For minutes of previous meetings, please see FoundationsTeam.
ColinWatson - chair
- Outstanding actions from last meeting
- Chris to upload openoffice.org3 to the ~openoffice-pkgs PPA (carried over)
- Alexander to write mobile broadband MIRs (carried over)
- Any business from activity reports
OpenJDK's compiler targets Java 1.6 bytecode
- What should we do with acpi-support?
- Good news
Outstanding actions from last meeting
- Chris to upload openoffice.org3 to the ~openoffice-pkgs PPA
- Chris said: "Still outstanding unfortunately, however I am in progress on it. I got 2.4.1-9ubuntu1 done and in the archive, and working on the 3.0 debs as we speak. I'm still without power so working off my much slower (than desktop) laptop." Lars offered remote access to a fast desktop if that would help; Chris accepted. (2.4.1-9ubuntu1 fixed two RC bugs, for which thanks.)
- Alexander to write mobile broadband MIRs
- Done (though still to be processed).
Actions from this meeting
- liw to give calc access to desktop machine for OOo builds
- calc to upload 3.0 to PPA by end of week
- doko to extract information about what we lose by targeting Java 1.4 and report back
OpenJDK's compiler targets Java 1.6 bytecode
Matthias raised the issue that OpenJDK's compiler targets Java 1.6 bytecode, which breaks compatibility with the Blackdown VM. We could explicitly build ant and all runtime dependencies with -target 1.4 (which would be quite time-consuming to change), or we could simply have OpenJDK default to 1.4 or 1.5. Matthias is continuing to investigate what we lose by targeting 1.4 bytecode across the board, in discussion with OpenJDK upstream and with Michael Koch.
What should we do with acpi-support?
Bryce asked this question. Colin said his understanding was that it was supposed to be superseded by hal/pm-utils, but that nobody has yet done the hard work of going through all the quirks in acpi-support and making sure they've been added elsewhere; without that, removing acpi-support runs the risk of introducing hardware-specific regressions (and indeed Steve gave the real-world example of his wireless key).
Bryce is currently going through recent Debian changes to acpi-support (a combination of new quirks and larger restructurings, the latter of which he has been avoiding), and was wondering whether to continue this. We agreed that it made sense to use the most current possible version as a base for the merge into pm-utils anyway, so this is not going to be wasted work.
- Upstream regression caused by invalid UTF-8 data in the password store, but no data lost (once the fix lands, passwords will be recovered automatically). This is on track and Alexander is monitoring it.
May not make beta, but an ifupdown plugin is in progress which should cover most standard setups. Anyone with a complex /etc/network/interfaces should try it (append ,ifupdown to the plugins= line in /etc/NetworkManager/nm-system-settings.conf and restart). We discussed some of the options in a little more detail, but further discussion should go to ubuntu-devel.
An OpenSUSE developer is working on this.
- Colin is working on this and should be finished in time for beta (although with a relatively quick-and-dirty fix).
- Bryce needs to retest this. It is not a regression - more of a wish for functionality that ought to be working by now.
- Waiting for user feedback since Bryce doesn't have the corresponding hardware.
- Allegedly recurred; needs some more analysis. (We were running out of time so there was no further discussion in the meeting.)
- We thought we'd nailed all of this down for Hardy. Colin will work on this.
Alexander posted some links to the Firefox EULA outcome:
Steve mentioned that the next release of the venerable Project Athena is being http://web.mit.edu/release/www/athena10/ based on Hardy.
- firefox EULA bug; multiple rounds of prototypes, screenshots, collecting feedback and finally upload fix together with 3.0.2+build6 update to intrepid.
- abrowser rdepends transition: mozillateam added abrowser to explicit depends of all packages currently rdepending on firefox meta package; on top we added a Provides: firefox to the abrowser meta package.
- finalizing security update bits for ffox 3.0.2 and 126.96.36.199 update; doing QA, fixing previously missed patches for dapper backport; support security team in uploading/releasing the bits: next thunderbird update and doing universe security releases.
- network-manager NEW bugs processing (didn't get far due to overload)
- writing MIR for libmbca and mobile-broadband-provider-info (review pending)
- more release discussion with network-manager upstream; getting feedback on outstanding issues with ifupdown plugin; next: defining final release blocker bugs.
- sponsoring queue: no progress due to unfortunate mix of EULA bug and security update business.
- working on new fontforge package
- discussion with the debian maintainers of the fontforge package about build options
- submitted new fontforge package to the debian maintainers for approval and integration into debian (Ubuntu should sync from there)
- discussions with jtv about import queue handling and batch approval of outstanding templates
- bug fix for ttf-arphic-uming to fix misrendering when using embedded bitmaps (LP: #272913)
- intrepid alpha 6 testing for font rendering and localization
Development / Bugs:
- Troubleshoot gradient banding problem on ATI HD3670. Experimented with all the dithering settings. Coded up a patch that seems to minimize the problem pretty well. FDO# 17094.
- failsafeX: Received patch to add a save-logs function. Closed several long-standing bugs, that the new system lets us avoid. Did some additional testing and fixed some other issues that I noticed.
- Hotkeys / ACPI - Investigation into various hotkey issues led to acpi-support. Spent a couple days reviewing and pulling in outstanding patches. But maybe acpi-support should just go?
- Sponsoring work - 268097, commented on others that need different work before they can go in.
- Planning for late -fglrx release. Decided with mvo, superm1, timo, that if we don't have it by Oct 16th, we should arrange to have upgraders automatically switched to -ati; this could be reconsidered if an adequately good -fglrx appears before release.
- Will be off Friday, Sept 26
- Hurricane Ike evacuation 
OpenOffice.org Release Status meeting
- Uploaded openoffice.org 1:2.4.1-9ubuntu1
- Uploaded openoffice.org-l10n 1:2.4.1-9ubuntu1
- Preparing openoffice.org 1:3.0.0~rc2-1ubuntu1
- Package openoffice.org 1:3.0.0~rc2 released yesterday
- openoffice.org 1:3.0.0~rc3 is scheduled for Sep 30
- openoffice.org 1:3.0.0 is scheduled for Oct 7
 I still have no power. I am staying with my sister now in her apartment and working off my laptop which is substantially slower (~ 3-4x slower) than using my desktop, this mess is getting ridiculous, they still have no firm time estimate for when my area will have power again. However, it does sound like they expect to have everything finished in a week or so though.
- On holiday Friday, September 19th.
Call for testing usb-creator (https://lists.ubuntu.com/archives/ubuntu-devel/2008-September/026557.html).
- Work on usb-creator bugs.
- Fixed a bug in the kde frontend to ubiquity and released a new version for the alpha.
- Looked into a bug wherein the kde finished dialog never shows and ubiquity jumps back to partman_commit. Unable to find the cause of this and will resume looking after usb-creator work.
- Sponsored 273749.
- Released bzr-builddeb 2.0.1 with an upload to Debian (thanks Reinhard), and synced it to Debian (thanks Cesare and Steve). This brings the "package" revision spec and a pile of bug fixes. It needs more testing, and may require a couple more fixes. I'll also consider it for backports depending on the outcome of the testing, as it would be good to make it easy for e.g. developers on Hardy to make use of it.
- Mini-sprint with Jono and Daniel in London. Discussed various development issues and helped Daniel film a few more videos.
- Bug triage and a few fixes for Intrepid. Consolekit has a few crashers that seem to be hitting quite a few people. Talked with seb128 and pitti about moving to 0.3 so that we can engage upstream to fix them.
- Bug fixes for system-cleaner. Got things ready to upload fixed version to intrepid, but will do so after waking up and not while asleep.
- Experimenting with installer to see about installing Ubuntu on a USB stick so that I can easily switch operating systems on the same box. Learned that grub hates me. (Virtualisation doesn't help me, since I wanted to try things that require real hardware, like a GPU.)
- Further study of Heapy.
- Tracked down, and uploaded fixes for the missing sound events for GNOME, i.e. no login sound.
- Patched alsa-lib to look for extra configuration files relating to pulseaudio. Also updated the pulseaudio alsa plugin in the alsa-plugins package to the latest code from alsa-plugins git, as it is much better performing code. Also cherry-picked a few fixes from alsa-lib git relating to documentation, and minor internal function fixes.
- Examined, and took some patches from the Mandriva pulseaudio package for some minor fixes that are addressed in 0.9.11 and later. Added configuration files to allow alsa applications to use pulseaudio for audio output only when pulseaudio is running. This means when a user uses GNOME, all alsa apps will use pulseaudio if it is running. This will not affect Xubuntu/Kubuntu users, so they will continue to use alsa exclusively.
- Uploaded a new libcanberra package revision to add some much needed files for GNOME's login/logout sound events.
- Refactored the ubuntu-sounds package to use the new freedesktop sound theme specification format/directory layout, preserving some older file names for use with gdm's sound playback. The package is not much smaller, but the size can be reduced further if space becomes an issue on disks.
- Spent time coming up to speed with some changes made upstrea for GNOME accessibility. A few more minor changes to casper are required to fully address these issues.
- Updated gnome-orca, accerciser, at-spi, atk1.0, and mousetweaks to their GNOME 2.24 releases.
- Merged some improved changes from Debian, who have now included my dmraid work into d-i. Also spent time with the debian dmraid maintainer ironing out a few bugs relating to my dmraid-activate script.
- Debugged why grub-installer doesn't make use of os-prober's output when dealing with dmraid arrays. Still need to code up a fix and test, but this should land prior to beta freeze.
- Found out only today that dmraid 1.0.0.rc15 is out, which has support for Intel Software RAID rebuilding, and RAID metadata creation. I haven't dug much further, but I still don't think there is proper handling of degraded arrays yet. The changelog is also a bit too brief in what has been changed.
- Spent more time wrestling with ardour, and the pile of junk that is scons. I've come to the conclusion that in order to get a buildable package, a lot of the translation files/rc files will have to be built prior to upload, and then copy them into place when the package gets built on the buildds. It is an absolutely awful hack, but whatever works at this point.
- Unfortunately no sponsoring this week. Head space and time have been taken up too much by doing final touches to audio and dmraid.
Off work most of this week.
- Intrepid alpha-6 released
- FFe processing
- disseminating information about e1000e hardware-damaging bug
- preparations for beta
- NBS transitioning: khelpcenter
- fix for bug #270328 in pam, minor bug in pam-auth-update
look at getting pam-auth-update in the System->Administration menu for intrepid, but likely to slip
- main sponsorship: bug #268765, 272383
- Monday archive duties