persiarbelem: Hey.  Haven't heard from you in a few days.  How is it going with the liquid stuff?04:38
rbelemhi persia 07:26
persiarbelem: Hey.  So, how's plasma-mobile?07:35
rbelempersia, i'm finishing your todo list :-)07:36
rbelempersia, these days i could not work overnight :-(07:37
rbelempersia, now finding the right build-deps07:41
rbelempersia, now it is 03:41 :-D07:42
persiaGo to bed.  Grab me during the day sometime.07:42
rbelempersia, i'm not sleepy now07:43
persiaOK, if you'd rather do it now than in the morning :)07:44
* persia is a morning person, so would generally prefer to go to bed and do it tomorrow07:44
rbelempersia, ok, 11:00 UTC is good for you07:48
persiaSure, if that's when you're up.07:48
persiaI just want to try to get all three packages in before the meeting on Tuesday.07:49
rbelempersia, plasma-mobile is almost done, so we can finish at least kdm tomorrow07:52
persiarbelem: Sounds good.  Any ideas about kwin?  Do we need to do something like in plasma-netbook?07:54
rbelempersia, do not know very well yet07:56
rbelempersia, maybe we should ask in kwin channel about "kwin for mobile", "What is missing to be done", "virtual keyboard"07:58
* persia grumbles about OpenGL-ES07:59
persiaAs long as we only use the subset of GL in OpenGL-ES, we ought be safe cross-platform.08:00
persiaIdeally, drivers that provided OpenGL-ES would have fall-back to mesa for unsupported calls.08:00
persiaThat way what can be accelerated is accelerated, and what can't be is still supported (if slowly).08:01
persiaBut, really, someone other than I is the right person to do that :)08:01
persiaFor the keyboard, would klavier work, or do we need something else?08:02
persiaI don't know anything about KWin: asking somewhere else might be a good idea.08:02
rbelempersia, i think we can get some idea from the matchbox08:04
persiaOn the keyboard, you could also try kvkbd, but I think klavier is a cooler name for a package, and so would probably try that first (on the unreliable basis that niftily-named software is better)08:04
persiaWell, I'd rather use the KDE solution if possible.  The way matchbox-keyboard works is very messy.08:04
persia(plus I still haven't merged the latest version from Debian)08:05
rbelempersia, hum...08:05
rbelempersia, how matchbox keyboard works?08:05
persiarbelem: it uses XEMBED to fit itself inside some other environment, and do the right thing.08:07
persiaThis gets messy if you want to handle use cases where people are switching focus a lot, etc.08:07
rbelempersia, there is a plasmoid called plasmaboard08:08
persiaHooking into either the accessibility toolkits or the input-method toolkits (for CJKV support), tends to be a more pleasing user experience for multi-application environemnts, because the user doesn't have to be concerned about which app has the keyboard embedded.08:08
rbelempersia, cool!08:09
rbelempersia, that's a very nice alternative i think08:09
persiaUsing the keyboard plasmoid sounds like a really good ieda, except "KEYBOARD DOES NOT WORK" is the top of the list of current known issues :)08:10
persiaWe may have some work to do.08:10
rbelemehehhe :-)08:10
rbelemthat's nice :-)08:11
persiarbelem: a11y or IME are the standard way DE software keyboards work.  The XEMBED stuff only happens in the embedded world.  So if we pick stuff from the desktop world (that runs fast enough on a handheld), we'll have more flexible solutions.08:11
persia(simply because for desktops, there is stricter separation between window-manager, session-manager, and application)08:12
rbelempersia, i have some gadgets to test this08:14
persiarbelem: Actually, it looks like KDE 4.3.95 just hit, so there's a decent chance that the keyboard plasmoid works in lucid.08:14
persiaExcellent.  None of my gadgets can run lucid currently.08:14
rbelempersia, if you need one, i can send it to you08:15
rbelempersia, i have one n810 extra08:15
persiaDoes the n810 run lucid?08:16
persiaI didn't think there were kernels.08:16
rbelempersia, i have one kernel ready for it ;-)08:16
persiaWell then.  Tell me how to make one :)08:17
rbelempersia, cool! i was working with a co-worker to get it running08:19
rbelempersia, i think these days he will publish the howto in his blog08:19
persiaThat'd be great!  I know there's always calls for more testing on armel, but right now there's not many users (but there are lots of n810s floating around).08:20
rbelempersia, and with 3d working :-)08:21
persiaReally!  With a proper X driver and everything?08:22
rbelempersia, texas released the driver08:22
persiaExcellent!  Now I have a non-virtual target for liquid :)08:23
rbelempersia, finished the search for build-deps08:26
rbelempersia, i will sleep a little bit more :-)08:26
rbelempersia, 07am i will wake up08:27
persiaOK.  I might still be around when you get up (depending on how late you sleep).08:27
persiaThat's not long at all :)  I'll be around then.08:27
rbelempersia, you will stay up until what time?08:28
persiaUsually around 15 UTC, but sometimes later, depending.08:28
persiaIf we're making good progress, I'll stay up late.08:29
rbelemcool :-D08:29
rbelemsee you in two hours08:29
persiaSleep well: I won't mind if it's three :)08:29
rbelemthanks persia 08:30
rbelempersia, uh! i slept a little bit more what i planned12:55
persiaThat just means you're more well-rested and ready do hack :)12:56
rbelempersia, yep ;-)12:56
rbelempersia, i will finish the debian/control descriptions12:59
persiaOK.  So last we talked, you were looking at debian/copyright, right?12:59
persiaAh, right, descriptions.  Aside from the KWin bits, those might not be completely wrong :)13:01
rbelempersia, i will get something to eat13:03
persiaHeh.  OK.13:03
asacah .. a sunday liquid sprint ;)13:10
rbelempersia, i'm ready!13:27
persiaOK.  You've made the editorial changes to control and fixed copyright?13:27
rbelemhi asac 13:28
persiaOK.  Did you build and test the package?13:28
asachi rbelem 13:28
rbelemusing pbuilder13:28
persiaThen I think we're probably done with plasma-mobile.  Push to review, and I'll double-check13:28
persia(and maybe we can get asac to review and upload, since he's around)13:29
persiaSo, next up: kdm-mobile13:29
persiaMy memory is that we were going to branch the kdm source, and apply special patches.  Is that correct?13:31
rbelempersia, yep13:31
persiaOK.  You've downloaded the kdm source, and you have the patch?13:32
rbelempersia, we need to patch the build system to build outside kdebase13:32
rbelempersia, hum... not yet :-(13:32
persiaOK.  That's first :)13:33
persiaOnce we have sources, we can trim out the package.13:33
persiaI have a feeling we're going to discover some issues with the headers and libraries provided by kdebase, and so need to file some bugs, but lets give it a try first :)13:33
rbelempersia, will we use the get-orig-source?13:34
* persia is still downloading the kdebase-workspace source, and will want to see what is done there to have an opinion13:34
rbelempersia, i already git svn the kde trunk :-)13:35
persiaDo you want to work from trunk or from the kdm in Ubuntu?  I'd rather work from the kdm in Ubuntu in hopes of defining some semi-automated process to stay in sync.13:36
rbelempersia, it took weeks to finish13:36
persiaOh my.  It's 80MB for kdebase-workspace Ubuntu source.  How long will that take you?13:37
rbelempersia, cool! kdm from ubuntu is better13:37
rbelempersia, i already downloaded13:38
persiaOK.  It might be some work, but I hope we can write a script that converts the kdebase-workspace source into our special kdm-mobile source.13:38
persiaThat way we can easily update after each kdebase-workspace upload.13:38
rbelempersia, kdebase uses bzr13:40
rbelempersia, so what is in bzr is what is in the package?13:41
persiaUpstream or lp:ubuntu/kdebase-workspace ?13:41
rbelempersia, let's use lp13:42
persiaI meant to ask "when you say "what is in bzr", do you mean some upstream bzr or lp:ubuntu/kdebase-workspace?"13:43
persialp:ubuntu/kdebase-workspace is just the same as the package contents, except in bzr.  It gets a fresh commit for each upload.13:44
persiaSince we're only grabbing the package to massively hack it and extract the tasty bits, I don't think bzr will help us.13:44
persiaThere may be some lp:kdebase-workspace which is a vcs-import of upstream svn (I haven't checked).  If so, I'd prefer not to use it for the same reason that I wanted to use the Ubuntu package initially.13:45
persiaSeems this contains KWin also.  Shall we do a single kdebase-mobile that contains both kdm and kwin?13:46
rbelempersia, i think is better use separated sources13:46
rbelempersia, so we can patch the build system and try to get it upstream 13:47
persiaYou think upstream would be interested in splitting these out?13:47
rbelemyep, they are moving to git13:47
persiaAh, so they will want to be separated sources anyway?13:48
persiaSo while we're working on kdm-mobile and kwin-mobile now, we might expect that in the future there would be a kdm source package that generated both kdm and kdm-mobile?13:49
rbelempersia, that's it :-)13:49
rbelempersia, after that we can help upstream to split all kdebase :-)13:51
persiaOK.  I still think we want to go with the extract-the-source-with-a-script-from-kdebase-workspace approach for now, just because of the caution of the kubuntu folk for LTS, but we can certainly structure things so that transition to a separated package is easier in the future.13:51
persiaIs this an upstream goal already?13:51
rbelemi will paste the link for you13:51
persiaWell, hrm.  That does sound tempting.13:51
persiaDo you already have some ideas for patching the build system, or are we starting from scratch?13:54
rbelempersia, it should be almost the same for plasma-mobile13:56
persiaOK.  Easy enough then.13:57
persiaCopyright should be easier this time, because we can use kdebase-workspace debian/copyright as a reference.13:58
rbelempersia, http://techbase.kde.org/Projects/MovetoGit13:59
persia"Main modules such as kdelibs and kdebase will each become one repository." - are you sure that splitting is going to be accepted?14:01
rbelempersia, i talked to a kde developer and he said yes14:02
persiaOK.  I'll trust you on that :)14:02
rbelempersia, i will confirm this on #kde-git14:02
persiaNo need to do so now.  Let's do the packaging, and look at that later.14:03
persiaAre you pushing plasma-mobile to REVU?14:03
rbelempersia, not yet14:03
rbelemi will push right now14:03
rbelempersia, i have to create another gpg key14:04
persiaWhat's wrong with the one you have?14:04
asacwhat ETA is  that review?14:04
rbelempersia, my gpg that is in ubuntu keyring is outdated14:04
asaci will be travelling for a few hours soon14:04
persiaasac: Soonish, or later, at your convenience.  I just won't upload without a second.14:05
persiarbelem: Do you still know the passphrase?14:05
asacpersia: i cn do it in abuot 3-4 hours14:05
asacjust post the REVU url to me i will pick it up14:05
persiaasac: Cool.  Thanks a lot.14:06
persiarbelem: Do you still trust the integrity of the expired key?  Is the secret key still secret?14:06
rbelempersia, yep14:06
persiaWell then, just edit the key and extend the expiration date.14:07
persiaI do this once a year, but leave it set to expire so that it dies when I can't make it work anymore.14:07
persiaOnce you've extended the key, you need to push it to the keyservers again, etc.14:07
asackeep the revocation cert and set to infinite ;)14:08
asaci usually use subkeys for the real signing and throw them away regularly14:08
persiaasac: I can't guarantee that I can retain the integrity of both my secret key and my revocation certificate for eternity.14:09
persiaFor instance, I could be hit by a bus, and no longer able to either revoke my key or enter my passphrase.14:10
persiaBut I tend to perhaps be overly careful :)14:10
asacyou should tell your lawyer to do that ;)14:10
asaclast wish: please revoke my key :-P14:10
persiaAmusing, but morbid.14:11
persiaBack to packaging!14:11
rbelempersia, ehehe14:11
rbelempersia, some of the e-mails, that is in the gpg key, i do not use anymore14:14
rbelempersia, you think this is a problem?14:15
persiaI believe in removing any identities from a key if those identities are no longer controlled by the key holder.14:15
persiaBut I know other people who have lots of leftover identities on their key.14:15
rbelempersia, and the key strength is only 102414:17
rbelempersia, i created it in 2004 before debconf4 :-)14:17
persiaAh.  Yes, you might want to create a new key :)14:18
rbelempersia, so after create a new key it will be not signed by anyone14:19
rbelempersia, may i mail to already signed my current key and ask to sign the new one?14:20
persiaYou may, but they may not sign it.14:22
persiaREVU doesn't require a WoT though.14:22
persiaSo your work is not blocked for today, but you'll want to get to a keysigning when you can.14:22
rbelempersia, cool!14:23
persiaI've just tossed together a quick wrapper package around the kdm sources from the kdebase-workspace source in lucid, and will see how badly it doesn't build :)14:25
persiacopyright is still 0-byte, and I've not patched yet.14:26
rbelemok! :-)14:26
rbelempersia, i will create a 8192 bit key14:33
persiaShould last a little while, although not terribly long :)14:34
persiarbelem: http://paste.ubuntu.com/361980/14:55
rbelempersia, unknown CMake command "install_pam_service"...14:57
persiarbelem: There's a FindPAM.cmake in kdebase-workspace/cmake/modules14:58
persiaI just don't know anything about cmake, so am relying on you to tell me what to do :)14:58
persiaMy debian/ is 0-byte copyright, minimal control building kdm-mobile binary, mostly copied from kdebase-workspace, except adding Conflicts/Replaces kdm, echo 7 > compat, and rules.tiny + --with kde15:00
persiaOh, and a stub changelog entry.15:00
rbelempersia, :-)15:00
rbelempersia, i think just need to copy the cmake dir to kdm15:01
persiaThe entire cmake dir?15:01
rbelempersia, just what is needed15:02
persiaSo, CMakelists.txt and FindPAM to start?15:02
rbelempersia, and add to main CMakeList.txt add_subdirectory(cmake)15:02
persiaOr do you think I need more of them?15:02
rbelemfor now just findpam15:03
rbelemi think15:03
rbelempersia, i'm compiling things with -j8 in home :-)15:04
rbelempersia, i setup yesterday a faster wireless network :-)15:04
rbelempersia, and installed icecc on each machine15:05
persiaHeh.  That goes fast, but I wonder about how well it matches the distro builders :)15:05
rbelempersia, all machines run ubuntu karmic here, just mine runs lucid15:07
rbelempersia, and icecc takes care of packaging the toolchain and send to others machines15:08
persiaAh, toolchain packaging.  Now I know why you need such a system :)15:10
persiaThere's very little I end up compiling that won't compile within half an hour, even on my little 512MB box that thrashes when compiling bunches of things.15:11
persiaBut I do need to do something about my configuration.  Having a local mirror would speed stuff up, since sometimes downloading build-depends takes a bit, and I do want to figure out a nice way to not have my build chroots be encrypted.15:12
persia(but that involves a full reinstall, which I never seem to have time to do)15:12
rbelempersia, i started to write a little script to create mirrors15:19
persiaAnyway, the build-deps are redownloading.15:19
persiaThere are a bunch of good mirror scripts.15:20
rbelembut based on germinate output15:20
persiaI just usually relied on my bandwidth to ignore them, but today I've been running as many as 7 simultaneous builds, and so found it a bit restrictive.15:20
persiaOoh.  You might want to get in touch with cjwatson if you make progress on that.  The ArchiveReorganisation effort will need those, and he'll know with whom to speak about it.15:21
rbelemi'm using apt-cacher as a simple mirror15:21
rbelempersia, i already talked to him and he commited some changes, but not this script yep15:22
persiahttps://wiki.ubuntu.com/ArchiveReorganisation/Components has some of the specs about how it might work.15:22
rbelemafter this script jigdo is dead :-)15:23
persiaBut doesn't jigdo also do the image creation?15:24
rbelemmy goal is to download only what is need and then build the image15:24
rbelemso, it will make easier the life of the people that what to create a new ubuntu flavor15:25
persiaI guess.  I usually try to do stuff with live images myself.15:26
persiaOnce one creates the meta, and tunes a system with deborphan, it ought be easy to recreate this with a livefs build.15:26
persiaAnd if one has a good flavour, and is supporting it, getting it built on the livefs builders doesn't seem especially hard.15:27
rbelemi did not used livefs yet :-(15:30
persiaOh.  I find it easier for creating install CDs.15:32
persiaAnd if one isn't doing an install, only the meta matters :)15:32
persiaAnd that's *with* the FindPAM in cmake/modules15:34
rbelempersia, can you send the debdiff to me :)?15:42
persiadebdiff against what?15:43
persiaI'll send you the tar.gz of the (currently native) package :)15:43
rbelempersia, cool! :-)15:43
persiarclbelem@ , right?15:44
rbelempersia, yep15:44
persiaOn the way.15:45
rbelempersia, gnupg finished the generation of my 8192bit key :-)15:45
persiaOnce we get copyright sorted, we can use bzr, which might be easier to collaborate with :)15:45
persiaCool.  Let me know when you stick it on REVU, and I'll take a look.15:45
rbelemdownloading the tar.gz15:49
persiaIt's just a quick hack: don't expect much from it :)15:49
rbelempersia, found the install_pam_service16:14
rbelempersia, http://paste.ubuntu.com/362026/16:15
persiaSo we just need to copy in more stuff from kdebase-workspace?16:15
rbelempersia, yep16:16
persiaExcellent.  So it builds now?16:18
rbelempersia, not yer16:18
rbelempersia, works!16:25
persiaDoes the produced binary work?16:25
rbelempersia, hum... got an error16:26
rbelempersia, http://paste.ubuntu.com/362034/16:28
* persia digs16:29
persiaI'm confused.  Shouldn't that be set in kdm/ConfigureChecks.cmake ?16:31
persiaspecifically, with set(KDM_LIBEXEC_STRIP 0)16:31
rbelempersia, yep16:31
rbelemi will paste the cmakelist.txt16:32
rbelempersia, http://paste.ubuntu.com/362038/16:33
persiarbelem: I wonder if we need something in KDE4Defaults or some such.16:35
persiaPerhaps we don't have strong enough build-deps?16:35
rbelempersia, that's true16:38
rbelempersia, let's apt-get build-dep kdm16:38
rbelemjust to get i building16:39
persiaAll of it!16:39
persiaI was trying to trim build-deps :)16:39
persiaBut yeah, just copy the build-deps line from kdebase-workspace if you want it all.16:39
rbelemafter that we trim it16:40
rbelempersia, found the solution16:46
rbelempersia, need to copy more files16:47
rbelempersia, it got a problem :-(17:12
rbelempersia, did not find bgrender.h in any package17:12
persiaShould be in kdm-mobile/kdm/background/bgrender.h17:14
rbelempersia, it is in kcontrol subdir17:14
persiaAh, oops!17:15
rbelempersia, kdebase-workspace-4.3.95/kcontrol/kdm/background/bgdefaults.h17:15
rbelempersia, it should be in kcontrol-dev package17:16
rbelempersia, but it does not exist17:16
persiarbelem: Well, we've two choices.  We can try to fix kontrol-dev (and wait), or we can copy the file for now.17:17
persiaI think we should copy the file for now, and fix kontrol-dev later.17:17
rbelempersia, ok let's do this17:18
rbelempersia, http://paste.ubuntu.com/362065/17:22
rbelempersia, we need the libraries17:22
persiaCan we get them with an install of kontrol-dev ?17:22
rbelempersia, yep17:23
persiaSo let's add that to build-depends :)17:23
rbelempersia, but kcontrol-dev does not exists17:24
persiaheh.  I wondered why I couldn't find it :)17:25
persiaSo, we can create kcontrol-dev, or we can copy more stuff.17:25
rbelempersia, we should add the kcontrol-dev to the kdebase-workspace debian/control17:26
persiarbelem: Well, we'd have to coordinate with all sorts of people for that.17:26
rbelempersia, :-)17:27
rbelempersia, what time is it in your timezone?17:27
persiaBut yeah, that makes sense.  It just means that we'll need to prepare a patch to expose libkcontrol and libkcontrol-dev, and then wait for acceptance.17:27
persiaSince there was just a big KDE upload today, I think it may be a bit.17:27
rbelemcool! :-)17:27
persia11 hours later than yours :)17:27
persiaBut don't worry about it.  Getting stuff done is more interesting than diurnalism.17:28
rbelempersia, 12am?17:28
persiaOh, I thought it was later there.  13 hours later than your timezone :)17:29
rbelemmine is -040017:29
persiaAh, mine is +0900, so yeah, 13.17:30
persiaSomehow I thought you were in -0300, and beause it was summer, maybe -0200.17:30
rbelempersia, in the other half of the brazil it is -0200 because of the summer17:31
persiaYou're out west?17:31
rbelempersia, i live in the heart of the amazon rain forest :-)17:32
persiaI hope that it's in a clearing :)17:32
rbelemmark already came to my city17:34
rbelemtogether with martin pool and kiko17:34
rbelemwe organized a very nice tour for then17:35
rbelempersia, i will build kdebase to get the output files to create the libkcontrol.install and libkcontrol-dev.install17:42
persiarbelem: Whilst that's building, how about uploading plasma-mobile to REVU?17:51
persia(or aren't the keyservers synced yet?)17:51
rbelempersia, ops... forgot17:51
rbelempersia, i will do the upload now17:51
persiaCool.  We ought be able to get it uploaded today then.17:52
rbelempersia, i will get a lunch18:17
rbelempersia, it is better for you get some sleep, hum?18:18
rbelempersia, i will ping you in my evening18:18
persiaOK.  Talk with you then.18:21
rbelempersia, sleep well :-)18:28
asac_rbelem: how are things going?23:40
rbelemhi asac 23:44
asacsorry... reconnect23:45
asacwell irssi restart ;)23:45
asacdid you write anything?23:46
rbelemasac, i think plasma-mobile package is ready23:46
asaci saw something blinking right when i closed the client23:46
asacis it in REVU?23:46
rbelemasac, but i'm having problems to setup my gpg key23:47
rbelemasac, i tried to create a 8192bit key23:47
rbelemasac, but the subkeys can just use 409623:48
asacis that a problem?23:48
rbelemasac, i will upload with my old key23:48
asacif the old key is good23:48
asacjust do that23:48
rbelemasac, ok23:48

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!