[12:14] <aSt3raL> when will X be fully merged?
[12:17] <Hobbsee> aSt3raL: when it's done.
[12:17] <aSt3raL> haha ok
[12:17] <Hobbsee> how's that for an answer?
[12:19] <aSt3raL> it was more of a 'quit bugging me' than an answer :p
[12:19] <Hobbsee> well...
[12:22] <aSt3raL> Hobbsee: so what do you do?
[12:22] <Hobbsee> aSt3raL: i work on the kubuntu side of things, do a fair bit of bug triage, and am a MOTU, so work on universe stuff too
[12:22] <aSt3raL> ok cool
[12:22] <Hobbsee> + the community-ish side of kubuntu too
[12:29] <geser> aSt3raL: the xorg meta package was uploaded with version 7.2 10 hours ago
[12:29] <Hobbsee> heya jono!!!
[12:30] <jono> heya Hobbsee:)
[12:30] <jono> hows are we tonight?
[12:30] <jono> or morning
[12:30] <jono> or whatever it is in that crazy timezone
[12:30] <StevenK> jono: 10:30am
[12:30] <Hobbsee> jono: it's morning!
[12:31] <jono> heh
[12:31] <Hobbsee> strange thing - i almost never see it...
[12:33] <jono> hehe
[12:33] <Hobbsee> even the customers comment...
[12:34] <sabdf1> Hobbsee: you must be a morning person :-)
[12:35] <Hobbsee> sabdf1: heh.  nope :P
[12:36] <ajmitch> hey jono, sabdf1 
[12:36] <sabdf1> howdy ajmitch
[12:36] <jono> hey ajmitch
[12:37] <Fujitsu> Hi sabdf1, jono.
[12:37] <jono> hey Fujitsu
[12:37] <Hobbsee> (or identified)
[12:37] <jono> Hobbsee: he has multiple personalities
[12:37] <Hobbsee> jono: oh dear.  that's plus the random lunatics that pretend to be him, but talk like 13 year olds, right?
[12:38] <ajmitch> how worrying
[12:38] <Hobbsee> ajmitch: yes.  thank goodness for op powers.  (for stopping people disrupting channels)
[12:38] <Hobbsee> Fujitsu: heh, sorry :)
[12:39] <jono> Hobbsee: not really, Mark contacted me the other in fact and said "Hi Jono, Is it true you have m/\d l33t sk1llz? Power to the brother. Ubuntu FTW. See my website myspace.com/sabdfl"
[12:39] <kylem> Hobbsee, if it makes you feel any better, he's connected to internal irc from the same host. :)
[12:39] <Hobbsee> kylem: right :)
[12:39] <jono> Hobbsee: that may in fact be lies what I just said
[12:39] <sabdfl> FTW?
[12:39] <Fujitsu> kylem: You /must/ have been hacked.
[12:39] <Hobbsee> jono: yeah yeah :P
[12:39] <Hobbsee> sabdfl: for the world, i believe
[12:39] <jono> sabdfl: For The Win
[12:39] <sabdfl> niiiiice
[12:39] <jono> sabdfl: the kids of today seem to use it
[12:39] <ajmitch> jono: I wonder how much we'd need to get him to drink to say that :)
[12:39] <jono> I don't get it either
[12:40] <jono> ajmitch: sounds like a challenge
[12:40] <Fujitsu> jono: I'm a kid of today, and I had no idea what it means.
[12:40] <Hobbsee> oh wah.
[12:40] <Fujitsu> So there.
[12:40] <jono> Fujitsu: obviously not "l33t" enough
[12:40] <Fujitsu> Obviously.
[12:40] <jono> Hobbsee: how much does the Kubuntu community post to the fridge?
[12:40] <Hobbsee> jono: close to nothing.
[12:40] <jono> Hobbsee: I am looking to get it fixed up
[12:41] <jono> Hobbsee: any particular reasons?
[12:41] <Hobbsee> jono: mainly because we post the stuff direct to kubuntu.org
[12:41] <jono> Hobbsee: right
[12:41] <Hobbsee> which only Riddell has access to, currnetly, incidently
[12:41] <Hobbsee> or Riddell blogs directly, so appears on planet
[12:41] <jono> Hobbsee: do you feel the fridge feels ubuntu/kubuntu/edubuntu or just ubuntu?
[12:41] <Hobbsee> argh.  this supertux bug...and email...
[12:43] <Hobbsee> jono: at the moment, i dont even read the fridge - it seems that planet has depreciated the fridge
[12:43] <Hobbsee> it likely hasnt, but it's not active enough to be a snapshot of what's going on
[12:44] <jono> Hobbsee: yeah, I get the same views from some people, looking to fix this up a bit, and see how the marketing team can feed it better
[12:44] <Hobbsee> it's not even mentioning that UDS has been announced - at the minimum, the stuff from ubuntu-devel-announce should probably be on there, tc
[12:44] <Hobbsee> right
[12:45] <jono> exactly
[12:45] <jono> this is what I am looking to fix
[12:45] <jono> the fridge team is a little informal right now, it needs a leader and a roadmap
[12:45] <jono> and its a sexy job, I am sure a loving leader is out there
[12:45] <Hobbsee> i'm trying to figure out what of kubuntu should go on there
[12:46] <Hobbsee> maybe our testing team, or something, once that's ramped up a bit.
[12:46] <Hobbsee> at the moment, it talks conferences, teams, and asking people questions, from what i'm gathering, browsing it now - how would kubuntu fit into that?  (apart from the obvious of one of us answering questions)
[12:48] <Hobbsee> jono: sexy job hey?  interesting description :P
[12:48] <jono> Hobbsee: I reckon so, sexier than glibc maintenance
[12:49] <Hobbsee> hehe
[12:49] <ajmitch> that's not too hard
[12:49] <StevenK> jono: Oh come on, toolchain hackers get all the chicks.
[12:50] <canine_kouji> I'm wanting to create a couple programs for ubuntu. I'm not sure if anyone has done this already, so I'm checking. Has anyone here been working on a program or configuration application for GNOME for Windows services integration?
[12:51] <ajmitch> StevenK: oh is that what I have to work on?
[12:51] <canine_kouji> I do know how to integrate by hand, but for others.. I want them to be able to drink the ubuntu kool-aid ;)
[12:51] <ajmitch> canine_kouji: what do you mean by windows services integration?
[12:51] <lotusleaf> canine_kouji: you mean like what kcontrol has for wine?
[12:51] <canine_kouji> ajmitch: set up certain settings like active directory.
[12:52] <ajmitch> oh right
[12:52] <canine_kouji> adding the machine to the network
[12:52] <StevenK> ajmitch: :-P
[12:52] <canine_kouji> is there anything I can work on?
[12:52] <canine_kouji> specifically with this. I would consider it a very valuable feature. MacOSX already has an application.
[12:53] <ajmitch> the code is in feisty, just not very good right now
[12:53] <canine_kouji> ajmitch: if I can read Indian code, I can read your code ;)
[12:53] <ajmitch> in universe
[12:53] <canine_kouji> ajmitch: would you have a application name I can use to find this application?
[12:53] <_ion> What is Indian code? :-)
[12:54] <canine_kouji> _ion: code which comes from Rent a Coders from India.
[12:54] <kylem> c code tikka masala
[12:54] <_ion> I'd think there are competent as well as incompetent coders in India just like in any other country. :-)
[12:55] <ajmitch> the package is authtool, it's written in python, and I'm rewriting some reasonable sized chunks right now
[12:55] <canine_kouji> _ion: more code I read comes from India, I'm sure the college kiddies here in the US are just as illiterate and incompetant.
[12:55] <canine_kouji> ajmitch: awsome :) My favorite language
[12:55] <_ion> from __future__ import ruby
[12:55] <canine_kouji> using gtk, right?
[12:56] <canine_kouji> _ion: I'm waiting on YARV
[12:56] <StevenK> _ion: Have you seen 'from __future__ import braces' ?
[12:56] <canine_kouji> _ion: koichi has been very calm in implementing the byte code compiler for rb -> rbc for me. Ruby is almost there.
[12:56] <_ion> stevenk: Is there an actual implementation? :-)
[12:57] <_ion> canine_kouji: Yeah, YARV should be good.
[12:57] <StevenK> _ion: It's a more a joke.
[12:58] <canine_kouji> ajmitch: will authtool be included in ubuntu by default in future releases?
[12:58] <StevenK> >>> from __future__ import braces
[12:58] <StevenK> SyntaxError: not a chance
[12:58] <_ion> Ah, hehe. :-)
[12:58] <ajmitch> canine_kouji: hopefully :)
[12:59] <canine_kouji> ajmitch: I hope so too. to clarify, does GNOME have a smb browser?
[12:59] <canine_kouji> ajmitch: the smb browser is another one of the points of joining the network, to be able to access computers.
[12:59] <kylem> nautilus with gnomevfs.
[01:01] <canine_kouji> :) authtool is a dream come true
[01:01] <ajmitch> heh
[01:02] <ajmitch> most of what I'm doing is configuring pam, winbind, nsswitch.conf, etc
[01:02] <canine_kouji> ajmitch: you know of anyone working on a "add printer" for "look in directory"?
[01:02] <ajmitch> I still have to do the little things like the joining of the domain with 'net ads join'
[01:02] <ajmitch> nope, I haven't dealt with printer stuff
[01:03] <Hobbsee> yay, X didnt die.
[01:03] <canine_kouji> its just one of those "another cool features" OSX doesn't have look in directory, but it can scan for the extraneous data the printers send across the network for detection.
[01:03] <ajmitch> good
[01:03] <ajmitch> yes, cups can do that on ubuntu
[01:04] <Fujitsu> ajmitch: CUPS doesn't do Avahi-stuff, does it?
[01:04] <ajmitch> eg in the printer config tool, there's a 'detect LAN printers'
[01:04] <ajmitch> Fujitsu: I can't recall if it does yet or not
[01:04] <canine_kouji> ajmitch: I see it ;)
[01:04] <Fujitsu> I believe it doesn't.
[01:04] <ajmitch> oh well
[01:04] <Lathiat> no
[01:04] <Fujitsu> 'tis unfortunate.
[01:04] <ajmitch> something for Lathiat to fix :)
[01:05] <canine_kouji> looking in active directory for printers would be one of those features needed for the corp env :/
[01:05] <canine_kouji> ajmitch: gremlins :)
[01:05] <ajmitch> nah, it's going to the main archive
[01:06] <ajmitch> I think because the nz mirror isn't usable for many people
[01:06] <ajmitch> due to isp's peering squabbles
[01:08] <Lathiat> is there an IX over there?
[01:08] <ajmitch> yep
[01:08] <Lathiat> whats it called?
[01:08] <ajmitch> and telecom & telstraclear are the main two who refuse to peer nciely
[01:08] <Lathiat> oh, yes
[01:08] <Lathiat> naturally
[01:08] <ajmitch> WIX, and APE are the 2 main ones
[01:08] <Lathiat> its the same here
[01:08] <Lathiat> the big ones don't peer
[01:08] <Lathiat> because they dont need 
[01:08] <Lathiat> to
[01:08] <ajmitch> exactly
[01:08] <Lathiat> telstra/optus dont peer at any of the WAIX or PIPEs
[01:08] <Fujitsu> Same in AU, neither Telstra nor Optus do.
[01:09] <Fujitsu> Ah.
[01:09] <Lathiat> because they have prestigous tier-1 status :)
[01:09] <Fujitsu> Yep.
[01:09] <Fujitsu> Sooo annoying.
[01:09] <Lathiat> yeh
[01:09] <ajmitch> we also have smaller ones, with rather humourous names
[01:09] <Lathiat> its dumb
[01:09] <Lathiat> its the big players throwing there wait around
[01:09] <Lathiat> at the detrement of the internet as a whole
[01:09] <Lathiat> it sucks
[02:31] <Hobbsee> damned thing
[03:55] <bddebian> Heya
[07:23] <saddev> hi, I've a development policy question
[07:25] <saddev> Mono is able to run most applications that were written using the  IDE editor Visual Studio .NET 2005. If I'm developing an application that helps administering Ubuntu, and I write it with VS.NET. Granted that it has to work under Mono, would that application be considered for inclusion in Ubuntu, or there is some policy that prevents this because of VS.NET?
[07:26] <Treenaks> saddev: It probably should be compileable in Ubuntu as well
[07:27] <Treenaks> using mono
[07:27] <Treenaks> as long as it does that, IDEs/editors don't really matter I think :)
[07:27] <saddev> ok, so as long as msc is able to compile it under ubuntu, it won't matter that the design part of the app will be full of autogenerated Microsoft code?
[07:29] <saddev> and can I ask you another question about the inclusion process?
[07:30] <Treenaks> hm
[07:30] <Treenaks> autogenerated code
[07:30] <Treenaks> you should read the license of VS.NET what the license of generated code is..
[07:30] <Treenaks> +and figure out
[07:30] <saddev> so forget about Mono, say that I develop an app for Ubuntu (be it Mono, Python, etc...). If I want it to be part of the standard Ubuntu repository, how hard is it?
[07:31] <Treenaks> saddev: not hard (but we prefer to get new# apps through Debian)
[07:31] <saddev> Microsoft license... poor me. It's going to be hell! I better just use glade or GTK#
[07:31] <Treenaks> saddev: monodevelop ;)
[07:31] <saddev> Trennaks: yes, that's what I'm looking into
[07:32] <saddev> Trennaks: so a built open source application can be submitted for inclusion in Debian
[07:32] <saddev> and then picked up by Ubuntu?
[07:32] <Treenaks> it will automatically be picked up by ubuntu on the next release, or if it's before the freeze, you can ask for it to be included.
[07:33] <Treenaks> (Ubuntu syncs from debian unstable when starting to prepare a new release)
[07:33] <saddev> Treenaks: sounds very reasonable
[07:34] <saddev> Treenaks: I know it sounds silly. But being a developer I'd like to contribute with some cool little app or utility. Do you have any idea about what I could develop? Something that is currently missing from Ubuntu. perhaps?
[07:34] <Treenaks> saddev: uh.. no idea :)
[07:35] <saddev> leaving the hard part to ESR who just joined Ubuntu :-)
[07:35] <saddev> Trennaks: thanks. In what do you program if i may ask?
[07:36] <Treenaks> whatever I need :)
[07:36] <Treenaks> not much atm
[03:36] <niktaris> cjwatson, hi, can ubiquity create a root account in feisty ?
[04:31] <bddebian> Heya
[04:33] <towolf> ugh, fc-cache keeps strangling my system after an X-related update yesterday. an infinite loop or something?
[04:36] <towolf> it must be related to Type1 fonts because when I move gsfonts and some others away it runs through.
[04:45] <sladen> is Jono going to show up at FOSDEM, people are asking for him
[04:45] <cjwatson> niktaris: not yet, no
[04:45] <mjg59> sladen: No
[04:45] <cjwatson> niktaris: might be a reasonable thing to add to the Advanced tab, or maybe somewhere else, dunno
[04:46] <sladen> mjg59: funky, ta.
[04:46] <mjg59> cjwatson: I've just been chatting to the parted maintainer - the current upstream code is supposed to handle synchronising the gpt and mbr tables
[04:48] <cjwatson> interesting
[04:48] <cjwatson> though we need to get refit building on amd64 *anyway*, so it may not be worth ripping out the code we already have in place
[04:48] <cjwatson> unless there are other relevant things we need from upstream parted
[04:48] <mjg59> Why do we need refit?
[04:49] <cjwatson> well, gnu-efi
[04:49] <cjwatson> refit's easy after that, aiui
[04:49] <mjg59> We've got no way of modifying EFI flags or usefully blessing it in the installer, have we?
[04:49] <cjwatson> I just have an aversion to upgrading parted to new upstreams after UVF
[04:50] <cjwatson> it has a habit of breaking
[04:50] <mjg59> Yeah I'll see if I can find the patch
[04:50] <cjwatson> I guess I'd be happy with that then, yeah
[04:50] <cjwatson> would have to rip code back out of partman-efi, but there
[04:50] <mjg59> I think worrying about gnu-efi is probably unnecessary
[04:50] <cjwatson> (and yes we still need partman-efi to handle the EFI system partition logic)
[04:50] <mjg59> Yeah
[04:51] <mjg59> Also need to send him the HFS+ resize patch
[04:51] <cjwatson> if Kyle's discussion with linux-ide goes well, we'll need to patch parted for HPA handling too
[04:52] <mjg59> Hm. Not getting anywhere in finding this code.
[04:53] <mjg59> Need to try to find Dave Cantrell
[04:55] <kylem> dave cantrel..
[04:55] <kylem> why is his name familiar.
[04:55] <mjg59> RH developer
[04:55] <kylem> ah.
[04:57] <kylem> cjwatson, assuming i get any discussion at all :/ (posting on a friday was probably a bad idea :)
[05:06] <cjwatson> maybe I should have deferred my followup ;-)
[05:06] <cjwatson> Alan's point was a good one, certainly, and needs to be dealt with
[05:08] <cjwatson> kylem: maybe post a followup early next week with a patch adding an ioctl to let you find out the non-HPA size, and see what they say?
[05:08] <kylem> yeah.
[05:08] <kylem> problem with adding an interface is it has to stand th etest of time
[05:08] <cjwatson> it does feel like a bit of an excessively specific hack; I don't know if there's a more generic concept that it could be glued into
[05:09] <cjwatson> maybe the right thing would be "suggested size of area covered by a new partition table on this block device"
[05:10] <cjwatson> I need to learn more about partitioning on arches other than the big three
[05:10] <mjg59> ioctl?
[05:10] <mjg59> I'd have thought a sysfs attribute was more fashionable
[05:10] <cjwatson> well, BLKGETSIZE is an iocttl
[05:10] <cjwatson> ioctl
[05:10] <kylem> mjg59, block device crap is all ioctls
[05:10] <kylem> leeeegacy
[05:10] <cjwatson> should really live in the same place as it's the same kind of thing
[05:11] <mjg59> cjwatson: Not utterly sold on that argument :)
[05:11] <mjg59> But yeah. Adding a sysfs interface to block device magic is probably a job for later.
[05:11] <cjwatson> is BLKGETSIZE available via sysfs?
[05:11] <cjwatson> if not, I don't think something like it should appear in sysfs until BLKGETSIZE does too
[05:11] <cjwatson> otherwise you end up with random scattered bits of interface, none alike
[05:12] <niktaris> cjwatson, how about using a checkbox in the user creation dialogbox? "Create root account?"
[05:12] <cjwatson> which OK is kind of what the kernel already is but I don't think it should be encouraged
[05:12] <kylem> perhaps "BLKGETMAXSIZE"
[05:12] <cjwatson> kylem: maybe but that would break old partitioning tools wouldn't it?
[05:12] <cjwatson> assuming that you mean "get the full size of the block device including HPAs"
[05:12] <cjwatson> old parted wouldn't be able to work with partition tables overlapping the HPA
[05:13] <kylem> could return a tuple? :)
[05:13] <kylem> dunno.
[05:13] <cjwatson> niktaris: could work, feel free to try it out and send a patch
[05:13] <cjwatson> niktaris: not sure I'm entirely happy with it being presented by default but that's not a firm no
[05:13] <kylem> i don't really have a preference, in fact, my preference would really be to just do what the legacy code did.
[05:13] <niktaris> cjwatson, hehe will try
[05:14] <cjwatson> kylem: what does that mean though? return the size of the block device including HPA (since AIUI that's what the old code did)?
[05:14] <mjg59> This network is shonky
[05:14] <mjg59> Legacy behaviour is for the HPA to be utterly disabled and all ioctls to behave as if it had never been enabled, AFAIK
[05:14] <kylem> right.
[05:15] <cjwatson> mjg59: meaning that parted will see a block device whose extents include the HPA?
[05:16] <cjwatson> or the other way round? for some reason I am finding the terminology confusing
[05:16] <mjg59> Disabling the HPA means that the host no longer protectes the area
[05:16] <mjg59> Resulting in it becoming available
[05:17] <mjg59> That is, yes, parted will see a block device whose extents include the HPA (ie, the entire disk)
[05:18] <cjwatson> right, so if it's enabled, how would parted work with a partition table created while the HPA was disabled and overlapping the HPA?
[05:18] <cjwatson> would there be a separate bit of interface to disable the HPA again, or would the kernel detect that such a partition table was there and disable the HPA automatically?
[05:19] <cjwatson> maybe I misunderstood Alan then - if the HPA is enabled, as far as creation of new partitions go, all I want to know is how big the block device is now - I don't want to deliberately create partition tables extending into the HPA
[05:20] <mjg59> drivers/ide will *always* disable the HPA before userspace starts
[05:20] <mjg59> There is currently no interface for userspace to know it was ever there, other than parsing dmesg
[05:20] <cjwatson> which means that legacy partition tables will be broken
[05:20] <mjg59> When moving to libata, yes
[05:20] <mjg59> That's the issue
[05:20] <cjwatson> I think the kernel should detect such legacy partition tables and re-disable the HPA
[05:20] <cjwatson> since nothing else can
[05:20] <kylem> can't.
[05:20] <cjwatson> why not?
[05:21] <mjg59> To maintain consistency, the libata behaviour should be to always disable the HPA
[05:21] <kylem> the hpa detection happens on device discovery
[05:21] <mjg59> parted could then be enhanced to suggest that you not create or delete partitions in the HPA
[05:21] <kylem> partition tables occur way afterwards
[05:21] <cjwatson> ok, then I agree that libata should always disable the HPA, since otherwise you're fucked
[05:21] <kylem> right.
[05:22] <kylem> the problem is SATA.
[05:22] <kylem> which is almost universally libata based.
[05:22] <kylem> which means partition tables are based on the reduced size of disk.
[05:22] <cjwatson> there's no compatibility issue with moving from an enabled HPA to a disabled HPA, is there?
[05:22] <mjg59> No
[05:22] <kylem> it won't break, but people might fuck their shit up when they see "free space" in their partition tables.
[05:22] <cjwatson> the partition table will just only specify a reduced range of the disk
[05:23] <cjwatson> they won't, the free space is bounded in the partition table, isn't it?
[05:23] <mjg59> parted will notice that the partition table doesn't cover the entire disk and mark the remainder as free space, no?
[05:23] <kylem> hmm. i don't know. could experiment and find out.
[05:23] <kylem> mjg59, that was my thought.
[05:23] <cjwatson> er, I need to remember the partition table format
[05:24] <mjg59> I thought the mbr table just contained partition sizes, rather than attempting to define disk geometry?
[05:25] <cjwatson> mm, you're right, it doesn't contain the disk size, annoyingly
[05:25] <cjwatson> you could funt BLKGETSIZE to lie
[05:26] <cjwatson> but again that would break partition tables that deliberately use the HPA
[05:26] <cjwatson> so you need to teach parted, fdisk, and everything else to know about the HPA double-quick ;-)
[05:27] <cjwatson> have fun
[05:27] <kylem> mjg59, does the HPA region typically exist in the partition table?
[05:27] <kylem> i can't recall what my ibm had.
[05:29] <cjwatson> or add a userspace interface to enable/disable the HPA and then at least a workaround is available
[05:29] <mjg59> cjwatson: Surely always disabling the HPA without providing any intelligence is no worse than the current situation?
[05:29] <cjwatson> http://lkml.org/lkml/2005/9/2/213
[05:30] <cjwatson> mjg59: depends how vital the contents of the HPA are
[05:30] <cjwatson> if the BIOS relies on it ...
[05:31] <kylem> iirc it's used for BIOS suspend on some old thinkpads.
[05:31] <kylem> er, hibernate.
[05:31] <mjg59> cjwatson: As I said, how is it worse than the current situation?
[05:31] <mjg59> If people were breaking stuff, we'd already have heard about it
[05:32] <cjwatson> with the current situation you can't break your HPA on fresh installation
[05:32] <mjg59> Yes you can
[05:32] <cjwatson> (if it's using libata)
[05:32] <mjg59> Right
[05:32] <cjwatson> I'm not convinced we'd have heard about it
[05:32] <mjg59> But you've been able to for anything using drivers/ide up until now
[05:32] <cjwatson> I have nowhere near enough bandwidth to keep up with all the installer bugs
[05:32] <kylem> the problem currently is that anyone upgrading to feisty and using libata-pata will suddenly not be able to use their disk.
[05:33] <kylem> and will scream in panic as they think their /home is fried...
[05:33] <mjg59> kylem: It's presumably not hard to limit this to pata, even if that's not what upstream want?
[05:33] <kylem> mjg59, it is indeed not terribly difficult.
[05:33] <mjg59> That then leaves us where we were pre-feisty
[05:33] <cjwatson> anyway, must go tidy the house some more
[05:33] <kylem> i can read cable type and limit to 40p/80p PATA.
[05:33] <mjg59> Ok
[05:34] <kylem> i can also add an ioctl to increase the size and revalidate the disk
[05:34] <kylem> that way if we wanted to let ubiquity or something detect it and provide a clickbox to disable it, we could.
[05:34] <kylem> ugh, then we have to put it into the initramfs and.. oh bother.
[05:35] <kylem> partitioning is hard. let's go shopping.
[06:06] <kylem> mjg59, i'm tempted to say we could just leave HPA enabled wholesale for new installs...
[06:20] <kylem> hmm. did ubuntu disable middle click in firefox? bummer.
[06:20] <Mithrandir> it's been disabled since warty or so.
[06:20] <kylem> gruntle.
[06:20] <Mithrandir> just set middleMouse.contentLoadUrl to true
[06:21] <kylem> heh, was looking through the changelog for that.
[06:21] <phoenix24> but simple tweaks can enable it.
[06:37] <mjg59> kylem: How do you define a new install?
[06:38] <kylem> thinking more about it, i don't think there's a single good way to do this.
[06:38] <mjg59> I think the least worst is to do what drivers/ide did in the kernel, and make userspace smarter
[06:40] <kylem> yeah.
[06:40] <kylem> i'm trying to determine call chain when the disk gets revalidated from ioctl.
[06:40] <kylem> to make sure i'm handling it correctly.
[06:41] <kylem> mjg59, one of the thoughts i had was to just stow way the full disk size, and if there's a BIO which gets rejected as within limited_size <= x <= hpa_size, disable hpa and retry the bio.
[06:42] <mjg59> Hm.
[06:42] <mjg59> kylem: Sounds distressingly hacky
[06:42] <kylem> no fucking clue how feasible it is.
[06:43] <kylem> alternatively, hack the msdos partition (ugh, i guess efi too) code to see if the results make sense and unclip the disk...
[06:43] <kylem> there's sata disks using drivers/ide too, which is also a problem with just basing it on cable type.
[06:44] <mjg59> Nf.
[06:44] <mjg59> We lose heavily.
[06:44] <kylem> everyone loses.
[06:44] <mjg59> Whoever wins, etc.
[06:45] <kylem> i don't know how much of this should be solved for feisty, maybe we should defer libata-pata until it's more mature...
[06:45] <mjg59> Well, given that we can trivially ensure that it has much the same level of functionality as it had in edgy...
[06:47] <kylem> i don't know that the error recovery is as good in drivers/ata yet.
[10:26] <tsmithe> hi - if i have a package waiting on NEW, but it has a bug, is it possible to have a fixed version uploaded on top?
[11:54] <jdong> mjg59: any emotional attachment to xinerama on xgl?
[11:54] <jdong> mjg59: we need new xgl from git to compile against xorg 7.2....
[11:54] <jdong> and that xinerama patch no workie