2010-03-02
For minutes of previous meetings, please see DesktopTeam/Meeting.
Meeting Minutes
Present
Main Meeting
- Rick Spencer (rickspencer3) - chair
- Alberto Milone (tseliot)
Arne Goetje (ArneGoetje)
- Bryce Harrington (bryce)
- Chris Cheney (ccheney)
- Chris Coulson (chrisccoulson)
- Didier Roche (didrocks)
- Jonathan Riddell (Riddell)
- Martin Pitt (pitti)
- Sebastien Bacher (seb128)
- Till Kamppeter (tkamppeter)
Eastern Edition
Luke Yelavich (TheMuso)
- Robert Ancell (robert_ancell)
- Chris Halse Rogers (RAOF)
Apologies
Ken VanDine (kenvandine)
Agenda
- Outstanding actions from last meeting
- Welcome!
- Partner Update
- Kubuntu Update
- mozilla update
- Release Bugs/Release Status
- X cursor at boot time
- Review activity reports
- Any other business
Actions from this meeting
- ACTION: pitti to drive "hiding X cursor" project
Outstanding actions from last meeting
- ACTION: rickspencer3 to arrange call/meeting with ara, pitti, and seb128 regarding upgrade testing
- RESULT: Not done yet
- ACTION: rickspencer3 to bring up stopping new work with statik
- RESULT: Done
- ACTION: rickspencer3 to follow up with robbiew on startup time targets
- RESULT: Done
- ACTION: rickspencer3 to start blueprint to define criteria and associated wiki page to capture list of Featured apps
RESULT: done, https://blueprints.edge.launchpad.net/ubuntu/+spec/desktop-lucid-featured-applications (thanks to robert_ancell)
Welcome!
- Please welcome to new Chris's to the team
- Chris Coulson, new Firefox and NM maintainer
- Chris Rogers, new UNE maintainer, also with a specialty in xorg maintenance
Partner Update
OLS
- Ubuntu One Music store
- libubuntuone and rhythmbox-ubuntuone-music-store have landed on the CD, thanks for all the help
- there will be bug fixes for both of those coming this week based on bugs filed during beta testing the store
- once those bug fixes are in, and tested by the beta group, they hope to turn on the production store. Perhaps end of this week.
- Ubuntu One client
- updates for the control panel are expected this week
DX
Kubuntu Update
- KDE 4.4.1 landed today
mozilla update
- libsoup2.4-gnome, libproxy, webkit are done
- working on epiphany
- expecting everything to be building this week, and functionality to be debugged next week
Release Status
http://people.canonical.com/~pitti/workitems/canonical-desktop-team-ubuntu-10.04-beta-1.html
- Most remaining work items of related to landing Dx deliverables
- U1 status
- Postponed some work items regarding control panel
- Should have UI for user defined folders any day now
- SFTS/Gwibber
- Making surface area for users more intuitive, but developers still should know it's "gwibber"
- A bit more integration work going on
- Still seeing some crashes
X Cursor at Boot Time
- sabdfl would like us to hide this
- who can take up this work item?
- long discussion ensued regarding options
- pitti to lead investigation and determine correct implementation strategy
Other Business
Activity reports
Alberto Milone (tseliot)
- Working on 16 colours support for Plymouth (LP: #526892)
- Fix packaging scripts for the next fglrx release
- Fix problem about empty modaliases in fglrx
- Add support for additional separate places for messages in plymouth (LP: #497311)
- Complete 1st release of the plymouth theme
Arne Goetje (ArneGoetje)
- Blueprint work:
- desktop-lucid-language-selector: all done, beta available, still need to fix some minor annoyances in the UI
- desktop-lucid-font-selection: done, preparing ubuntu-desktop-fonts package in order to replace some seeded packages on the Live CD.
- sick on Friday
- filed MIRs for poppler-data and pinyin-database
- generated new language-pack updates for Lucid
Bryce Harrington (bryce)
- Sponsoring
- Sponsor Sarvatt's patch to quell error message in -evdev
- Sponsor apport fixes to xorg by RAOF
- Bug triaging
- Run process-incomplete-bugs.py against -intel; plan again next week
- Send weekly bug list to Intel
- Triage all xorg bugs tagged 'lucid'
- Symptom tag all -intel bugs tagged 'lucid'
- Run process-incomplete-bugs.py against -intel
- Apport hook work
- Set up apport/udev for X freeze capture
- Improve apport hook for X freezes
- Add info for freeze dump apport hook
- Ensure xorg apport hook adds 'lucid', etc.
- Add some prompts for X apport hook
- failsafeX / xdiagnose development
- Ensure failsafeX works well with KMS
- Start new project xdiagnose to house failsafex scripts
- Update xdiagnose blueprint
- Arsenal scripting improvements
- Bugs with Patches report
- Patch page is showing bug age with []
- In patches report, exclude bugs tagged 'patch-needswork'
- Exclude Fix Committed bugs in patch report
- Improve xorg package retargeter
- Detect bugs mentioning wacom or tablet in description or title and file them against wacom driver
- Analyze why retargeter misfiled 479948 to -ati
- Improve xorg retargeter to be better at figuring out the right package to move bugs to
- Make retargeted email indicate Incomplete w/ Response, and omit showing Incomplete w/out response
- Improve process-incomplete-bugs.py
- Make it better able to handle drive-by commenters
- Make it ignore non-Ubuntu bug reports
- Switch how status is updated in process-incomplete-bugs.py
- Improve process-tagging.py
- Fix unicode bugs in release name tagging script
- Make script to tag bug reports 'lucid' if they mention lucid in title or description
- Detect kubuntu or xubuntu in description and set a tag accordingly
- Modify bugs-symptoms.py to tag symptoms
- Make arsenal stop excessively renaming -intel bugs with chipset
- Add a hook and cron job to run report-intel-upstream
- Debug why high-karma-bugs report is not generating
Update xorg_pkg_list to pull xkeyboard-config versions from http://xlibs.freedesktop.org/xkbdesc/
- Recode milestones bug report in python
- Bugs with Patches report
- Nouveau
- Test out nouveau on my nvidia hardware
- Document procedure for installing and reverting nouveau
- Sponsor themuso's patch to disable nouveau on ppc
- Close out all -nv bug reports filed before transition as obsolete due to move
- Review/follow-up on RAOF's remaining nouveau work email
- Fix nouveau to allow Depends on -pae kernel (LP: #524792)
Voodoo generator - http://lists.freedesktop.org/archives/nouveau/2010-February/005137.html
- Test for X issues on Dell Mini 10v
- File bug on Dell Inspiron 1011: lshw on console makes screen blue+green
- Decide on and procure new laptop to for my 3-yr anniversary
- Brainstorm some SOC-like student projects for Ubuntu-X
- Review proprietary driver test results
Chris Cheney (ccheney)
- Backported libsoup2.4-gnome, libproxy, webkit
- Working on backporting epiphany-browser
- Met with security team about questions on support for webkit/epiphany-browser
Didier Roche (didrocks)
- WI B1 related (7 news):
netbook-launcher-> see how to get caching when n-l draw the bg and g-s-d isn't triggered
-> DONE, new release of netbook-launcher containing cache
fix live system -> installed system background cache copying for installer-only mode: WAITING FOR REWIEW
-> casper old hook removed -> bin taken from CANCELED
- sync gthumb to debian unstable to get last release working with automounting: DONE
-> gthumb is needed as we removed f-spot and we need one app for importing photo from camera (see bug #530437)
- Not WI related:
- UNE specifc:
- Other:
- GNOME updates: gimp, seahorse-plugins, yelp
- pushed lernid in universe
Jonathan Riddell (Riddell)
Ken VanDine (kenvandine)
- Ubuntu One Music store testing
- Empathy update and bug fix
- gwibber fixes
- weekly DX uploads
- fixes for indicator-me to handle latest ubuntuone-client and empathy changes
Luke Yelavich (TheMuso)
Accessibility
- Uploaded a new revision of gnome-orca to remove the hard dependency of libgail-gnome-module. This package is still needed for packages like evolution that use libbonoboui/libgnomeui, but it has now been moved into the seeds, where it will be easy to kill off in the GNOME 3 world.
- Continue to debug accessibility not working in only-ubiquity mode on the live CD. I am using KVm for my testing. I am now wondering whether I should use real hardware to test, to see is behavior changes. Still want to get this fixed for lucid final.
Audio
- Audio bug triaging, bugs in question were against the alsa userspace apckages, pulseaudio, and the kernel.
- More audio package bug trawling, this time to try and get an update on old bugs. If nobody replies in a couple of weeks or so, I'll close them, to try and get bug counts down somewhat.
- Involved with discussion about users wanting to urn off the login sound. This is something that gdm config utilities need to address.
- Help the mobile team troubleshoot audio issues with pulse/gstreamer etc on arm hardware. This was in the form of requesting tests via a bug report.
Misc
- Alpha 3 CD testing.
Martin Pitt (pitti)
Only a 4-day week (swap day on Friday)
Assigned blueprints: All implemented
Done:
- Bug fixing: rhythmbox-ubuntuone-music-store, udev, udisks, yelp
- Version updates: udisks, gvfs, gnome-disk-utility
- Boot speed improvements:
- udisks: add profiling, defer ATA SMART handling
- libnotify, nm-applet: initialize notify daemon lazily
- some conf calls
- work items tracker: fix burndown charts for milestones
- Tracked down the reason for the loss of kernel ddebs; applied a workaround, and asked kernel team to fix package names
- TB meeting
- Worked out release team status with Sebastien
Sponsoring:
- lsb, rhythmbox
Robert Ancell (robert_ancell)
Sebastien Bacher (seb128)
- desktop updates: gnome-media libgweather gnome-panel totem poppler cairo gtk rhythmbox gtkhtml evolution-data-server evolution evolution-exchange
- sponsored several csd updates from Cody
- backported pidgin upstream change to fix icq login
- new rhythmbox snapshot
- new libgnome-keyring snapshot to fix new keyrings registration issue
- review and sponsored gnome-bluetooth port to application indicator
- gstreamer updates synced from debian
- tracked codec installation issue which was again bluez breaking things
- tested alpha3 iso, ran into several ubiquity issues, spent some time filing bugs and trying to debug a crasher
- tracked ubiquity crasher to webkit and then enchant, got that one fixed by sponsoring an enchant change
- f-spot bugfixing sprint (backport git changes and sponsored different patch to fix hundredpapercut issues)
- some conf calls, prepared weekly r-t meeting with pitti
- sponsoring: rhythmbox-ubuntuone-music-store gnome-doc-utils indicator-application gnome-settings-daemon libdbusmenu libindicator indicator-application rhythmbox-ubuntuone-music-store ibus gimp gnome-control-center gnome-settings-daemon enchant gnome-settings-daemon libindicate gnome-control-center
Till Kamppeter (tkamppeter)
- HPLIP: Updated to version 3.10.2 final.
- lsb: Made the "lsb" package not installing an MTA any more (extra daemon, configuration questions) by adding a dummy binary package ("lsb-invalid-mta") with a "sendmail" command which simply errors out (LP: #141641)1. Let the "lsb" package require "lsb-printing" as the printing stuff is core part of the LSB (LP: #386128).
system-config-printer: Applied fixes so that automatic download of drivers and PPDs from OpenPrinting works correctly.
- ghostscript: Applied upstream patch to fix a bug which prevented Ghostscript from printing the Ubuntu test page (Upstream bugs #691137, #690909).
- Answered and triaged printing-related bug reports.
Chris Coulson (chrisccoulson)
- First day yesterday
- GNOME updates: gnome-settings-daemon, gnome-power-manager, gnome-control-center, gnome-user-share, libwnck
- Fixed bug in gnome-about-me which prevented saving any fields in to the addressbook
- Fixed a crash at startup in gnome-power-manager late last night, which seemed to affect everybody except for me
IRC Logs
[17:04] <pitti> https://wiki.ubuntu.com/DesktopTeam/Meeting/2010-03-02<<BR>> [17:05] <rickspencer3> here are the the actions from last week are done
[17:05] <rickspencer3> except:
[17:05] <rickspencer3> ACTION: rickspencer3 to arrange call/meeting with ara, pitti, and seb128 regarding upgrade testing
[17:05] <rickspencer3> RESULT: Not done yet
[17:05] <rickspencer3> I'll get to this this week I hope
[17:06] <rickspencer3> thanks to robert_ancell for doing a really really wonderful job with driving the Featured Apps spec
[17:06] <rickspencer3> let's finalize that in next week's meeting
[17:06] <rickspencer3> so
[17:06] <rickspencer3> next item
[17:06] <rickspencer3> welcomes
[17:06] <rickspencer3> we have quite the explosion of Chris's on the desktop team
[17:06] <rickspencer3> Chris Coulson, new Firefox and NM maintainer
[17:07] * chrisccoulson waves
[17:07] <rickspencer3> hi chrisccoulson, based in England, correct?
[17:07] <chrisccoulson> that's correct
[17:07] <rickspencer3> I'm sure everyone in the Ubuntu community knows chrisccoulson for his many years of contributions
[17:08] <rickspencer3> our other new Chris is Chris Rogers, new UNE maintainer, also with a specialty in xorg maintenance
[17:08] <rickspencer3> based in Sydney
[17:08] <rickspencer3> so he's asleep atm
[17:08] <rickspencer3> and goes by RAOF
[17:08] * pitti makes a mental note to ask for deciphering his nick tomorrow morning
[17:08] <rickspencer3> thanks pitti
[17:09] <rickspencer3> perhaps that doesn't require an ACTION :)
[17:09] <pitti> lol
[17:09] <bryceh> and btw found out his official last name is "Halse Rogers" not just Rogers
[17:09] <rickspencer3> ooops
[17:09] <pitti> oh
[17:09] <rickspencer3> my bad
[17:09] <rickspencer3> thanks bryceh
[17:09] * pitti updates his muttaliases
[17:09] <tseliot> :-)
[17:09] <rickspencer3> now I have to wonder who this Chris Rogers guy we hired is then?
[17:09] <bryceh> rickspencer3, heh
[17:09] <rickspencer3> ok
[17:09] <rickspencer3> so kenvandine is not available atm
[17:09] <rickspencer3> so no partner update today
[17:10] <rickspencer3> I will mention U1 briefly in the release status section though
[17:10] <rickspencer3> Riddell, kubuntu update?
[17:11] <rickspencer3> ok
[17:11] <rickspencer3> I forgot to remind Riddell, so let's move on
[17:11] <rickspencer3> ccheney, mozilla update?
[17:11] <pitti> apparnetly KDE 4.4.1 is just being landed
[17:11] <rickspencer3> urk
[17:12] <rickspencer3> ok
[17:12] <ccheney> rickspencer3: i got the libsoup2.4-gnome, libproxy, webkit bits done working on epiphany itself now
[17:12] <rickspencer3> ccheney, what's the status of the backporting?
[17:12] <rickspencer3> good
[17:12] <rickspencer3> ccheney, does epiphany seem as straight forward as asac was hoping?
[17:12] <ccheney> not too bad but certainly not straight forward
[17:13] <rickspencer3> mm
[17:13] <rickspencer3> ok
[17:13] <rickspencer3> ccheney, will you be done by eow?
[17:13] <ccheney> it requires forward porting code from hardy version to the new version since the new version uses gtk bits that moved out of it in new epiphany/gtk
[17:13] <ccheney> i think i will be done by eow yea
[17:13] <rickspencer3> great!
[17:13] <ccheney> i don't know if the resulting stuff will run, so there will likely be debugging after i finish the code
[17:13] <rickspencer3> I will sleep better when that is done :)
[17:13] <rickspencer3> oh
[17:13] <rickspencer3> well, yeah
[17:14] <rickspencer3> we should consider that as part of the project as well ;)
[17:14] <ccheney> but i think i can have the stuff at least preliminary finished by eow
[17:14] <rickspencer3> thanks ccheney
[17:14] <rickspencer3> next is release status
[17:14] <rickspencer3> I have a few items, but as usual, I would like to hand the mic to pitti to cover bugs and work items
[17:14] * rickspencer3 hands mic to pitti
[17:14] <Riddell> yes we're in the middle of releasing 4.4.1
[17:14] <rickspencer3> thanks Riddell
[17:15] <rickspencer3> feel free to finish with the Kubuntu team and check back later!
[17:15] <rickspencer3> first things first, after all
[17:15] <pitti> I don't have a lot this week
[17:15] <rickspencer3> music to my ears
[17:15] <Riddell> prio
[17:15] <pitti> my most important message is to congratulate all of you for alpha-3
[17:15] <pitti> http://people.canonical.com/~pitti/workitems/canonical-desktop-team-lucid-alpha-3.html looks exemplary
[17:15] <pitti> at the beginning we said this was a very optimistic list of WIs
[17:15] <pitti> and we didn't drop too much
[17:16] <pitti> for now, I'd just like everyone (the new CHrises in particular) to be aware of http://people.canonical.com/~pitti/workitems/canonical-desktop-team-ubuntu-10.04-beta-1.html<<BR>> [17:16] <pitti> so kenvandine has like 70% of the WIs, and everyone else can twiddle thumbsWWfix bugs
[17:17] <rickspencer3> fix bugs!
[17:17] <kenvandine> hey... most of those are in the dx integration one
[17:17] <kenvandine> which dbarth was going to clean up
[17:17] <pitti> kenvandine: how much of SFTS is bonus at this point? I guess the "launch by defualt" bits aren't
[17:17] <rickspencer3> kenvandine, we know
[17:17] <rickspencer3> :)
[17:17] <seb128> I did clean quite some dx integration ones
[17:17] <bryceh> I like that desktop-bugs is helping us do WI's now
[17:18] <pitti> bryceh: oops -- artifact from using linked bugs..
[17:18] <rickspencer3> ok
[17:18] <pitti> nothing else from me this week
[17:18] <rickspencer3> so to answer your question pitti
[17:19] <kenvandine> seb128, he needs to assign the individual WIs to his team members
[17:19] <rickspencer3> I don't know the specific work items, but sabdfl was quite excited to finish out sfts with this extra level of integration effort
[17:19] <rickspencer3> so I think they are not-optional
[17:19] <seb128> kenvandine, ok
[17:19] <kenvandine> rickspencer3, there are like 2 of those i think
[17:19] <rickspencer3> kewl
[17:19] <rickspencer3> ok
[17:19] <pitti> but there's also some "add exceptiion handling to foo" bits
[17:20] <rickspencer3> right
[17:20] <pitti> which I'm not sure about; they sound like robustness fixes
[17:20] <kenvandine> pitti, that is currently blocked
[17:20] <rickspencer3> I am betting that those come back to us as bugs that need to be fixed ;)
[17:20] <kenvandine> i am waiting for ryan to land some changes...
[17:20] <rickspencer3> changes?
[17:20] <rickspencer3> *sigh*
[17:21] <kenvandine> yeah... i warned you there would be a FFE
[17:21] <kenvandine> :)
[17:21] <rickspencer3> yeah yeah
[17:21] <rickspencer3> but c'mon everyone
[17:21] <pitti> kenvandine: for a new gwibber version?
[17:21] <rickspencer3> we need to lock down so we can stop churn and fix bugs
[17:21] <kenvandine> right now there is no way for us to capture individual failures
[17:21] <kenvandine> like an operation that fails for just one service
[17:22] <kenvandine> ryan promised it like over a week ago...
[17:22] <rickspencer3> POSTPONED
[17:22] <rickspencer3> let's move on
[17:22] <kenvandine> it shouldn't be high risk... but it blocks the last thing i want to do for exception handling
[17:22] <kenvandine> ok
[17:22] <rickspencer3> so also wrt to release status
[17:23] <rickspencer3> the U1 team is locking down bits for Lucid
[17:23] <rickspencer3> they have postponed some work items regarding control panel
[17:23] <rickspencer3> Should have UI for user defined folders any day now
[17:24] <rickspencer3> we already touched on Gwibber
[17:24] <rickspencer3> note that we will not call it "Gwibber" in teh UI
[17:25] <rickspencer3> I forget what it will be called in the UI, I presume it is specified in the blueprint
[17:25] <rickspencer3> anything else wrt release status, or shall we move on?
[17:26] <rickspencer3> ok
[17:26] <rickspencer3> x cursor at boot time
[17:26] <rickspencer3> so currently, plymouth loads to show some boot progress
[17:26] <rickspencer3> then x loads
[17:26] <rickspencer3> at which point it displays the spinng cursor
[17:27] <kenvandine> note: it will be called gwibber, but the window title bars will be more generic like "Broadcast messages"
[17:27] <rickspencer3> sabdfl asks that we hide this cursor
[17:27] <rickspencer3> bryceh, I assume this is not trivial, but also not terribly complicated
[17:27] <rickspencer3> thoughts?
[17:27] <seb128> seems an xorg task, could be for RAOF maybe if he doesn't have too much to do already?
[17:27] <pitti> it could also be done by gdm, I suppose?
[17:28] <rickspencer3> seb128, that's what I was thinking, but I would like kenvandine and tseliot to provide guidance and oversight
[17:28] <rickspencer3> pitti, hmmm, good point
[17:28] <pitti> setting the mouse coursor to an invisible one, and changing to the normal one once it's done loading?
[17:28] <seb128> pitti, gdm doesn't fix the gdm to desktop stage though
[17:28] * pitti doesn't know whether that's possible in gdm, though
[17:28] <kenvandine> i know bratsche looked into this in karmic
[17:28] <bryceh> rickspencer3, you can set the mouse cursor to an invisible cursor pretty easily
[17:28] <tseliot> rickspencer3: we could move the cursor below the logo perhaps?
[17:28] <seb128> dx came with a gdm workaround in karmic
[17:29] <rickspencer3> sounds like there are some options
[17:29] <kenvandine> yeah, i bet bratsche can give us advice
[17:29] <didrocks> seb128: we still need the cursor once gdm is loaded to select user, session, and so on…
[17:29] <pitti> tseliot: that sounds complex; it's a sprite on top of everything usually, isn't it? and hardware rendered?
[17:29] <seb128> didrocks, that's busy cursor
[17:29] <seb128> the issue there was, what happen when plymouth is not used
[17:29] <seb128> or xsplash
[17:29] <seb128> gdm should be spinning cursor in those cases
[17:29] <pitti> should it?
[17:30] <pitti> it doesn't seem plymouth related at all to me
[17:30] <seb128> otherwise you get an empty screen and no activity
[17:30] <seb128> well, when you have an animation you know things happen
[17:30] <seb128> when you get an empty stalled screen not so much
[17:30] <tseliot> maybe it's something that the plymouth upstart job can do?
[17:30] <tseliot> (hide the cursor)
[17:31] <seb128> also what happen for people using an another login manager?
[17:31] <seb128> ie kubuntu
[17:31] <tseliot> so that it's done only if you pass splash on boot
[17:31] <didrocks> I agree with seb128, when gdm take some time and you don't have plymouth, it's good to have a feedback that the system isn't frozen
[17:31] <tseliot> good point
[17:31] <seb128> I would prefer to have that change in right place that a gdm specific workaround
[17:31] <rickspencer3> ok, who can be on point to make this happen, even if RAOF ends up with the implementation, I don't feel it would be fair to ask him to design a solution
[17:31] <rickspencer3> ?
[17:31] * rickspencer3 hears scampering noises
[17:31] <rickspencer3> :)
[17:32] <pitti> seb128: hm, I had actually thought that it would be safer to do in gdm
[17:32] <pitti> so that we don't break kdm, XFCE, and whatnot
[17:32] <seb128> well I would prefer having plymouth going over mouse
[17:32] <seb128> if that's possible
[17:32] <seb128> but I don't know if that is
[17:32] <chrisccoulson> pitti - isn't it only the greeter which has a connection to the display?
[17:32] <chrisccoulson> (in which case, it would be too late wouldn't it?)
[17:32] <pitti> seb128: that sounds hard to do (since it's rendered by hardware), but ICBW
[17:32] <seb128> or mouse turned off during plymouth run
[17:33] <rickspencer3> hehe
[17:33] <pitti> chrisccoulson: the slave starts X
[17:33] <seb128> like the plymouth script could turn curser off and on
[17:33] <seb128> well, I said what I had to say
[17:33] <pitti> chrisccoulson: it could pass options to X, or run programs in it (which it already does with the greeter)
[17:33] <tseliot> seb128: when you see the cursor it means that plymouth has been already deactivated and left the bootsplash in the framebuffer
[17:33] <seb128> I will let people who know figure ;-)
[17:33] <chrisccoulson> pitti - oh, ok, i didn't realise that
[17:33] <pitti> well, it's one of these little things which are again regression prone and nontrivial to implement
[17:33] <pitti> (sounds like 20 new indicators *cough*)
[17:34] <tseliot> heh
[17:34] <rickspencer3> hmm
[17:34] <seb128> I know we had the "hide cursor" discussion during dapper ui sprint
[17:34] <seb128> ie a while ago ;-)
[17:34] <seb128> and it's coming back regularly since and nobody came with a good way to do it
[17:34] <pitti> rickspencer3: I guess we also don't really have an option to say "no", right?
[17:34] <seb128> so it's probably not trivial
[17:34] <rickspencer3> pitti, well
[17:34] <rickspencer3> I can't say "no"
[17:35] <rickspencer3> but I can say "it's hard, costly, and buggy"
[17:35] <rickspencer3> and based on this conversation I will do that
[17:35] <rickspencer3> however, we still need to try
[17:35] <bratsche> As I recall seb128 is right that the cursor should be disabled by Xorg to begin with.. if you disable it in gdm then you're still probably going to see the cursor briefly before gdm starts.
[17:35] <pitti> I think I could give me a hard limit of one hour to ask people and play around with that
[17:35] <seb128> if nobody comes with a good idea I guess I would go with the gdm hack
[17:35] <rickspencer3> pitti, would you be willing to coordinate a plan of attack?
[17:36] <pitti> bratsche: ok, thanks; that helps already
[17:36] <rickspencer3> I'll summarize a list of options from this discussion to discuss with sabdfl
[17:36] <pitti> but if we change X to hide the cursor by default, we'd break all non-gdm setups
[17:36] <tseliot> how about making the cursor invisible?
[17:36] <tseliot> and making it visible after we're done with gdm, etc.?
[17:36] <pitti> tseliot: invisible, hide, still the same problem -- something has to turn it back on later
[17:36] <bratsche> pitti: I'm not sure.. don't most other display managers set the cursor at some point?
[17:37] <pitti> i. e. fix gdm, kdm, xdm, ldm, raw X, custom setups without a WM, etc.
[17:37] <pitti> bratsche: xdm certainly doesn't :)
[17:37] <bratsche> Oh, hmm.
[17:37] <pitti> and I get bug reports where people use startx
[17:37] <pitti> which is fine for specialized setups, embedded stuff, and whatnot
[17:38] * mclasen recommends asking upstream...
[17:38] <tseliot> pitti: we can make the cursor visible again when stopping the plymouth upstart job
[17:38] <pitti> hence my original point to limit it to gdm itself
[17:38] <Keybuk> isn't there an X patch around somewhere that adds a command-line option to hide the cursor by default
[17:38] <Keybuk> in which case, you could add that option from the existing gdm plymouth transition patch
[17:38] <Keybuk> (likewise kdm)
[17:38] <pitti> so that gdm says "start X with cursor hidden" and enables it later on
[17:38] <tseliot> Keybuk: yep
[17:38] <pitti> so that we can avoid changing the default X behaviour
[17:39] <Keybuk> that'd mean that the cursor is *only* hidden in the cases where we've already got a patched window manager
[17:39] <Keybuk> and in the case where plymouth was running
[17:39] <pitti> Keybuk: exactly
[17:39] <Keybuk> that way you then just need a second half of the patch to turn the cursor back on ;-)
[17:39] <seb128> can we do what mclasen recommends too
[17:39] <seb128> and ask upstream if they have an idea on how to do that too
[17:39] <Keybuk> sure
[17:40] <pitti> ok, sounds good
[17:40] <rickspencer3> all
[17:40] <rickspencer3> I will be happy for this conv. to carry on
[17:40] <rickspencer3> but let's window down the meeting first
[17:40] <rickspencer3> then I can let you get back to the discussion
[17:40] <pitti> I think we covered all we can in the meeting
[17:41] <rickspencer3> any other business?\
[17:41] <seb128> yes
[17:41] <Keybuk> Fedora seem to have a patch for this
[17:41] <seb128> yes was a reply to pitti's comment
[17:41] <pitti> rickspencer3: I'll take the action to coordinate this, file bug, investigate options, etc.
[17:41] <seb128> no other business
[17:41] <rickspencer3> pitti, thank you!
[17:41] <rickspencer3> ok
[17:41] <rickspencer3> that's a wrap
[17:41] <rickspencer3> thanks all
[17:41] <seb128> thanks
CategoryDesktopTeam
Back to DesktopTeam.
DesktopTeam/Meeting/2010-03-02 (last edited 2010-03-02 21:20:53 by 224)