functionality. I'm confident
that we can get the required features into Mir and wire it up to
X(Mir) to support Xrandr flawlessly.

> Full version of issues with details:
>
> ISSUE 1: Works, but jumpy on text being typed, a bit slow and all text
> from session writes to console
>
> VGA compatible controller: Advanced Micro Devices, Inc. [AMD/ATI]
> Wrestler [Radeon HD 6320]:
>
> XMir starts up, screen flickers and is laggy on events like changing
> workspaces and tab-changing application windows - often showing
> previous application for a moment. In xfce4-terminal it prints
> characters and then will delete them and show them again, when
> shutting down the text I typed in my session shows up in the console
> during shutdown (including paswords!) video of all here:
> http://www.youtube.com/watch?v=AfroweheIEU
>
> A lot of reports were like this:
>
> Laptop - ATI Mobility™ Radeon® HD 5870 - Useable, text input is very laggy.
>
> AMD Athlon2 X2, Radeon HD 4200 integrated graphic card: too much lag
> when typing. Otherwise, display (fonts, icons, etc.), sound and mouse
> behavior are perfect.
>
> Ati mobility Radeon 4000 family. Hp g42 notebook: At one point when i
> pressed a letter it printed on the screen only when the next was
> pressed. Same with back space and space. Everything else seems
> allright (only 5 min testing)
>
> Jumpy text bug: https://bugs.launchpad.net/mir/+bug/1208242
>
> Someone already reported the all-text-typed-in-console issue:
> https://bugs.launchpad.net/ubuntu/+source/plymouth/+bug/1198741
>

I think Issue 1 is captured in:

  https://bugs.launchpad.net/mir/+bug/1199450
  https://bugs.launchpad.net/mir/+bug/1102756

Do you think that the bug report accurately describes the issue?

> ISSUE 2: Poor multi-monitor support
>
> NVIDIA Corporation GF119 [GeForce GT 520] (rev a1)
> dual monitors where mirrored by default
> i see no option to change its mode, http://i.imgur.com/yHSAH2F.png
>
> Note: Multiple monitor setups are very important to many Xubuntu
> users.  This is considered a fundamental issue to one of our
> developers as to whether or not we can successfully deploy XMir.
>
> It has been reported that the multimonitor setting has not been merged
> to any mir session yet. (
> http://irclogs.ubuntu.com/2013/08/03/%23ubuntu-mir.html#t17:14 )
>

As noted earlier: Yes, missing right now and actively worked up.
Confidence is high that we can deliver a multi-monitor experience that
does not regress when compared to vanilla X. I'm happy to keep you
posted on our progress here. Please refer to the following bug report,
too:

  https://bugs.launchpad.net/mir/+bug/1102760

> ISSUE 3: Does not work at all
>
> Netbook had a string of apport failures on getting to desktop.  It
> booted but was useless.
> Hardware described here: https://wiki.ubuntu.com/skellat/netbook-profile
>
> Podcast production platform tested with newest ISO resulted in this
> on-screen image of uselessness using this hardware profile:
> https://wiki.ubuntu.com/skellat/audiobox-profile - Video of the whole
> boot process: http://www.youtube.com/watch?v=2Q1AdXsJj5E
>
> Other reports had similar fragmented screen errors, making it unusable:
>
> Intel Mobile Corp Mobile 4 series: http://imagebin.org/266445
>
> PC - GF GT 240 graphics card: http://lderan.co.uk/xmir.jpg
>

I'm triaging and looking into those reports over my day now, will come
back with more detailed information asap.

> Based on feedback so far, it seems at least 50% of the casual testing
> reports we're getting thus far are left in this position of having it
> be completely unusable.  The machines are unusable to the point that
> cameras pointed at screens are the only way we can gather data as we
> cannot harvest logs.
>

Again thanks for your feedback and the testing you are carrying out
with XMir. Fortunately, the issues you are reporting are almost all
known and being worked upon right now.

HTH,

  Thomas

> Bug here: https://bugs.launchpad.net/mir/+bug/1208250
>
> --
> Elizabeth Krumbach Joseph || Lyz || pleia2
> http://www.princessleia.com