[02:01] <persia> ian_brasil: What's the best way to check if accessibility works?  I'd like to have accessible input working for the no-keyboard use case.
[02:04] <ian_brasil> persia: well i install python-dogtail to check..not sure if this is the best way 
[02:06] <ian_brasil> this needs the accessibility layer working
[02:06] <ian_brasil> if everythin is working well then 
[02:06] <ian_brasil> >>> import pyatspi
[02:06] <ian_brasil> >>> reg = pyatspi.Registry
[02:06] <ian_brasil> >>> print reg.getDesktopCount()
[02:06] <ian_brasil> 1
[02:06] <ian_brasil> >>> desktop = reg.getDesktop(0)
[02:06] <ian_brasil> >>> print len(desktop)
[02:06] <ian_brasil> 0
[02:07] <ian_brasil> this last value will not be 0 and the apps are registering themselves to the acc layer
[09:00] <rhp> so, having installed ubuntu intrepid (and windows xp again), I just tried to create an updated image. However it seems I'm out of luck these days, as it could not get to completion.
[09:00] <rhp> I got a crash in the mksquashfs process.
[09:07] <rhp> Any ideas whether 8.04 will also give me a system where I can update images? intrepid seems a bit over the edge for me...
[09:21] <persia> ian_brasil: I see, so it's a registration thing.  Hrm.  That sounds fairly complicated, as the only actually running process is supposed to be hildon-desktop: the other apps aren't that independent
[09:21] <persia> (and sorry for the delay: my network was having a bad day)
[09:22] <persia> rhp: How are you creating an updated image?
[09:23] <rhp> using a script I found on ogra-maemo's blog.
[09:23] <rhp> Or somwhere else... I do not remember.
[09:23] <rhp> Reading through it, it seemed logical enough...
[09:25] <persia> rhp: OK.  I'm not sure why it would crash.  There are new daily images available each day, and rsync seems to be smart about the bandwidth of updates.  This is the method I tend to use.
[09:28] <rhp> rsync?
[09:28] <rhp> You don't just have the update manager update your system?
[09:30] <persia> I reinstall my system 5-6 times a day, but my use is certainly not normal.
[09:30] <persia> You got a crash with update manager?  Running a live session, or post-install?
[09:31] <rhp> I got a couple of crashes actually. One running update manager -- telinit crashed.
[09:32] <rhp> One while updating het mid image -- in mksquashfs... but the crashreport message was quite busy...
[09:32] <rhp> Both post-install
[09:33] <rhp> You do not happen to know whether 8.04 also works for updating the image?
[13:48] <lool> Cool, -mid doesn't prompt on boot anymore
[13:53] <persia> lool: But grub doesn't get installed :/
[13:53] <persia> It's stuck live for now.  Trivial to fix Friday.
[14:12] <persia> So, I've been looking at the forums to try to figure out where to point people for Ubuntu Mobile.
[14:13] <persia> It was suggested that "Hardware & Laptops" might be the right place.  Apparently, it's possible to get some "prefix" for Ubuntu Mobile stuff (I still need to track down what that means, and how).
[14:13] <persia> Anyone have any thoughts on that, or ideas?
[14:13] <lool> persia: Hey could you tell me more about the grub issue?
[14:14] <lool> StevenK: Could you drop the union=unionfs thingy some isolinux.cfg?
[14:14] <lool> StevenK: or syslinux.cfg
[14:14] <lool> StevenK: Only for -mid though
[14:14] <lool> amitk: Thakns for the aufs update in linux-lpia; works like a charm; when it the one in -generic landing?
[14:14] <lool> amitk: This would fix the issues with the -mobile images
[14:15] <persia> lool: ubiquity now supports grub for installs, but /usr/lib/ubiquity/ubiquity/components/grubinstaller.py isn't installed on lpia.
[14:15] <lool> persia: Is it too late to fix this?
[14:15] <persia> Basically, I fixed the code, and tested the code patch, but failed to fix and test the packaging.
[14:16] <persia> Basically.  I have a local fix, but as I don't yet have the icon for it, and it means *everyone* else needs to rebuild their disks and start testing from scratch again, I think it's not enough to push.
[14:16] <persia> I've not actually spoken with the testing and release teams, aside from a quick, retracted, query to the installer team.
[14:17] <persia> While I'm at it, I'm adding support for ntfs partition mounting, which is probably interesting for those with Windows installed on the devices.
[14:17] <lool> persia: Can be get the fix in ubiquity now, and hope they have another reason for an upload before beta?
[14:17] <amitk> lool: talking to kernel/distro folks to see if aufs can be updated in a kernel upload
[14:18] <lool> amitk: What's your current list again?  lrm-lpia and then?
[14:18] <persia> lool: Likely.  The plan is to push a branch for review by the installer team within the next few hours.  Whether there is an upload or not isn't something I really want to push.
[14:18] <lool> amitk: ath5k doesn't seem too far from working   :-/
[14:19] <persia> amitk: Any ideas for bug #274704?
[14:20] <amitk> lool: after aufs to base kernel, there is lrm (almost done)
[14:22] <lool> Ok
[14:22] <ian_brasil> persia: Nokia recently opened the repo to the Hildon Test Automation project http://hildon-test-aut.garage.maemo.org/...there are scratchbox i386 debs in the repo so it should be possible to get it working on MID
[14:23] <persia> ian_brasil: Nifty.  Is it only i386, or also armel?  It may be that my Zaurus is a better testbed for that.
[14:24] <ian_brasil> there are armel debs too https://garage.maemo.org/svn/hildon-test-aut/trunk/ 
[14:25] <persia> ian_brasil: I'm currently a bit caught up with the fact that I need to recover from some mistakes from last week, but thanks for the pointers.  I'm definitely going to try to have a play with that.
[14:27] <ian_brasil> np..i had it working on scratchbox a while back ..the i386 debs on lpia complained about hildon-fm or hildon-1 if i remember correctly
[16:00] <ian_brasil> is the intention to have an ubuntu-mobile iso on http://iso.qa.ubuntu.com/ or does that not make much sense
[16:19] <persia> ian_brasil: I've just been discussing having it there with the testing team.  We're a little late for beta, but if we prepare some test cases in the next couple days, we can be there for RC.
[16:20] <persia> ian_brasil: Are there any specific test cases you think should be added, or do you think we're good with just the basic install testing as an initial set, and we can add more once we have some specified features, etc.?
[16:22] <ian_brasil> persia: that is great news. i have been looking at the Basic, Advanced and Compliance test and maybe we should agree upon some targtes for the compliance tests?
[16:22] <ian_brasil> i ran LSB last night but it seems very based towards rpm
[16:22]  * ogra would appreciate QA testing
[16:23] <persia> That's a *lot* of tests, and I'm not sure we can hit them all.
[16:23] <ogra> but i guess its hard to find devices
[16:23] <ogra> (in the QA team that is)
[16:24] <ian_brasil> the current stats are:
[16:24] <persia> QA team is highly fluid.  If there are users, there will be QA team members.
[16:24] <ian_brasil> With platform in Hibernate state, time from user pressing power on button to desktop UI being fully responsive <= 16 sec 
[16:24] <ian_brasil> Time from user pressing sleep button to platform going to S3 state (not just display going blank) <= 7 sec 
[16:24] <ian_brasil> With platform in S3 state, time from user pressing wake-up button to desktop UI being fully responsive <= 7 sec 
[16:24] <ian_brasil> When the platform is Cold booted (i.e. when device probe, etc is required), the time from user pressing power on button to desktop UI being fully responsive <= 37 sec
[16:24] <ogra> 7sec doesnt look so bad
[16:24] <persia> What HW?  Which image?
[16:24] <ogra> i would rarely expect a real boot
[16:25] <persia> Yeah.  Real boots are for people with HW that doesn't support hibernate.
[16:25] <ogra> well, even hibernate is for whimps :P
[16:25] <persia> And hibernate is for people with HW that uses too much power in suspend.
[16:25] <ogra> just have a big enough battery in your device 
[16:26] <persia> Well, I have one laptop with a 40 minute runtime with the stock battery (no, I've never used the stock battery, and don't intend to do so).  For this, I think hibernate beats suspend.
[16:26] <ogra> i had pondered initially to disable hibernate in ubuntu-mobile 
[16:26]  * ogra tests the latest mobile image ... hopefully we have mouse and kbd back
[16:27] <persia> It also depends on the device.  For one of my laptops, I'm not supposed to connect/disconnect the port replicator without another pass through the BIOS.  I haven't quite figured out why, and the manufacturer carefully doesn't support Linux on this model (they do for other models, but haven't done the internal testing for this one)
[16:27] <persia> ogra: You regenerated the image?
[16:28] <ogra> after the meta change, indeed
[16:28] <ogra> 30.1 should be better 
[16:28] <persia> ian_brasil: For initial tests, I had planned Install (auto-resize), Install (entire-disk), Install (manual partitioning), and Live Session.  Do you think we need a deeper set at this point?
[16:29] <persia> ogra: Refreshing.  No idea if I'll get to a test cycle tonight though: I need to sort the default password.
[16:30] <ogra> YAY
[16:30] <ogra> mouse, touchscreen kbd and joystick all work
[16:30] <ian_brasil> persia: No, that sounds Ok to me
[16:31] <persia> ian_brasil: I'm all for more testing, but right now we seem to have, at most, maybe 10 people actively involved, which makes me not want to add too much extra burden to everyone.
[16:32] <persia> Mind you, if I can ever figure out how the forums work, and the people from the mailing list go to the forums, that might change, but I've not yet found the right set of information.
[16:33] <ian_brasil> persia: right...i like the idea of the tragtes to try to aim for like fottprint 500MB , memory 192MB etc
[16:33] <ian_brasil> s/tragtes/targets
[16:33] <ian_brasil> s/fottprint/footprint
[16:34]  * ian_brasil hates his new keyboard
[16:34] <persia> Footprint 500MB!  I don't think we're anywhere near there, nor has that been one of the things I've been watching.
[16:35] <persia> For -mobile, I think it's safe to look at >4G, >384MB.  For -mid, maybe >2G, >256MB, but less than that will require a lot of changes.
[16:39] <persia> OK.  Forum setup now in responsible hands.  We should have forums later in the week sometime.
[16:40] <ogra> for mobile/mid ?
[16:40] <ogra> cool, who will read them ? 
[16:40]  * ogra is not a forums guy
[16:41] <persia> Erm.  I was hoping one of the people who said "cool" when I announced it :)  Maybe we'll get better response from the mailing list, when it's all sorted.
[16:42] <ogra> heh
[16:42] <persia> Maybe I'll read it occasionally, but since I live on IRC, I doubt I'll give it the attention it needs.
[16:44]  * ogra lives on mailing lists and IRC
[16:45] <persia> Yeah.  I hear email is the coming wave in communication :)
[16:46] <ian_brasil> persia: plus testing of the main apps too like xournal, liferea ..this is a good candidate for automation hence the dogtail question earlier
[16:49] <ogra> persia, really ? i thought twitter replaces IRC and mail together :)
[16:49] <persia> ogra: :)
[16:50] <ian_brasil> i am a little unclear ..exactly what parts of hildon does ubuntu-mobile use? ..if nothing then dogtail should just work on ubuntu-mobile 
[16:50] <ogra> nothing :)
[16:50] <ogra> ubuntu-mobile is plain gnome with different desktop setup and themeing
[16:51] <ogra> well, if you call midbrowser hildon that might count though :P
[16:51] <persia> I'm not sure it's worth running dogtail testing against each image, but perhaps running throughout the cycle as a source of bugs would make sense.
[16:52] <ian_brasil> well dogtail only works with the plain gnome desktop session so now it depends on what you did to the theme!
[16:53] <ogra> differnet panel layout, different applet layout on the panels ...
[16:53] <ogra> and pure gtkrc changes
[16:53] <ogra> nothing fancy that should get in your way
[16:53]  * ian_brasil is still waiting for ubuntu-mobile.img to download so he can play with it
[16:53] <ogra> the 30.1 one seems good 
[16:54] <ian_brasil> does it use compiz?
[16:54] <ogra> apart from the broken browser and missing autologin preseeding
[16:54] <ogra> yup
[16:54] <ian_brasil> ogra: then it will not work
[16:54] <ogra> ah
[16:55] <ogra> well, you can switch to metacity
[16:55] <ian_brasil> ah great
[16:57] <ogra> its the same desktop ubuntu-desktop uses
[16:58] <ogra> apart from the tweaks mentioned above
[16:59] <persia> Which means the same HW requirements as Ubuntu Desktop (>4G, >384MB)
[16:59] <ogra> well, i might cut down the default app selection a bit 
[16:59] <ogra> so we get to 2-3G
[17:00] <ogra> and 256M with intrepid ;)
[17:00] <ogra> we have compcache by default now
[17:00] <ogra> 384 was only for the heavy reqirements of the livecd
[17:06] <ian_brasil> persia: the idea is to set up scripts for all the main applications and just run them automatically
[17:08] <persia> ian_brasil: Right.  I think it's good to do that.  I'm just not sure if I want the release managers asking "Why haven't you had four independent dogtail runs for Alpha 3 yet?" in January.
[17:08] <persia> In summary; Yeah!  Let's dogtail it.  Let's just not make that a QA requirement for release until we can pass.
[17:10] <ian_brasil> i would suggest a dogtail run after the debian merge to catch regressions and then one for the Beta maybe
[17:11] <persia> I think we can be more agreesive than that.  I think one at DIF, one a FF, one at Beta, and one at RC is probably sensible.
[17:11] <persia> I don't think we need multiple runs per milestone.
[17:12] <persia> So, roughly, late January, late February, mid March and mid April.
[17:12] <persia> (unless someone monkeys more with the release schedule than I expect)
[17:14] <ian_brasil> that seems sane..it gives some time to actually write the scripts too...not that that is particularly difficult
[17:19] <persia> ian_brasil: Would you mind adding that to the list of tentative jaunty specs?
[17:19] <ian_brasil> persia: ok
[17:19] <persia> Thanks!
[17:22] <ogra> hmm, no bluetooth on the Q1 
[17:23] <ogra> at least i dont see a single device there
[17:26] <persia> No BT!  That's unfortunate.  You need a dongle.
[17:26] <ogra> hmm, the n800 doesnt see my headset either 
[17:26] <ogra> weird
[17:27]  * ogra just bought a bunch of BT devices so he can test
[17:29] <persia> ogra: Did you get any USB network or USB serial devices?  I haven't figured out how to test those with my collection.
[17:29] <persia> s/USB/BT/g
[17:29] <ogra> no
[17:29] <ogra> a headset and a gps reciever
[17:29] <ogra> and by end of the week i'll get a freedom keyboard
[17:29] <ogra> BT as well
[17:30] <ogra> but i wonder why the Q1 doesnt see anything
[17:30] <ogra> is our BT stack so broken ? 
[17:32] <persia> Yes.  It's completely broken.  Either install crevette's 0.28 (maybe applying mterry's patch for the pairing wizard), or use the ~bluetooth PPA for superm1's work on 4.x
[17:32] <persia> The former only fixes a couple of things, the latter seems to work fairly well (at least for my equipment)
[17:34] <ogra> are we likely to get either into intrepid ? 
[17:35] <ogra> seems serious to me that we show the BT applet but it has no function at all
[17:35] <persia> Yes.  slangasek is watching activity on the 4.x stack work, and 0.28 will be pushed as backup if it is deemed unsuccessful.
[17:36] <ogra> ah, good
[17:37] <persia> Excellent.  You can upload then :)
[17:45] <ogra> after beta i guess
[17:46] <persia> Yes, after beta.  StevenK and superm1 were chasing the rdepends earlier.  I'm planning a cycle retest once I can figure out how to preseed blank passwords.
[17:47] <ogra> cjwatson should be able to help you
[17:47] <ogra> (and tell you if it works at all)
[17:48] <ogra> might be thats not possible due to security considerations
[17:54] <persia> No, it's possible.  I have suggestions.  I'm just not quick with debconf and state machines yet.
[18:01] <ogra> oh, sweet, maemomapper works with google sattelite maps
[18:22] <lool> StevenK: Could you please remove the empty password for the user from preseeded data?
[18:23] <lool> StevenK: This isn't working, so this will at least allow to install with a password prompt
[18:23] <persia> lool: No point.  The empty stuff is ignored.
[18:23] <lool> persia: Oh so you tried and it's ignored
[18:23] <persia> StevenK: Please don't bother to change that: it will complicate me fixing it, and has no negative impact in the current state.
[18:24] <persia> lool: Yes, that's why I'm trying to hack user-setup to accept more preseeding options.
[18:24] <lool> persia: That's fine
[18:24] <lool> I'm asking because cjwatson wanted me to test it, and it saves me the testing
[18:25] <persia> Essentially, --automatic doesn't work.  If there is a ubiquity upload, the install works.  cjwatson's latest bzr changes were byte-for-byte compatible with my branch (although independently generated)
[18:25] <persia> What were you asked to test?
[18:28] <lool> Removing the preseeded data to see whether I can install
[18:29] <persia> With which version of ubiquity?
[18:29] <lool> Ok, I think everything is quite confused
[18:29] <lool> Allow me to take a deep breath and recap
[18:29] <persia> Yeah.  What are you trying to accomplish?
[18:30] <lool> a) I understand installer is currently broken due to some grub-ish problem specific to lpia b) you tell me some hours earlier today you'd fix this today c) this is fixed in upstream ubiquity d) cjwatson and you weren't sure this fix was useful to upload as some other thing I didn't know about (empty password string) is also breaking something
[18:31] <lool> I don't know much avec the breakage in d), I seem to understand that preseeding "" will be refused in the same way that an empty password would be refused if entered by the user
[18:32] <lool> Anyway, it seems cjwatson was suggesting I should test whatever I'd like to test to fix the password issue d) before he uploads ubiquity
[18:32] <persia> The reason the install fails is because the ubiquity grub-installer controller wasn't included in the lpia build.
[18:32] <lool> But then he decided it was easy enough to upload ubiquity
[18:32] <lool> Also, one of the proposed fixes to test was to remove the value from the preseeded data, but it seems this would get us exactly the same behavior
[18:33] <lool> At this point, I don't think I have anything else to do
[18:33] <persia> Oh.  OK.  In that case, I'll stop fussing with the password thing, and do the casper hack to put in the .desktop file for the non-automatic install with ubiquity 0.10.2.
[18:34] <persia> No, telling me that there's going to be a new ubiquity (which I hadn't heard), is the key thing you needed to do.
[18:34] <persia> It changes what I'm doing tonight.  I'll take a walk, clear my head, and do the other hack to add the Install icon to kourou *only* in the live environment.
[18:35] <persia> That needs a change to the image build, but ought to be able to be done for images tomorrow.
[18:36] <persia> I'm just not confident I'm going to absorb enough information about debconf to complete the password trick before I get too tired.
[18:36] <lool> persia: So --automatic breaks if there's a question which hasn't been preseeded such as the password?
[18:37] <persia> No, it's just insufficiently automatic, and the user has to enter a password, and then there is a password, which means that users without keyboards can't use admin tools, which is annoying.
[18:37] <persia> The fix is to enable the accessibility infrastructure, and get the input tools to use that for input, but that's *not* going to happen for intrepid.
[18:37] <lool> persia: I don't understand why you want to add a desktop file for non-automatic install then?
[18:38] <lool> persia: Is it because the current keyboard doesn't allow typing in e.g. gksu?
[18:38] <persia> Oh, right.  I suppose I could do it in automatic.  Doesn't matter.  I'm just not thinking as clearly as I might.
[18:39] <persia> The reason that we don't want a password is because the current keyboard doesn't allow typing in e.g. gsku, yes.
[18:39]  * ogra thinks thats rather a gksu issue
[18:39] <persia> That said, there's no reason for me to set it for non-automatic mode, because it works with --automatic, just not the way I want, because I can't preseed the blank password.
[18:39] <ogra> it should provide hooks for a11y apps
[18:40] <persia> ogra: Nope, because you can use a11y tools with gksu in -desktop.
[18:40] <ogra> yeah, -g might help
[18:40] <persia> ogra: The issue is that the a11y stuff isn't set up for -mid: ian_brasil has been complaining about this for some time.
[18:41] <lool> persia: So, with the new ubiquity we should get installable images starting tomorrow, except we will have to type a password and this will be a pain with gksu, correct?
[18:41] <ian_brasil> :)
[18:41] <ogra> well, make the gksu call in your .desktop file use gksu -g
[18:41] <ogra> then your input isnt totally locked
[18:41] <lool> And what remains TBD is that ubiquity needs to be launched by hand, and you're working on the casper bits to show a menu entru?
[18:41] <persia> lool: Yes, and the last bit waits on me sorting the password as a hack for intrepid, and we can look more agressively at a11y for jaunty.
[18:41] <ogra> if you have an onscreen kbd it will work
[18:42] <ogra> i just tested with cellwriter
[18:42] <persia> lool: Right, after I take a walk, I'll add the .desktop file.
[18:42] <persia> ogra: In -mid?
[18:42] <ogra> no, on my laptop
[18:42] <ogra> but it doesnt matter
[18:42] <persia> ogra: Yes it does.  Remember, all apps run *inside* hildon-desktop.
[18:42] <ogra> if you have a kbd that can be used for text input using -g with gksu will give you the ability to use it
[18:43] <persia> Well, that didn't work for me for update-manager for hardy.
[18:43] <ogra> did you try with -g ? 
[18:44] <persia> Yes.
[18:44] <persia> I haven't tried with -g in intrepid.  Maybe it's different.
[18:44] <ogra> GEEZ !
[18:44] <ogra> gksu even has a gconf setting for taht
[18:45]  * ogra adds that to ubuntu-mobile-default-settings
[18:45] <ogra> yay
[18:45] <ogra> keyboard input prob solved :D
[18:45] <ogra> why
[18:45] <ogra> didnt
[18:45] <lool> Ok folks, I'll leave there for today
[18:45] <ogra> i 
[18:45]  * lool waves
[18:45] <ogra> look that up earlier
[18:45] <ogra> ciao lool 
[18:47] <ogra> yippie, that WORKS !
[20:21] <persia> I need a filename that is guaranteed to exist *only* for ubuntu-mid images.  Anyone disagree that /usr/share/ubuntu-mid-default-settings/mid-gui-start might be such a file?
[20:26] <persia> Or maybe I don't.  Hrm.
[20:29] <lool> persia: /usr/share/doc/ubuntu-mid/?
[20:29] <persia> Grr.  I hate it when I spend hours troubleshooting something and find out that someone already fixed it months ago, and the documentation is out of date.
[20:29] <lool> ah eh
[20:29] <lool> I didn't do it!
[20:30] <persia> lool: No.  I just found a bug in kourou: for some reason it isn't showing the install icon.  That's the only missing bit to everything working for the non-automatic way.
[20:31]  * lool goes in non-computer mode for tonight now
[20:31] <persia> If we want to go automatic, we need to add a stanza to casper-bottom/10adduser to sed the .desktop file or something, but that's not beta critical.
[20:31] <persia> Good night lool.
[20:34] <persia> ogra: Just for double-confirmation, everything looks good for me in the latest -mobile image.
[20:40] <persia> StevenK: What's your feeling about adding ubiquity.desktop to Home?  How about Preferences?  It is currently parsed by Settings, but I think we're ignoring that menu, from what I can tell.  Separately, do you think we want to enable anything in settings.menu?  (And, yes, I don't mind waiting several hours for an answer)
[20:41] <ogra-maemo> persia, thanks
[20:42] <persia> ogra: One odd bit is that the close icon is third from the right, to the left of the brightness and volume controls.  Is it that way for you?
[20:42] <persia> Perhaps because I'm running at 800x600 right now?
[20:43] <ogra-maemo> hmm, yeah, likely... we can play with the setup after beta
[20:44] <persia> OK.  Do you want a bug for this, or is it just because I'm using a resolution no longer in vogue?
[20:44] <ogra-maemo> you cann force applets to the very right
[20:44] <persia> Well, I have to unlock them before I can move them, but if we can force hint it, I think the experience would be better.
[20:45] <ogra-maemo> i'm not sure it will work, but we can try
[20:49] <persia> OK.  I need to get some network adaptor on this working somehow :/  At the current time, it's *really* hard to send screenshots.
[21:43] <rhp_> I'm not getting anywhere with building an image myself in intrepid. Seems as if intrepid is both slow and buggy, could this be true?
[21:44] <rhp_> Running the latest standard image still gives me no wifi, so let's try to get that figured out. What would I do to get that fixed?
[21:50] <persia> rhp_: You're starting with a standard intrepid daily image?  This is -mid or -mobile?  (my apologies: it's late here, and my memory isn't what it ought be)
[21:52] <rhp_> -mid from cdimage.ubuntu.com/ubuntu-mid/intrepid/current/
[21:54] <persia> OK.  And you have *no* wireless.  Have you tried booting the same machine with an i386-based flavour?
[21:56] <persia> I think it's probably bug #274832, but if you might have a different bug of the class of bug #274704.
[21:57] <persia> If it's 274832, booting an i386 based flavour (e.g. -mobile) should have wireless.  If you can confirm this, then you're just waiting for packages that were last reported to be available "soon".
[21:58] <persia> If it's something like 274704, you'll need to collect information about the hardware, and post it to a bug.  If you can find a working wireless driver on the net somewhere, and point to it, that might happen faster.
[22:05] <rhp_> I doubt whether it is 274704, because the wifi nic is recognised, just not as a wifi card. The system thinks it is a wired card.
[22:05] <rhp_> So, am I correct that -mobile is for i386 devices and -mid is not?
[22:07] <persia> Well, the images that are available today happen to match those architectures.
[22:07] <persia> That's largely in part due to what was tested for what when, etc.  I hope we'll see both flavours architecture-neutral for jaunty.
[22:08] <persia> Current -mobile is built for i386 and current -mid is built for lpia.
[22:08] <persia> That said, I have an lpia device that works better with -mobile.  While I don't own one, I know of i386 devices that would be better with -mid.
[22:08] <rhp_> ok, confusing world
[22:09] <persia> At this point in time, the architectures haven't differentiated enough that you can't run i386 on lpia or lpia on (most) i386.  I don't know about the future.
[22:09] <rhp_> I'll have a go with -mobile. Currently I'm looking whether X will play nice with the fpit driver, which I got installed in the mean time.
[22:10] <persia> There's also a special kernel for lpia, the "lpiacompat" flavour, that allows one to run lpia on i586.  I don't know how well it works (I don't have any i586).
[22:10] <persia> You got -fpit to work!  Excellent.  Did you need to do anything special?
[22:12] <rhp_> I got fpit installed. Looking now whether it actually works.
[22:12] <rhp_> Aparently not directly, but I might have to tweak xorg.conf a bit...
[22:12] <persia> WIth the new Xinput2 hotplug, I suspect you'll need to feed HAL a .fdi file to make it work properly.  You may be able to hardcode stuff in xorg.conf, but there's some reports of that not working for some devices.
[22:13] <rhp_> oh, good to know.
[22:19] <rhp_> persia: it seems that none of the changes I made locally running ubuntu-mid (on the laptop) got stored during a reboot. That is correct?
[22:21] <rhp_> another silly question: since I cannot boot directly from usb stick, I need to copy the kernel and initrd.img of the hd first before I can boot. Would it also be possible to boot from pxe? That might make setting up the thing much easier. Then I can just put in a network cable for now...
[22:30] <rhp_> From the sight of it, that should be easy enough...
[22:31] <rhp_> I'm really having to get used to finding keyboards and not doing everything over ssh again ;-)
[22:32] <persia> You ought be able to boot from PXE, and it ought to work.
[22:33] <rhp_> Nice.
[22:33] <rhp_> ok, trying -mobile now...
[22:35] <persia> Any changes are lost on reboot.  There is a persistent mode to keep data between reboots.  It's something like that documented at https://help.ubuntu.com/community/LiveCD/Persistence?action=show&redirect=LiveCDPersistence
[22:36] <persia> My memory is that the image is persistent by default, so you can skip the bit about using F6.
[22:46] <rhp_> ah, ok. That was wat persistent was for... I skipped that when I created the grub bootline. I knew I should not have...
[22:47] <rhp_> Things are looking good! the -mobile version (which looks awesome, by the way) recognizes my wifi card just fine.
[22:48] <persia> OK.  Then you are probably hitting 274832, which is in progress, and ought be fixed soon.
[22:48] <persia> Which Stylistic did you have again?
[22:57] <rhp> So then... there I am!
[22:57] <rhp> Great work, to you all...
[22:58] <ogra-maemo> grr, Q1 wlan doesnt work after suspend
[22:59] <persia> rhp: Also, if you like -mobile for your size and resolution, the install works.  For -mid, it ought work with the next image respin.
[23:00] <rhp> respin is not the same as nightly build, I guess.
[23:00] <rhp> I kinda liked the user interface that -mid offers, can I enable that in -mobile too?
[23:01] <rhp> Although the graphics look way cool in -mobile.
[23:02] <persia> Not really.  The reason -mobile is i386 and -mid is lpia is because of some compilation differences between -mid and -mobile that need to get cleaned up.
[23:02] <rhp> But after that it would be possible?
[23:02] <persia> Anyway, subscribe to bug #274832: when it gets closed, you can switch back to -mid, and still have wireless.
[23:03] <rhp> ok, I'll do that.
[23:03] <persia> Yeah, after that it will be possible to switch between -mid and -mobile for any architecture, although there might be some leftover packages in either direction.
[23:03] <rhp> Funny though that it does recognise it as a wired nic...
[23:03] <ogra-maemo>  The reason -mobile is i386 and -mid is lpia is because i'm lazy :)
[23:03] <persia> But that's a jaunty thing: I'm guessing it will be in decent shape in February or so.
[23:04] <persia> ogra-maemo: It's not just you: some other people were lazy too.
[23:04]  * persia included
[23:04] <rhp> ok. First thing I'll do now is setup pxe-boot, so I can use the usb stick to install to.
[23:05] <rhp> I just tried to install -fpit, but apt-get update already ran out of space...
[23:05] <rhp> For now... time to go to bed.
[23:05] <rhp> See you again tomorrow.