Fwd: My Contribution For The Next QA Status Report : IRC Release Meeting

[Null Ack]

As part of discussions on IRC Sunday, it was mentioned that its policy
to mark package update requests as wishlist importance. I've since
looked up the process policy and the wording is:

Wishlist: a request to add a new feature to one of the programs in Ubuntu.

    * These aren't always bugs, but can be ideas for new features
which do not yet exist.
    * These can also be requests to have software packaged for Ubuntu.
    *  If it is non-trivial to implement, it should rather be written
as a feature specification, see FeatureSpecifications

This isnt best practice because:

1 It creates a false impression to users about the requests
"importance" and the actual meaning of importance in this case is
non-intuitive to them.
2 It provides no guidance to packagers about how important this is for
the user experience

I intend to discuss this with the Bug Squad team, gain agreement and
change the doco unless I get some feedback from QA people here. So
speak up if you dont agree.

I'd like to comment on some themes from Sunday's IRC talk in relation
to this....I understand that voluntary devs will scratch their own
itch and work on whatever they want. It would be a cop out of the QA
team in my view though not to help improve processes (which is listed
as one of the teams objectives) and encourage focus on how the user
experience is effected when looking at priorities. There's no harm in
implementing priorities for package update requests - it does not
restrict the freedom for voluntary contributors but it does provide
guidance and help emphasise the importance of thinking about the user
experience. Priorities exist for other bug types, and for good reason.

cheers



---------- Forwarded message ----------
From: Null Ack <nullack at gmail.com>
Date: 2008/9/13
Subject: My Contribution For The Next QA Status Report : IRC Release Meeting
To: ubuntu-qa at lists.ubuntu.com


As the weekend may put a delay on getting a response on how best to
provide my contirbutions for the QA status report I'm providing it
here.

Generally there was a range of things that disappointed me with the
quality aspects of the 8.04 LTS release and even a casual look at user
feedback on the ubuntu forums shows some strong negative feedback from
many people. This extended to upstream critics as well, such as the
audio implementation for Hardy basically being labelled as botched by
upstream. This is why I am especially keen to contribute to this part
of the QA team's work for future releases.

Also while I have accepted the reasons for my first application for
membership into the QA team being declined, I have a sense that the
amount of work I'm doing is not clearly visible to you all. By
contributing to the QA status report it is one way of showing the
significant amount of time thats gone into my work.

1. Items I Consider Critical for Resolution Prior To Beta:

257403 - Update Flash plugin 10 to the new RC - wishlist
246911 - [Wishlist] please add libnspr4-0d to ia32-libs

These two bugs really should not have a priority of wishlist. The user
experience on flash is a core user experience that people naturally
expect to just work. It bothers me how people deciding on the priority
have become so disconnected with the user experience that we label
fundamentals as wishlist - especially when the current revisions are
unstable development revisions not production releases. I don't have
time to create a litany of known issues with the current version of
flash, but suffice to say its not good and the path to a good flash
experience involves synching to upstream.

251910 - gvfsd-trash crashed with SIGSEGV in g_idle_funcs()

This occurred for me today again, so its still current in Intrepid. Im
unsure if it should go upstream or not but since its the weekend I
probably wont see Sebastien on IRC until Monday. There is problems in
a wider context with debugging and how users are engaged for
backtraces which I am working on separately to this email.

258743 - NM 0.7 Fails To Set Custom MTU

Im unable to set valid maximum transmission unit values in NM. More
seriously, the assignment of static IP addresses is not persistant.
Right now Im forced to use DHCP with mac reservation (which I dont
want to do) to have an easy network experience. A related but seperate
bug is number 256054.

243803 - Grub Installs To Wrong Root Path

On both ubuiqity and alternate there is a hardware specific problem
that results in the wrong boot path and no boot. Ive seen this on my
old test machine, the ubuntu forums and in the blogsphere. I no longer
have test access to that hardware so I can't personally help with
testing any longer but it is a real bug that is problematic for new
users to diagnose and resolve.

Generally for beta my view is that users should be able to reliably
install the software, setup their network, browse websites including
flash and not have processes segfaulting.

2. Items I Consider Critical For Resolution Prior To Production:

246269 - Switched from vesafb to uvesafb, but uvesafb can't work without v86d

I cant run high resolution modes in virtual consoles and on bootup
there is kernel log errors about the framebuffer.

256494 - compiz turned on again during update after I had desktop
effects turned off

An annoyance where compiz does not respect my wish to be off where it
turns itself on at each reboot

262027 - Nvidia GLX Upgrade Erroneously Reports Success to Synaptic on Fail

Bug in driver install where it aborts with a resolve manually yet
tells Synaptic the install went ok

259017 - Audigy2 No Audio By Default

Someone changed the default toggle in the audigy2 driver to be
digital. Most users have an analogue speaker setup and will not have
any sound working until they manually tell alsamixer they dont have a
digital amplification setup for their speakers.

255929 - Totem Fails On Multiple Audio and Subtitle Streams

Video experience is broken for content that has multiple audio streams
or multiple subtitle streams

257818 - Totem Fails To Deinterlace

Totem is not deinterlacing content that is interlaced using the
standard gstreamer back end even though the gstreamer framework has
the capability to do so

263153 - gstreamer0.10-ffmpeg caused crash on some ASF file and many
other problems to do with an old ffmpeg

Again I dont have time to create a litany of the issues caused by
having an old ffmpeg but its suffice to say that the gstreamer ffmpeg
plugin is a core component to a users video experience on Ubuntu
because it contains demuxers and decoders that are widely deployed on
the internet. There is a bunch of bugs in LP that have the root cause
of not being up to date with SVN of ffmpeg. Ubuntu needs to achieve a
process that routinely synchs to ffmpeg SVN on a regular basis.

3. Items That Would Be Nice To Resolve That Shouldnt Take Much Work

My little wishlist if your will:

263819 - BluBuntu Meta Package Misses Ubuntulooks
263817 - GPAR2 Incorrectly Uses The Ubuntu Icon

Thanks

Nullack