People who speak below:

otavio    : Otavio Salvador
bubulle   : Christian Perrier
fjp       : Frans Pop
fs        : Frederik Schueler
zinosat   : Davide Viti
h0lger    : Holger Levsen

Hours are European Time (UTC +0100):


16:53 < bubulle> OK, people coming for the meeting you can begin to say you're here...:-)
16:54 -!- bubulle changed the topic of #debian-boot to: D-I team meeting: http://wiki.debian.org/DebianInstallerMeetings
16:55 < otavio> I'm here!
16:56 < bubulle> otavio: so we're two of us...nice..:-)
16:57 < otavio> bubulle: yes. It's growdy ;-)
16:57 < bubulle> given that I remembered only 2 hours ago that winter time means that 16:00UTC is 17:00 for me, i Guess that many others may have the same problem
16:58 < fjp> here
16:59 < bubulle> aaaaah..:-)
16:59 < bubulle> anyone else *really* here?
16:59 < bubulle> Houston, we have a problem
17:00 < fs> I'm here too :)
17:00 < otavio> oh!
17:00 < otavio> getting better!
17:00 < bubulle> OK, folks this is officially meeting time...but I'm afraid we're really a few of us
17:00 < otavio> fjp: did you tested a image using parted 1.6.25.1? worked fine?
17:01 < bubulle> fjp: I give you the call: do we still hold the meeting or do we reschedule it?
17:01 < fjp> Well, with only 4 ppl here there does not seem to much point to follow the agenda.
17:02 < bubulle> yep, I also think so
17:02 < bubulle> we have two options:
17:02 < otavio> bubulle: I think if we start to reschedule it'll be a mess since we won't try to stay at the schedule time
17:02 < fjp> If anybody has questions or points they'd like to raise, let's do those.
17:02 < bubulle> we can wait for 17:00 UTC in case some other people made the same mistake than me
17:03 < fjp> otavio: Just did an installation for sid_d-i. No problems.
17:03 < otavio> fjp: nice! good to kno
17:03 < otavio> fjp: next upstream version will have a lot of changes
17:03 < fjp> otavio: Only basic i386 though, nothing fancy.
17:03 < bubulle> fjp: we can at least talk about the user-setup-udeb thing, can we?
17:03 < otavio> fjp: sure
17:03 < fjp> bubulle: Sure.
17:03 < otavio> bubulle: go ahead
17:03 < bubulle> OK
17:03 < fs> I have prepared linux-di-amd64-2.6 based on 2.6.14-3
17:04 < fjp> fs: Nice. It'll have to go through NEW though.
17:04 -!- bubulle changed the topic of #debian-boot to: D-I meeting in progress. Topic: moving password questions in stage1. Keep this in shadow or move it to a udeb controlled by the D-I team?
17:04 < bubulle> OK, let's keep the micro-meeting on track, then
17:04 < fjp> That's also what we're waiting for for i386
17:04 < bubulle> about this user-setup thing
17:04 < bubulle> I have something nearly ready now
17:04 < otavio> fjp: what?
17:05 < otavio> bubulle: for user-setup-udeb or shadow itself?
17:05 < bubulle> fjp: my intent was asking whather someone objects that I move it in main D-I tree
17:05 < bubulle> otavio: for user-setup-udeb
17:05 < fjp> bubulle: No, don't see any reason not to.
17:05 < otavio> bubulle: i think makes more sense to have  it inside of d-i tree and also as a module
17:06 < bubulle> this is my opinion also...
17:06 < otavio> bubulle: later, it can also do other kinda  of setup for end-user and then would be great to use one module forit
17:06 < bubulle> yep...
17:06 < bubulle> so, I'll move from d-i-people/bubulle to packages/user-setup
17:06 < otavio> bubulle: also having it outside of shadow make eaiser to CDDs to change it
17:06 < otavio> bubulle: otherwise, force us to maintain shadow package
17:07 < bubulle> I'll keep the current code of passwd.config until beta1 is declared unsupported
17:07 < fjp> bubulle: Did you already talk to Kamion about it?
17:07 < bubulle> fjp: not yet
17:07 < bubulle> I imagined he would show up here
17:07 < otavio> bubulle: right
17:07 < bubulle> apart from the name, user-setup is barely what they do in Ubuntu anyway
17:07 < fjp> s/barely/merely/
17:07 < bubulle> and he said that the current way they do it is "ugly"
17:08 < otavio> bubulle: so would be good to try to merge and also replace the ubuntu package later
17:08 < otavio> bubulle: merge, if possible, of cours
17:08 < bubulle> OK. Well, about this topic I see no other point to talk about
17:08 < bubulle> Let go through other topics and see if we can discuss some..:-)
17:08 < otavio> bubulle: but currently, it'll only setup passwords, right?
17:08 -!- bubulle changed the topic of #debian-boot to: D-I meeting in progress. D-I BSP (joeyh) (10mins)
17:09 < bubulle> well, without joeyh, might be hard...
17:09 < bubulle> unless someone volunteers to organize a D-I BSP
17:10 < otavio> bubulle: why not do it together with normal BSPs?
17:10 < fjp> Joey mentioned it to me this week. He would have liked to hold the BSP this weekend, but because of other things decided not to.
17:10 < bubulle> next w-e?
17:10 < fjp> No, thanksgiving or something I believe.
17:11  * bubulle organized his own micro-BSP on localechooser...:-)
17:11 < otavio> hehe
17:11 < bubulle> fjp: so mor eor less joeyh suggested he could organize it anyway?
17:11 < fjp> Yes.
17:12 < fjp> Personally I doubt it will bring very much...
17:12 < bubulle> OK, then. Let's conclude this....this is the only thing we can do right now
17:12 < otavio> fjp: how is going d-i (dfb)?
17:12 -!- bubulle changed the topic of #debian-boot to: D-I meeting in progress. status of switch to 2.6.14
17:12 < bubulle> otavio: let's try to stay on track..:-)
17:13 < bubulle> fjp: we probably are missing a lot of people for that, don't we?
17:13 < otavio> bubulle: sure. I thought you're going to leave the rest of meeting for later. That's why I asked
17:13 < fjp> Hmm. Thought we said not much point in going through agenda...
17:13 < bubulle> fjp: yes maybe....I was trying to see whether at least one of the points could be discussed..:-)
17:13 < fs> I would like to commit some changes to kernel-wedge, so linux-di-amd64-2.6 can just include pcmcia-modules and scsi-extra-modules
17:14 < fjp> Basically most things are in place. Waiting for 2.6.14 for i386 to pass through NEW.
17:14 < fjp> fs: I think joey already OKed that?
17:14 < fs> he did
17:14 < fs> I'll commit then
17:14 < bubulle> what about geert's suggestion to break base-installer and have a kernel-installer?
17:14 < otavio> fjp: I saw a commit about reverting base-installer to  initrd-tool while initramfs-tools went in testing
17:14 < fs> should I wait for joeyh to upload the new package?
17:15 < otavio> fs: maybe, you could upload it?
17:15 < otavio> fs: and then put your code in?
17:15 < otavio> fjp: any problem with what I proposed to fs ?
17:16 < fs> then I'll might add myself to uplaoders and upload it
17:16 < bubulle> seems we don't go very far on that one too..:-)
17:16 < fjp> fs: Your call. If you're sure of the changes, go ahead. No need to add to uploaders if there's no bugs to be closed.
17:16 < otavio> i think
17:17 < fs> ok, even better
17:17 < fs> then I'll commit and upload
17:17 < otavio> fs: go ahead and upload it then ;-)
17:17 < bubulle> fs: in general I would say this is anyway the moment for "risky" changes as long as they don't break testing installs with beta1
17:17 < bubulle> fjp: do you agree?
17:17 < fjp> Yes.
17:17 < fs> bubulle: the changes won't break anything
17:17 < bubulle> this is why I commited a few longstanding patches to localechooser
17:17 < fs> in this case =)
17:18 < bubulle> fs: so it's even better
17:18 < fjp> Although we should keep things working enough so we can test 2.6.14
17:18 < fjp> bubulle: See my commit yesterday?
17:18 < otavio> fs: are any fix waiting on kernel-wedge for upload?
17:18 < bubulle> yeah, right...it seems we already have enough release goals for beta2 anyway
17:18 < bubulle> fjp: which one?
17:18 < CIA-6> debian-installer: fschueler-guest * r32270 packages/kernel/kernel-wedge/ (3 files in 2 dirs):
17:18 < CIA-6> debian-installer: Marked the following modules optional:
17:18 < CIA-6> debian-installer: - scsi-extra-modules: aha152x dpt_i2o dtc ibmmca sym53c416
17:18 < CIA-6> debian-installer: - pcmcia-modules: i82365 tcic
17:18 < CIA-6> debian-installer: Prepare for release.
17:18 < fjp> Sorting language list to keep C on top of list.
17:19 < bubulle> fjp: ah, yeasterday...yes, I saw it
17:19 < bubulle> seems fair by me. I'll release a new localechooser soon so that all this stuff may be tested widely
17:19 < fs> otavio: joey did one other change to kernel-wedge
17:19 < bubulle> OK. Anything else for "switch to 2.6.14"?
17:20 < fjp> Not really. Main thing will be to see what happens when/if discover is disabled.
17:20 -!- bubulle changed the topic of #debian-boot to: D-I meeting in progress. status of switch to udev
17:20 < bubulle> seems the transition is obvious
17:20 < fjp> And how the new initrd generators hold up.
17:21 < bubulle> not user we have around ppl with advice about this udev thing
17:21 < fjp> Is really part of the same thing.
17:21 < bubulle> yep, think so...nothing to add, then, right?
17:21 < otavio> about discover, I'll do a upload probably today
17:21 < otavio> fjp: is possible to migrate discover1* to testing now that beta1 was release?
17:22 < fjp> otavio: Already done I think.
17:22 < otavio> fjp: no
17:22 < fjp> otavio: Please ask debian-release then.
17:22 < otavio> fjp: doing it right now
17:23 -!- bubulle changed the topic of #debian-boot to: D-I meeting in progress. status of apt-setup
17:23 < bubulle> This is joeyh's baby
17:23 < fjp> One thing about discover. Should we switch to discover2 (mainly for 2.4 kernels)?
17:23 < otavio> fjp: did
17:24  * fjp has noticed weird effects from apt-setup
17:24 < otavio> fjp: I'm trying  to merge the databases and then start to drop discover1
17:24 < otavio> fjp: not now
17:24  * bubulle has not widely used apt-setup right now
17:24 < fjp> otavio: Should be done fairly soon if it needs to be done before Etch.
17:25 < otavio> fjp: also because discover2 is bigger
17:25 < otavio> fjp: I'll probably need to write a xml parser to be used in d-i
17:25 < fjp> otavio: Yes, I saw that. Could be a problem.
17:25 < otavio> fjp: so we don't need to suply libxml2
17:26 < otavio> fjp: but the major issue right now is to have both databases merge
17:26 < fjp> otavio: http://lists.debian.org/debian-boot/2005/11/msg00959.html
17:26 < otavio> fjp: so will be easier to maintain the packages
17:26 < fjp> Would be nice to get some response to that from discover1 and discover2 maintainers.
17:27 < bubulle> otavio: it seems to me that discover(s) lack manpower. Am I wrong?
17:28 < bubulle> fjp: but isn't discover1 maintainers just "us" (the d-i team)?
17:28 < otavio> bubulle: you're right
17:28 < otavio> bubulle: no
17:28 < otavio> bubulle: I was one of people working on discover1 and discover2
17:28 < otavio> bubulle: but now I'm also lacking time to work on it since I started to work hard on GRUB that was in bad shape
17:29 < bubulle> and how high is Progeny's work on discover2. I have no idea right now
17:29 < otavio> bubulle: after grub and parted (now both are in good shape again), I want to come back to discover
17:29 < otavio> bubulle: pere was one of people that did a lot of development of discover1
17:30 < fjp> otavio: IMO s/390 parted support has higher prio.
17:30 < otavio> fjp: discover2 is now part of debian
17:30 < otavio> fjp: yes? ok
17:30 < otavio> fjp: i'll work on that
17:30 < bubulle> anythign else about these things about discover(s)?
17:30 < otavio> bubulle: I'll probably release packages of both today
17:31 < otavio> bubulle: but nothing sshould change for d-i by now
17:31 < bubulle> something about apt-setup? Seems that no-one has things to say about it....
17:31 < otavio> fjp: about discover2 move, i think is better to wait while we merge the databases and see how to build a small version for d-i
17:32 < fjp> Just noticed that apt-setup causes the base-install for a netinst to be done over the network instead of from CD...
17:32 < fjp> Filing bug report.
17:33 -!- bubulle changed the topic of #debian-boot to: D-I team meeting in progress: status of graphical installer as part of daily builds and release
17:33 < fjp> otavio: Yes. Smaller version would be good. Also need to add sparc sbus support I think?
17:33 < otavio> fjp: after the meeting I  want help with hercules. OK
17:33 < fjp> OK
17:33 < otavio> fjp: yes
17:33 < bubulle> fjp: where are we with g-i in d-i builds?
17:33 < otavio> fjp: but sparc sbus wouldn't be hard
17:34 < zinosat> we need udebs for the fonts
17:34 < fjp> One blocking issue open: the udeb dependency resolving during build
17:34  * bubulle learns that zinosat is here..:-)
17:34 < otavio> fjp: dpkg patch waiting for inclusion, right?
17:34 < fjp> fonts are less important IMO as we have a good working workaround
17:34  * zinosat wawes! I arrived late
17:34 < fjp> otavio: "waiting" is too optimistic
17:35 < otavio> fjp: hehehe
17:35 < otavio> fjp: ok ok
17:35 < bubulle> fjp: but this is still a reasonable goal for beta2, right?
17:35 < fjp> I think I'll try to patch dpkg-dev with the patch and see if it really solves the problem.
17:35 < otavio> bubulle: probably will depends of dpkg issue
17:36 < fjp> And then try to push it.
17:36 < otavio> bubulle: could you add s390 parted support as beta2 goal too?
17:36 < otavio> bubulle: I'll try to manage it
17:36 < fjp> Alternative is to manually change all dependencies in the problem packages.
17:36 < fjp> otavio: Is already there.
17:37 < fjp> otavio: http://wiki.debian.org/DebianInstallerEtchBeta2Prep
17:37  * bubulle was impressed recently by the g-i look and feel
17:38 < bubulle> I'll make demos of it at a booth we'll have next week
17:38 < fjp> I plan to put together a wiki page with open issues and TODO list and then the announcement to d-d-a. Hopefully tomorrow.
17:38 < bubulle> and the Debian booth at Solutions Linux will feature of of my "babelboxes" running g-i
17:38 < fjp> Nice.
17:38 < otavio> fjp: where I can get a recent g-i image?
17:38 < bubulle> otavio: fjp home on gluck
17:39 < otavio> bubulle: ok. Let me grab for test
17:39 < bubulle> even if these font problems are not a big issue we should try not forget them
17:39 < bubulle> and still puch font packages maintainers to make udebs for us
17:39 < bubulle> push
17:40 < zinosat> the main usability proble IMO is the crash when switching to VT2
17:40 < otavio> bubulle: nobody provided patches for the packages yet?
17:40 < otavio> zinosat: this is serious
17:40 < bubulle> I think that Attilio works very close with GTKFB upstream about this, right?
17:41 < zinosat> especially on AMD64 where it seems impossible to get around it unless you start with the newt fe and use nc or something like that do redirect syslog over the network
17:41 < fjp> The fonts is mainly making very clear to maintainers what we want. What we don't want is udebs containing all fonts in a package, but only those we actually use.
17:41 < otavio> fjp: and this was already pointed out?
17:42 < fjp> Also, it would be nice to get some people thinking about how to strip fonts so that they only include characters we need.
17:42 < fjp> No.
17:42 < zinosat> bubulle, yes and looks like Mike is getting closer to the source problem (memory allocations)
17:42  * otavio getting current miniiso
17:42 < bubulle> zinosat: I'm confident that you ppl will successfully manage to get thourgh this
17:42 < zinosat> bubulle, me too :)
17:42 < bubulle> seems that G-I became a very nice exposure case for GTKFB
17:43 < bubulle> any idea of who could work on the themability?
17:43 < zinosat> I think that when this pb will be fixed, upstream will realize how important would be a more recent upstream release ( I alreadi raised th backport problem)
17:44 < fjp> I'll mention that in the announcement.
17:44 < bubulle> fjp: any idea when you can make the announcement?
17:44 < fjp> k
17:44 < fjp> [17:38:36] <fjp> I plan to put together a wiki page with open issues and TODO list and then the announcement to d-d-a. Hopefully tomorrow.
17:44 < otavio> bubulle: fjp said tomorrow
17:45 < zinosat> bubulle, BTW, have you seen the bugreport about themes and color-blindness (one of the many excuse to work on that)
17:45 < bubulle> fjp: /me hides..:-)
17:45 < bubulle> zinosat: yep, saw it...this is why I asked about who can work on it
17:45 < fjp> (Please ignore my apt-setup remark. Was using wrong image...)
17:45 < bubulle> zinosat: the color-blindness is the good excuse but CDD would also benefit from it
17:46 < otavio> bubulle: sure
17:46 < otavio> bubulle: I'm interested on it
17:46 < zinosat> one question: how should we handle g-i bugs? we cannot always blame it on cdebconf-gtk :)
17:47 < bubulle> otavio: interesting in trying to implement themability in gtk-debconf-udeb or whatever it is called?
17:47 < bubulle> zinosat: interesting quesiton yes...
17:47 < otavio> bubulle: interested to use it. I can't work on it now :(
17:47 < bubulle> fjp: which brings the g-i installation-reports issue
17:48 < bubulle> how will we gategorize them? usertags?
17:48 < bubulle> s/g/c
17:48 < fjp> Usertags for installation reports seems to be dead. I implemented it and reported it to the list, but see nobody picking it up.
17:49 < bubulle> hmmm, too bad
17:49 < bubulle> at least g-i could be a good use case...maybe a very short HOWTO for us lazy people?
17:49 < fjp> There's a huge backlog in answering installation reports anyway.
17:50 < bubulle> yep, that topic always shows up in meeting without real solution
17:50 < bubulle> I don't even know whether the <beta3 install reports have been closed
17:51 < bubulle> anything else to add on the g-i issue?
17:52 -!- bubulle changed the topic of #debian-boot to: D-I meeting in progress. adding partman-crypto? adding partman-auto-lvm?
17:52 < fjp> Not really. I'm probably going to limit my involvement to getting the localudebs isse resolved.
17:52 < bubulle> not sure we have someone with an advice here
17:52 < fjp> (after the announcement)
17:52 < fjp> auto-lvm has been uploaded
17:53 < fjp> crypto: not heard from xam yet if it is ready or not.
17:53 < bubulle> k
17:54 < bubulle> seems we cannot ad dmor ehere
17:54 < bubulle> wow, I'm tired
17:54 < bubulle> OK. las topic
17:54 -!- bubulle changed the topic of #debian-boot to: D-I meeting in progress. Switch to using codename instead of suite in sources.list for installed system
17:54 < bubulle> IIRC, fjp has ideas about this
17:54 < fjp> Yes.
17:55 < bubulle> From the last meetings minutes it seems that a majority would favor this
17:55 < fjp> Basic support to determine codename has been added to choose-mirror and cd/iso scan
17:55 < fjp> This has been uploaded.
17:55 < fjp> Next step is to change base-installer and apt-setup to use code-name instead of suite.
17:55 < fjp> So, no problem really.
17:55 < bubulle> this probably impacts several documents we have in Debian, right?
17:56 < fjp> Not really. Mostly next release notes I think.
17:56 < bubulle> yep, I was about to add it will certainly deserve a release notes entry for etch
17:57 < bubulle> oh, btw, I remember that all-UTF8 is supposed to be a realease goal for beta2
17:57 < fjp> Actually having codename will make upgrades more controlled IMO.
17:57  * h01ger just waves - sorry i'm too busy for d-i atm
17:58 < fjp> Don't know about UTF-8. If someone wants to implement it, that's fine with me but that person will also have to organize testing for it...
17:59 < bubulle> fjp: the trickiest part is console display and input in second stage and the installed system
17:59 < fjp> Impact on installation itself is probably minimal. Impact will be on installed system.
18:00  * bubulle can't tell this will be a very easy way but it seems that all i18n people agree to make the move
18:00 < fjp> bubulle: I have no real opinion on it. Encoding is not a real issue for English and Dutch (my languages)
18:01 < bubulle> I think we have enough clueful people such as eugen and/or the CJK people to handle the problems
18:01 < fjp> bubulle: If you can organize it, then fine by me.
18:01 < bubulle> we'll try..:-)
18:02 < bubulle> I'll also try to puch the switch for locales to Denis belocs-locales
18:02 < bubulle> push, damn
18:02  * chicky is thinking about creating wiki page with list of utf-8 broken apps
18:02 < fjp> I think the beta2 release should be aimed for early Januari. So if you can do it before then, fine.
18:02 < fjp> bubulle: Fine by me, but that discussion IMO is not a d-i discussion, but a d-devel discussion.
18:02 < bubulle> yep
18:03 < bubulle> OK, then. Seems we (badly) covered all topics of the meeting
18:03 < fjp> bubulle: Also a size issue. belocs-locales is quite a bit larger.
18:03 < bubulle> so I'll hereby declare the official meeting over
18:04 < bubulle> last topic as usual: next meeting
18:04 < fjp> bubulle: Also belocs-locales may be a problem for buildds as it would mean replacing the regular locales package.
18:04 < zinosat> bubulle, new years day 00:00 :)
18:04 < bubulle> according to what we decided in September, this should be a wednesday, either 14th or 21th of December
18:05  * fjp suggests 14 as 21 is too close to christmas.
18:05 < bubulle> and 20:00UTC (or 21:00 maybe.....just to keep the same "real" hours
18:05 < bubulle> fjp: I tend to agree for 14
18:06 < bubulle> anyone objecting for Dec 14th 21:00UTC?
18:06 < zinosat> bubulle, very fine for me
18:06 < otavio> fine tome
18:06 < bubulle> so, this is decided at the majority of the 6 people who are here..:-)
18:06 < bubulle> and I hereby declare the meeting over