2011-11-22

Weekly Summary

  • Most blueprints are approved now; one to be written, one to be approved; work items are within confirmed bounds. (pitti)
  • versions.html is now working again, and against precise. (pitti)
  • Ported session-installer to GTK 3 and dropped gnome-codec-install, getting rid of some old libraries. (pitti)
  • Lots of progress on Apport client-side duplicate detection. (pitti)
  • Adding more test to OneConf, still some to go (didrocks)

  • Worked on tricky Qt accessibility bug, preparing an SRU for this week (agateau)

  • Packaged massif-visualizer, filed ITP on Debian and also pushed to REVU (agateau)
  • Started on CUPS filters Blueprint, joining CUPS filters discontinued by Apple with the PDF CUPS filters for new CUPS filters upstream package (tkamppeter)
  • Oneiric SRUs for Ghostscript, now printing to PostScript printers under Oneiric should work correctly (tkamppeter)

  • In preparation of empathy 3.3, filed MIR for clutter-gst (ken-vandine)

Agenda

  • Insert things to discuss at meeting here..

Partner update

Unity

Software Center

Kubuntu

JonathanRiddell

  • I'm back! Did you miss me?
  • Currently working in UTC-4 timezone
  • Going through the Qt/KDE merges, lots to catch up on
  • TODO: get Kubuntu approved as LTS by tech board
  • TODO: oneiric langpack update to include rekonq translations

X.org

IRC log Western edition

IRC log Eastern edition

hey bryceh RAOF robert_ancell and TheMuso meeting time!! https://wiki.ubuntu.com/DesktopTeam/Meeting/2011-11-22
 [8:35 AM] <RAOF> Good morning all :)
 [8:35 AM]  * bryceh waves
 [8:35 AM] <jasoncwarner_> great. lets get this thing started. 
 [8:35 AM] <jasoncwarner_> bryceh RAOF X update? 
 [8:35 AM] <TheMuso> Good morning.
 [8:36 AM]  ⇐ dbarth and mterry quit  
 [8:36 AM] <bryceh> jasoncwarner_, today working on an xdiagnose update, mostly clarifying dialog wording and such
 [8:37 AM] <bryceh> yesterday got 5 systems updated to precise, ran into various bugs including #876499 which seems to be bad - broken nvidia on fresh installs
 [8:37 AM] <jasoncwarner_> bryceh:  only fresh installs? upgrades are fine? 
 [8:37 AM] <bryceh> also worked through several other non-X bugs like bug #458482 and bug #893700 and others.  Think fixes for those are now int he works
 [8:37 AM] <ubot2> Launchpad bug 458482 in usb-creator "usb drives imaged with usb-creator fail to boot on some systems" [High,Triaged] https://launchpad.net/bugs/458482
 [8:37 AM] <ubot2> Launchpad bug 893700 in debootstrap "pbuilder create of a precise image fails on perl-modules / perl conflict" [Undecided,Fix released] https://launchpad.net/bugs/893700
 [8:38 AM] <bryceh> jasoncwarner_, I did one upgrade and ran into one bug where if I had extras enabled it failed; didn't look to see if that's reported already
 [8:39 AM] <bryceh> also ran into a slew of problems trying to get the installer to format the drives the way I wanted, it kept crashing.  But then just let it format them the way it wanted and all was good.
 [8:39 AM]  → robert_ancell (~bob@ppp121-44-212-7.lns20.syd7.internode.on.net) joined
 [8:39 AM] <bryceh> anyway, good news is other than the nvidia bug, X seems solid.
 [8:39 AM] <bryceh> and I think the nvidia bug might be an error in ubiquity.
 [8:39 AM] <bryceh> jasoncwarner_, on topic of multimonitor
 [8:39 AM] <RAOF> I think that nvidia dislikes something about the 3.2 kernel, too.  Or possibly something in UEFI/efifb/3.2.  It mostly fails to actually display the xserver.
 [8:40 AM] <bryceh> I've associated a slew of bug #'s to the blueprint (probably more than is reasonable; we will likely want to prune it down as w ego forward)
 [8:41 AM] <bryceh> I've also started work on the refactoring for libxrandr-utils.  Still waiting on my FDO registration but got the first few bits pulled into a local git tree
 [8:41 AM] <bryceh> next step there is I'm going to implement a test suite; seems there are two choices in X
 [8:41 AM] <bryceh> 1. 'check', used by libxcb
 [8:41 AM] <bryceh> 2. 'glib' used by xserver and a couple other things
 [8:42 AM] <RAOF> glib is also used by bits of unity, but I've played with check before.
 [8:42 AM] <bryceh> looks like #2 is the more "official" but since it depends on gtk looks like it needs additional packaging magic to make it condutional
 [8:42 AM] <bryceh> figure I can set up glib and if we need to move to check later it's not a big deal
 [8:43 AM] <RAOF> Yeah, that should be reasonably easy.
 [8:43 AM] <bryceh> as to bugs in general for precise,  they're light right now as would be expected
 [8:43 AM] <jasoncwarner_> bryceh: I'll check out the BP for multi-monitor later. Do you think the BP covers the scope of the problem for precise? 
 [8:43 AM] <bryceh> just 3 open, including the nvidia one I mentioned above.   That's really the only serious one.
 [8:44 AM] <bryceh> jasoncwarner_, I do.  I think the bulk of the "problem" that people are complaining loudly about right now are more to do with Unity usability problems, and I gather the Unity guys are working on those issues
 [8:45 AM] <bryceh> we have them linked to the blueprint so I plan to keep an eye on them, and raise priority of issues that seem not to be getting neough attention.
 [8:45 AM] <jasoncwarner_> bryceh: ok, thanks. 
 [8:45 AM] <jasoncwarner_> bryceh: anything else? 
 [8:45 AM] <jasoncwarner_> if not, TheMuso you are up :)
 [8:45 AM] <bryceh> jasoncwarner_, yes a couple things
 [8:45 AM] <jasoncwarner_> bryceh: go for it
 [8:46 AM]  → chronossc (~quassel@c906d5e9.virtua.com.br) joined
 [8:46 AM] <bryceh> in linking multimonitor bugs, I also found a lot of config tool issues (or even just regular X problems), that have been open for some time, so even aside from Unity we have plenty of work to be done to improve things.
 [8:47 AM] <bryceh> the other thing, beyond X, did some work on launchpad stakeholder issues last week.  Anyone with issues they'd like flagged for escalation just let me know.
 [8:47 AM] <bryceh> jasoncwarner_, ok done.
 [8:47 AM] <TheMuso> Ok, started to make good progress on pollishing accessibility for precise.
 [8:48 AM] <TheMuso> Indicator work is well under way, mostly in making sure that the most common indicators, and even some of the less common ones, have accessible descriptions, either by filing bugs, or writing patches directly, i.e brasero.
 [8:49 AM]  ⇐ kamstrup (~kamstrup@188.114.151.81) quit: Quit: Ex-Chat
 [8:49 AM] <TheMuso> Still trying to catch Ted to talk about dbusmenu changes for accessible descriptions for network strength etc in the network manager menu. I could start on the work, but there is a coule of ways that the code could be implemented, and I need to know what Ted preferrs. tedg, if you're around, can we please have a chat about this?
 [8:50 AM]  ⇐ ejat (~fenris@ubuntu/member/fenris-) quit: Read error: Connection reset by peer
 [8:50 AM] <TheMuso> The other big bug bear I am tackling atm is getting notifications fully accessible again. I'm mostly there, and am at the point where I may need to change something on the Orca side as well. I hope to have that nailed today and propose a branch for review by EOD.
 [8:51 AM]  → ejat joined  
 [8:52 AM] <jasoncwarner_> thanks TheMuso 
 [8:52 AM] <jasoncwarner_> TheMuso: anything else you wanted to note? 
 [8:52 AM] <TheMuso> Only that I hope we make a decision re GTK version soon, as the latest GTK in the 3.3 series has some a11y updates I would like us to have.
 [8:53 AM] <jasoncwarner_> TheMuso: ah, ok....
 [8:53 AM] <jasoncwarner_> thanks TheMuso 
 [8:53 AM] <TheMuso> np
 [8:54 AM] <jasoncwarner_> robert_ancell and RAOF want to update everyone on your testing efforts this week? RAOF you can go first then hand off to robert_ancell :)
 [8:54 AM]  ⇐ MrChrisDruif and seb128 quit  
 [8:55 AM] <RAOF> I've been looking at the distro-priority unity bugs, picked https://bugs.launchpad.net/unity-distro-priority/+bug/807921, and wrote a test & fix for it.  And a couple of extra tests of associated functionality while I was down there.
 [8:55 AM] <ubot2> Launchpad bug 807921 in unity "Unity tooltips doesn't show Hebrew characters" [Undecided,Confirmed]
 [8:57 AM] <RAOF> That problem was not actually a rendering problem, so neither were the tests, but today I'm going to be writing some rendering tests for the problem.  That'll make it easy to write further automated rendering tests, 'cause there'll be infrastructure there for starting a window, running your test, and pulling data from it.
 [8:58 AM] <robert_ancell> RAOF, what are you using to do the rendering?
 [8:58 AM] <RAOF> Currently the running X server + GL stack.
 [9:00 AM] <bryceh> RAOF, would xserver 1.11 make any difference?  (And what's the merge status on that?)
 [9:00 AM] <RAOF> That's acceptable for a “make check” run on a developer's system, but is obviously insufficient for headless testing.  It's not clear to me that we require headless testing, though.  If we do, it shouldn't be too hard to wrap this up in a standalone dummy server + software rendered GL.
 [9:01 AM]  ⇐ chronossc and kiwinote quit  
 [9:02 AM] <RAOF> bryceh: xserver 1.11 wouldn't make any difference, although some of the patches going into 1.12 around making running the xserver with dummy input and video drivers might.  The merge status is as last week - it builds, but my forward-ported XI 2.1 patches break grab behaviour, so need to be fixed.
 [9:04 AM] <jasoncwarner_> RAOF: anything else or should we give robert_ancell a chance? 
 [9:04 AM] <RAOF> I think we can give Robert a chance now :)
 [9:05 AM] <TheMuso> heh
 [9:06 AM] <robert_ancell> so, I'm currently extending the test X server code used in LightDM to support the messaging required for Compiz.  It's now in lp:xig.  I'm currently in the slow, no end in sight process of filling the gaps and the easy part of adding lots of tests comes after (was the same with lightdm).  I hope to have one window placement test working by the end of the day.  smspillaz is working from the other direction adding uni
 [9:06 AM] <robert_ancell> t tests.
 [9:08 AM]  → seb128 (~seb128@ubuntu/member/seb128) joined
 [9:08 AM] <robert_ancell> (eof)
 [9:08 AM] <bryceh> robert_ancell, can you elaborate briefly on what the goals are with xig?
 [9:09 AM] <robert_ancell> bryceh, to have an X server implementation which we can analyse all the state of the X server synchronously and easily generate clients/events to test an X client (specifically compiz)
 [9:12 AM] <RAOF> robert_ancell: Now that I think of it, have I pointed you at xcb?  Having xcb-generated server-side protocol handling code is something various people in X have talked about for a while; it might make your job easier?
 [9:12 AM] <robert_ancell> RAOF, I've used xcb as a client, but not as a server
 [9:12 AM] <TheMuso> /c/c
 [9:14 AM] <bryceh> robert_ancell, huh, interesting thanks.
 [9:15 AM] <RAOF> IIRC the xcb libs are generated from the protocol description.  If writing server-side protocol handling is becoming a chore, doing the same for the server implementation will probably work :)
 [9:15 AM]  → AfC (foobar@gateway.syd.operationaldynamics.com) joined
 [9:15 AM] <robert_ancell> bryceh, RAOF, there's also some overlap here with xvfb/its replacement.  Not 100% sure which direction I should have tackled this from but the current technique is working
 [9:15 AM] <bryceh> robert_ancell, I take it this is in an effort to improve automated testing of compiz?
 [9:15 AM] <robert_ancell> bryceh, yes
 [9:16 AM] <robert_ancell> bryceh, in particular to test window placement, stacking and client/wm race conditions
 [9:16 AM] <bryceh> robert_ancell, you may be right about some overlap with xvfb, but probably depends ultimately on what type of functionality you're going to be testing
 [9:17 AM] <RAOF> xvfb should be replaced with the full xserver + dummy input & video drivers; there are some patches in 1.12 that make this easier.  I don't think users of xvfb will need to care about this, though.
 [9:17 AM] <robert_ancell> it's the race conditions that will get hard with a black box X server
 [9:17 AM] <RAOF> It also sounds like xig offers significantly more opportunity to test.
 [9:17 AM] <RAOF> Right.
 [9:18 AM] <bryceh> wonder if xig ought to have a testsuite itself?  ;-) ;-)
 [9:18 AM] <RAOF> It should probably have XTS run against it?
 [9:19 AM] <jasoncwarner_> thanks, robert_ancell and RAOF 
 [9:19 AM] <jasoncwarner_> anyone want to comment on anything else? 
 [9:20 AM] <jasoncwarner_> ok, sounds like it! thanks everyone 
 [9:20 AM] <bryceh> thanks!

DesktopTeam/Meeting/2011-11-22 (last edited 2011-11-22 22:51:35 by jasoncwarner)