**** BEGIN LOGGING AT Sat Feb 27 02:59:56 2010 Feb 27 04:51:06 question: does the maemo sdk emulator support connecting to exchange 2007? Feb 27 04:52:47 buehler? Feb 27 04:54:05 lol Feb 27 04:54:49 :) Feb 27 05:00:15 hmmm Feb 27 05:00:37 no takers on the maemo sdk emulator supporting connecting to exchange 2007? Feb 27 05:12:04 asking about maemo in meego channel. son I am disappoint. Feb 27 05:17:22 pardon me, i must have forgot the meego infrastructure is 100% operational, my bad Feb 27 06:10:48 morning Feb 27 06:13:08 meh Feb 27 06:20:27 moin Feb 27 06:37:18 :) Feb 27 08:42:27 morning david! Feb 27 08:52:23 morning all Feb 27 08:52:29 hey Carsten Feb 27 10:35:27 hello everyone Feb 27 10:37:41 do you think we will have a MeeGo discussion somewhere about quality standards of UI etc.? Feb 27 10:41:16 Amby, Probably yes. But I would guess that it would be based much on http://www.forum.nokia.com/info/sw.nokia.com/id/eb8a68ba-6225-4d84-ba8f-a00e4a05ff6f/Hildon_2_2_UI_Style_Guide.html Feb 27 10:42:28 or not Feb 27 10:42:59 you don't know if the Maemo 6 UI will be anything like that Feb 27 10:43:04 in fact, not even I know that Feb 27 10:43:55 thiago, well, thats why I didn't say "it would be" but "i would guess" so yes, true I don't know. Feb 27 10:44:21 Plux: how much community involvement was there for the UI style? Feb 27 10:46:37 Amby, I'm not really sure about that for either moblin or maemo. Feb 27 10:46:47 I especially miss a stronger "key principles" part. Feb 27 10:51:46 my guess is that there won't be one MeeGo UI only Feb 27 10:51:54 there will be slightly different ones depending on the form factor Feb 27 10:52:12 thiago_home: right, I was more talking about the reference UI Feb 27 10:52:21 reference UI per form factor Feb 27 10:52:31 yes Feb 27 10:53:27 i wonder if one could use other toolkits for meego development, e.g. wx? Feb 27 10:55:43 sure Feb 27 10:55:44 it's X11 Feb 27 10:56:06 use whatever you will Feb 27 10:56:59 but don't expect it to come preinstalled in the devices, though Feb 27 10:58:53 sure. that's no problem. considering i plan to use wx (with haskell for desktop app), i'd like to write 'lite' version for mobile device using wx and not being tied to qt-only...that's all Feb 27 11:00:44 * ShadowJK is guessing anything that pulls in tens of megs of wx wouldn't be considered 'lite' by most people ;-) Feb 27 11:02:14 ShadowJK: heh, devices like N900 have lot of mem and here it is important to have 'lite' version running on mobile device anyhow Feb 27 11:02:45 is qt liter than wx? Feb 27 11:03:07 me = n00b, sorry Feb 27 11:03:44 alden: probably it is, considering that wx is a layer on top of another toolkit, e.g. gtk+ Feb 27 11:04:11 Gour: What are you working on? Feb 27 11:04:31 jrayhawk: planning to work on vedic astrology program Feb 27 11:04:50 jrayhawk: cool :) Feb 27 11:05:01 err Feb 27 11:05:05 :-D Feb 27 11:05:06 hee :) Feb 27 11:05:06 gour: cool :) Feb 27 11:05:09 it didn't "feel" light on my desktop, wonder what it'd be like on a mobile device :) Feb 27 11:06:32 ShadowJK: well, for now the only option is app running on old palm..so everything running on meego is huge plus for me Feb 27 11:06:46 ah Feb 27 11:08:07 it is interesting that before i went on holidays (end of jan), i was researching which platform to use for mobile development and was not easy to choose between maemo and moblin. now when i returned i 'discovered' that the two are merging :-) Feb 27 11:08:29 gour: :-) Feb 27 11:09:10 Gour: A friend of mine made a very simple "make a window and start mucking about directly in Cairo" Haskell library Feb 27 11:09:20 prior to that, moblin got some advantage and i was thinking to wait for some moblin-based smart-phone...now it should be easier and hopefully we'll see more meego devices Feb 27 11:09:28 you can clone http://www.aremonadscheating.org/git/bees.git/ if you want to take a look at it. Feb 27 11:09:41 jrayhawk: i plan to use wxhaskell Feb 27 11:10:32 talking like that will make poor cworth cry :( Feb 27 11:10:36 with moblin, one could just play with clutter bindings...now i wonder that will happen with clutter Feb 27 11:11:37 Intel OSTC bought the developers out, so I assume they're still interested in it. Feb 27 11:12:01 I haven't been paying attention, but I assume some of them are even in here. Feb 27 11:12:09 what will nokia say about it Feb 27 11:21:42 And, of course, someone should warn you that compiling Haskell on ARM is a bit... um... tedious, to be generous. Feb 27 11:22:33 because of glibc? Feb 27 11:23:50 I doubt it. Feb 27 11:24:28 Bootstrapping GHC6 takes around a week, last I heard. Feb 27 11:25:37 ghc is getting llvm backend... Feb 27 11:26:20 otoh, i was running ghc on moblin image without any problem Feb 27 13:39:40 hello all Feb 27 14:14:12 http://www.theviralscript.com?id=3nfv45uhihvz7ulahmir5ftd4jainb Feb 27 16:07:52 hi everybody Feb 27 19:36:24 evening qgil Feb 27 19:37:12 hi Stskeeps Feb 27 19:41:55 hey Feb 27 19:43:14 hi CosmoHill Feb 27 19:44:15 o/ Feb 27 19:45:47 "Harmattan, originally slated to become Maemo 6, the 6th incarnation of Nokia's Maemo mobile operating system, is now considered the first MeeGo release, and branded as such by the Maemo community." Feb 27 19:46:02 http://en.wikipedia.org/wiki/MeeGo Feb 27 19:46:03 i want to write an app with qml, anyone know if there will be major changes that may affect my app for meego/harmattan? Feb 27 19:46:30 How it is branded is no big deal Feb 27 19:46:47 Internally, it is still Maemo6 though, isn't it? Feb 27 19:46:49 I'm not supposed to edit the MeeGo page at wikipedia since I might have a conflict of interest but that sentence sounds wrong... Feb 27 19:47:16 harmattan isn't the first MeeGo release right? Feb 27 19:47:17 Well ask the General and he will edit it for you Feb 27 19:47:23 RST38h no "Maemo 6" used inside Nokia Feb 27 19:47:42 Was "Maemo 5" used on the inside? Feb 27 19:47:52 harmattan =! MeeGo as OS Feb 27 19:48:10 this starts sounding like a mess :) Feb 27 19:48:59 lol Feb 27 19:49:47 source: http://talk.maemo.org/showpost.php?p=529073&postcount=14 Feb 27 19:49:48 To me it sounds like they read the press release and tried to translate marketdroid->english. Lost in translation I guess :) Feb 27 19:50:06 no wonder Feb 27 19:51:18 HI Quim, did you draw any conclusions from the TMO thread about how to launch the x-terminal in Harmattan/MeeGo? Feb 27 19:51:47 so if harmattan won't be called "maemo 6" but it won't be "meego" either, what will it be called? Feb 27 19:51:56 zaheerm: no, there shouldn't be Feb 27 19:52:06 richieee72 no conclusions, we'll see Feb 27 19:52:16 ali1234: it will be "meego enabled" Feb 27 19:52:33 hmm ok, fair enough Feb 27 19:52:45 that's a term that doesn't mean "is MeeGo" but can be applied to just about anything that Nokia chooses to :-P Feb 27 19:52:48 still that doesn't sound like a name as such... Feb 27 19:53:00 it's not a name Feb 27 19:53:21 users care mostly about the name of the device Feb 27 19:53:29 how about "the operating system formerly known as maemo" Feb 27 19:53:45 ali1234: we'll suggest that :-) Feb 27 19:53:48 qgil: they do? Feb 27 19:53:52 qgil: why do you think so? Feb 27 19:54:13 qgil: especially considering that internet is basically "maemo this ... maemo that... " Feb 27 19:54:19 an important thing when naming a device is how it translates Feb 27 19:54:21 ah, forget it, it's saturday night :) Feb 27 19:54:32 from New York? Feb 27 19:54:45 i think there was a car that when you translated it from something to english spent cu... Feb 27 19:54:57 spelt* Feb 27 19:57:56 Harmattan is not MeeGo but I *personally* guess it will be fair to call it "based on MeeGo" or "MeeGo compatible" Feb 27 19:58:39 as explained in the link above, the differences will matter to a minority of users/develoers that probably will know well the deal and the details Feb 27 19:59:06 in other words: you Feb 27 19:59:21 also note that it will not be the first device either. It may be the first smartphone, though. Feb 27 19:59:23 that clears a bit of the confusion, thanks, qgil Feb 27 19:59:39 qgil: okay thanks. So long as you can still get the x-terminal some way in Harmattan/Meego, that'll be fine. Feb 27 19:59:44 the first device would be a netbook / nettop right? Feb 27 20:00:25 once there is a release out... Feb 27 20:00:43 richieeee72, and rootsh! Feb 27 20:00:43 the first device can be an existing device with an installable image for it Feb 27 20:00:56 what was the first "Debian device"? Feb 27 20:01:12 about new devices with MeeGo out of the box Feb 27 20:01:32 I don't expect the Nokia device to be the first one Feb 27 20:02:34 ShadowJK: Yeah, but I think the new security platform puts a spanner in the works with rootsh in closed mode anyway Feb 27 20:03:18 sucks :( Feb 27 20:03:45 get your facts right before concluding that something sucks Feb 27 20:04:02 ShadowJK: so long as you can still use the x-terminal at user level and SSH that'll do me. I very rarely use root on the N900 Feb 27 20:04:11 qgil, well actually we don't know anything about it yet, so :) Feb 27 20:04:18 so :) Feb 27 20:04:26 we can only make vague handwavings against the approaching evils :) Feb 27 20:05:19 I guess it depends both on how crippled it'll be in the closd mode, and how limited the closed kernel is (if you need a new one for NAT and such like on N900, or if the default 'closed' one is fine) Feb 27 20:06:13 crippled in non-closed mode I mean... if everything stops working when you flash your own kernel :) Feb 27 20:06:29 have you tried asking at http://wiki.maemo.org/Maemo_security + related Talk thread? Feb 27 20:09:29 It's going to be a lot more depressingly relevant to try to get Nokia to integrate patches. Feb 27 20:09:31 also "Company/developer" info at http://en.wikipedia.org/wiki/MeeGo feels not completely right Feb 27 20:09:33 Stupid DRM Feb 27 20:09:52 will it be possible to use the ubuntu moblin as the same/similar base for meego too Feb 27 20:09:58 MeeGo trademark and project are from the Linux Foundation Feb 27 20:10:41 Intel and Nokia are the companies putting together a huge amount of "company" and "development" but it's open to others Feb 27 20:11:05 lcuk define "use" Feb 27 20:11:14 lcuk, MeeGe is lower than Moblin Feb 27 20:11:22 well in the same way that m6 will be pretty much compatible Feb 27 20:11:30 and efine "ubuntu moblin" btw :) Feb 27 20:11:37 lcuk, when people say "Moblin", they often mean something more like "Hildon" Feb 27 20:11:39 using ubuntu moblin + the qt stuff should be the same effect Feb 27 20:11:43 lcuk: if you write everything in Qt it will work fine on any distro that ships Qt, which means any distro Feb 27 20:11:52 theres a ubuntu build with moblin ui ontop Feb 27 20:12:03 is "ubuntu mobile" supporting Qt and Web Runtime? Feb 27 20:12:43 ubuntu moblin is the moblin UI repackaged to run on top of ubuntu. ubuntu ships Qt 4.6 in the next release which will probably be out before meego Feb 27 20:12:52 ali1234, that would be the idea - but if thats the case it should be possible to use meego on n810 Feb 27 20:12:59 since we can put qt there Feb 27 20:13:20 lcuk: sure. we can put ubuntu arm or debian or any other arm supporting distro on the n810 too Feb 27 20:13:23 Er, yeah. . . . Feb 27 20:14:59 lcuk it depend on what you are looking for: MeeGo API compatibility or MeeGo UX look & feel (or both) Feb 27 20:15:07 i have no idea about web runtime, but if it is open source, and it becomes popular, you bet it will show up in all distros quickly Feb 27 20:16:46 qgil, that was sorta the point of the question, as long as meego on m6 is compatible enough it wont matter to most people and thats a good thing Feb 27 20:17:44 that's the entire point of MeeGo Feb 27 20:18:01 make everything underneath the applications common, shared and open source Feb 27 20:24:06 thiago_home, i was at the fosdem (and maemo summit) talks on security and ive read up quite a bit about them, any idea how they can work if the whole stack is open source and hence recompilable by users Feb 27 20:24:29 I don't have any idea Feb 27 20:25:31 lcuk, I think the bootloader isn't open ;p Feb 27 20:25:32 lcuk: by the apps and kernel needing to DRM not being so? Feb 27 20:25:54 there's also a security feature on the ARM chip too Feb 27 20:26:52 * lcuk nods Feb 27 20:28:33 lcuk, I was at the security talk in the maemo summit in amsterdam, they talked about a first-stage boot loader checking the crypto signature of the firmware and disabling decryption keys in the hw chip if it wasn't a blessed fw Feb 27 20:28:46 so you end up being able to do anything except decrypt DRMed content Feb 27 20:29:08 needless to say the first-stage bootloader isn't end-user-replaceable Feb 27 20:29:29 that usually isn't Feb 27 20:29:38 otherwise you could brick the device Feb 27 20:29:38 it'll be interesting to see if people will start finding bugs in it Feb 27 20:29:42 fascinating reading: http://www.xbox-linux.org/wiki/17_Mistakes_Microsoft_Made_in_the_Xbox_Security_System Feb 27 20:29:49 you need a cold-flash toolkit to do it Feb 27 20:30:15 yeah but if i can put the meego stack ontop of any os that supports qt what effective difference will lack of bootloader lead to Feb 27 20:30:20 I liked the arguments about how many people wouldn't have bothered trying to break the xbox if MS allowed them to run linux on it from the start Feb 27 20:31:08 many technical people will ignore DRMed content and so won't have a motive to start cracking the framework, as long as they can get root and install their own kernels and whatnot Feb 27 20:33:22 MeeGo Thoughts http://zchydem.enume.net/?p=622 but his comments about Ovi services are a bit confusing Feb 27 20:33:32 qgil: by curiosity, do you get emails about travel advisories to the US? Feb 27 20:33:57 i mean the article is good, but the Ovi part is confusing Feb 27 20:34:17 thiago_home yes, they are hillarious Feb 27 20:34:51 you fly to Portland in 2 weeks and you get an email warning you about a tsunami in Hawaii...... Feb 27 20:34:55 just got one about tsunami forecast Feb 27 20:35:58 hmmm Feb 27 20:35:58 I remember one warning you about protests in thed Basque country. My flight was to Gran Canaria. (2000 km away or more) Feb 27 20:36:06 * ShadowJK read the security wiki discussion and talk threads... Feb 27 20:37:09 I sure hope Ovi will be able to distinguish between closed and free maemo^Wmeego instance, and not offer things that depend on jail :D Feb 27 20:38:45 ShadowJH, I don't get you Ovi comment either (maybe it's me then) Feb 27 20:39:35 * mgedmin is kind of boycotting Ovi since it requires a username/password even for free apps Feb 27 20:40:06 hi Feb 27 20:40:26 hey Feb 27 20:40:29 ten minutes with moblin image creator and i feel like rewriting or improving it.. Feb 27 20:40:46 qgilN900, ah I mean if you don't have the closed kernel anymore (and from the comments it seems restoring would require total reflash), but software/media which do not use any of the DRM/securestorage is still supposed to work, what happens if you'd for example accidentally pay for something that turns out not to work on your device? Feb 27 20:41:02 how long does it take to download everything with moblin image creator ? Feb 27 20:41:31 In a sense there'd be two kinds of , drm-free and drm-enabled, which are slightly incompatible with eachother :) Feb 27 20:41:46 what happens if you pay for DRM content when you don't want DRM in your device, you mean? Feb 27 20:42:23 This would also be relevant for maemo.org repositories, or the future equivalent. For example the current mobile-hotspot (iirc) software would, in its current form, kill DRM on the device or just not be installable on the DRM-enabled devices Feb 27 20:42:27 and you won't re-enable the default behavior with DRM either? Feb 27 20:42:53 I got the impression "re-enabling" would be full reflash? Feb 27 20:44:02 no idea but what if you have to reflash? Feb 27 20:44:30 DRM stand for Direct Right Management or Direct rendering manager ? Feb 27 20:44:39 the former Feb 27 20:44:47 Digital Restrictions management? Feb 27 20:44:54 VLJ: both, but this discussion was the former Feb 27 20:44:59 ok Feb 27 20:45:06 qgilN900, Well then you lose whatever you downloaded anyway? ;-) Feb 27 20:45:31 Stskeeps, how would you change it for the better. is the new version for meego underway, would your notes and observations make for a clean start for the proper meego one? Feb 27 20:45:38 your purchases are recorded in the Ovi Store, so you can redownload them. Feb 27 20:45:42 ah Feb 27 20:45:51 mmm no idea in a Harmattan device but if you look the N900 flashing doesn't touch your eMMC Feb 27 20:46:18 actually, it does erase /home Feb 27 20:46:22 flashing touches installed apps tho Feb 27 20:46:24 not /home/user/MyDocs Feb 27 20:46:25 lcuk: once i'm done whining about it i'll be constructive :) Feb 27 20:46:34 Stskeeps :) Feb 27 20:46:44 no idea where your DRM media goes Feb 27 20:47:04 most people dont just get media tho, ovi store is mainly for apps atm isnt it? Feb 27 20:47:17 the talk at fosdem was good re: security framework for haramattan Feb 27 20:47:48 yeah zaheerm it seemed to flow better than the summit one, seemed to clarify a lot more than the first Feb 27 20:47:52 was more technical than the summit one Feb 27 20:47:57 so, regarding DRM, can i use it for my own evil purposes? Feb 27 20:48:12 no, the store also has other stuff, like ringtones, videos, wallpapers, etc. Feb 27 20:48:22 ali1234, she explained in the talk that user apps can use the encryption keys Feb 27 20:48:23 and was 4 months further in development Feb 27 20:48:31 sure ali1234, if its your digital media and you can wonga up enough cash for a certificate Feb 27 20:48:33 eg can i make it so that my N900 will only run code signed with my key, and not anyone elses? Feb 27 20:48:57 the security framework is for harmattan not for fremantle Feb 27 20:49:00 s/N900/what ever the new thing is called/ Feb 27 20:49:02 ali1234 meant: eg can i make it so that my what ever the new thing is called will only run code signed with my key, and not anyone elses? Feb 27 20:49:30 ali1234, thats a bit fruitless just dont share your code/apps with anyone Feb 27 20:49:38 no you dont understand at all Feb 27 20:49:48 i dont want to lock my apps to my specific device Feb 27 20:49:50 not sure how much you'll be able to modify the policies Feb 27 20:50:02 i want to prevent execution of anything i didn't sign myself Feb 27 20:50:10 at the bootloader level Feb 27 20:50:25 i doubt it Feb 27 20:50:53 sure you can as long as your write/build/compile the bootloader yourself you get to choose the encryption algo/key Feb 27 20:50:55 the bootloader will either run kernel images signed by nokia and hence allow access to the encryption keys or run any other kernel Feb 27 20:51:16 so i can't use the bootloader to boot my kernel with my keys and my DRM Feb 27 20:51:17 device won't start if you replace the loader Feb 27 20:51:32 ali1234, the keys are stored in hardware# Feb 27 20:51:54 qgilN900, it just looks to me like it'll split the devices into two categories, where some software (wizard-mounter, mobile-hotspot, joikuspot, for example) would only work on an open kernel, and presumably some other software that requires DRM features would only work on the closed kernel devices? Feb 27 20:52:17 so i can only use the DRM/encryption with nokia's keys? Feb 27 20:52:26 with your device's keys Feb 27 20:52:58 but i wont know what those keys are Feb 27 20:53:15 therefore they are nokia's Feb 27 20:53:16 but you'll be able to use them to encrypt, sign decrypt etc. Feb 27 20:53:32 but i wont be able to use them to sign my own binaries Feb 27 20:53:40 because that would totally defeat the purpose of the DRM Feb 27 20:53:48 shadowJK why are you assuming that those apps won't run with your default config? Feb 27 20:54:16 (note: I don't know them) Feb 27 20:54:20 qgilN900, maybe because they'd require kernel mods... Feb 27 20:55:07 ShadowJK, boot to task switch! Feb 27 20:55:26 can i replace the keys with my own? (i don't care that that means i can't use nokia's kernel or watch movies) Feb 27 20:55:48 ali1234, you won't be able to modify the hardware Feb 27 20:56:25 well it sounds to me like this can only be used against the user Feb 27 20:56:33 qgilN900, well the ongoing trend has been feature culling :) Feb 27 20:56:50 MeeGo has no trend Feb 27 20:56:55 as for yet Feb 27 20:58:33 I guess your questions will have easier answers after the first MeeGo and Harmattan releases including kernel and security framework Feb 27 20:58:39 Well let's hope that we can get many useful features included in MeeGo so that vendors don't get too tempted to throw out half the kernel :) Feb 27 20:59:50 qgilN900, has there been any timeline committed for that? Feb 27 20:59:54 in the times of Nokia being the sole vendor and maintainer of MeeGo, it made sense to remove tricky parts that are not very relevant to your next products Feb 27 21:00:11 btw i highly doubt ever device will have a unique key stored in hardware :) Feb 27 21:00:13 * thiago_home missed oprofile on the N900 Feb 27 21:00:31 ali1234, that was what was said at the security talk at fosdem Feb 27 21:00:43 thiago_home: have yo used perf (and sysprof) yet ? Feb 27 21:00:54 but in MeeGo there is platform development targetting many vendors, with maintenance open to whoever cares about a specific feature or component Feb 27 21:00:58 you;d need to make a unique signed upgrade for every device Feb 27 21:00:59 arjan: no, I haven't Feb 27 21:01:09 thiago_home: worth a shot.. it's actually very nice Feb 27 21:01:14 arjan: actually, I just wanted to access the performance counter registers on ARM Feb 27 21:01:18 actually nicer than oprofile by quite a bit Feb 27 21:01:27 thiago_home: that's what perf is for :) Feb 27 21:01:33 we can read the TSC in userland on x86, but not the equivalent on ARM Feb 27 21:01:40 qgilN900, will vendors be able to deviate from the meego default kernel and still call their device a meego device? Feb 27 21:02:24 that is a good question for someone sitting closer to the MeeGo kernel - I don't know Feb 27 21:02:31 ali1234, no you wouldn't the policy wiould be that public keys of nokia and other trusted parties would be there such that your device would be able to verify the signature Feb 27 21:02:45 * arjan works on meego & kernel ;-) Feb 27 21:02:51 how about fun: whats the biggest screensize we are gonna have by default for meego :D Feb 27 21:02:52 zaheerm: I suspect the answer is "it depends" Feb 27 21:03:11 hmm yeah i suppose that is is true Feb 27 21:03:12 arjan: please turn on the performance registers for me to read in userland :-P Feb 27 21:03:15 people have said for ages about variety of machines for maemo - with meego this is practical Feb 27 21:03:22 zaheerm: if you want to add some drivers etc, that's very very different than going 6 versions back and having a different ABI Feb 27 21:03:24 lcuk: counting TVOut cable? ;) Feb 27 21:03:46 qgilN900, nooo it will have to be wireless, we know of the limitations of standing near the projectors Feb 27 21:04:03 :) Feb 27 21:04:08 those cables are the only frustrating part Feb 27 21:04:17 otherwise its perfect :D Feb 27 21:04:21 touché Feb 27 21:04:23 arjan, so would device manufacturers be able to remove support for netfilter and nat etc. assuming those are in the kernel? Feb 27 21:04:29 lcuk, two n900's, one running x11vnc, the other running full-screen vncviewer Feb 27 21:04:36 zaheerm: removing is a much harder question than adding Feb 27 21:04:42 mgedmin, indeed Feb 27 21:04:42 that would probably suck over wireless Feb 27 21:04:43 zaheerm: since that means you might break app compatibility Feb 27 21:04:46 especially at a conference Feb 27 21:04:50 theres more than one way to skin a cat tho Feb 27 21:04:52 zaheerm: which would defeat half the point of meego Feb 27 21:05:00 arjan, the concern of developers would be the minimum subset of options Feb 27 21:05:10 zaheerm: why would you want to delete pieces btw? Feb 27 21:05:17 to save 20 Kb of code space? Feb 27 21:05:19 lean n mean Feb 27 21:05:24 arjan, i wouldn't but nokia or random chinese manufacturer might Feb 27 21:06:06 you missed an s Feb 27 21:06:07 I guess we're just using the N900 kernel's lack of NAT and [network] filesystems as example :) Feb 27 21:06:16 zaheerm: I don't think there's any official answer yet (this is all work in progress), but I suspect that "removing options that are userspace visible" will be rather frowned upon Feb 27 21:06:29 (while options that add, or are not userspace visible... much easier story) Feb 27 21:06:59 define removing, making into a module should be fine Feb 27 21:07:01 breaking app compatibility => evil Feb 27 21:07:11 Stskeeps: if it's application visible... ;) Feb 27 21:07:13 arjan, so then the important things are, what will be the minimum subset of enabled kernel options Feb 27 21:07:26 zaheerm: and that part has not been defined at all ;-) Feb 27 21:07:30 zaheerm I don't know Nokia but "random manufacturer" might just go for the change if really needed for their busines and simply claim the MeeGo "compatibility" or simple take the OS and sell products without calling it MeeGo Feb 27 21:07:38 I guess "application visible" only matters for applications in the base meego, too :) Feb 27 21:07:55 ShadowJK: eh... and ovi store and other stores and ... and ... Feb 27 21:08:04 qgilN900, random manufacturer that hasn't called its platform meego wouldn't be so important for devs Feb 27 21:08:06 * CosmoHill wonders if he could recompile meego for his own stuff Feb 27 21:08:59 CosmoHill: i sure hope you can :P Feb 27 21:09:02 * mgedmin wonders what "meego apps compatibility" means when there are devices using different architectures, screen sizes and input/output paradigms Feb 27 21:09:02 qgilN900, so i am guessing meego will have to certify the manufacturers' devices before they can use meego in their marketing material Feb 27 21:09:12 * CosmoHill has compiled a linux system from source files Feb 27 21:09:46 zaheerm think of Asian manufacturer selling a device without really caring much about the OS brand, or even abojt a certification Feb 27 21:10:06 qgilN900, then i as a 3rd party developer don't need to worry about them Feb 27 21:10:21 a different story goes for devices targeting EU / America Feb 27 21:10:22 oh yeah, there are some selling devices claiming "Maemo" :) Feb 27 21:10:43 zaheerm indeed, and they don't care much about you either Feb 27 21:10:46 zaheerm: anyway, general objective is to get to a situation where app writers can assume a basic set of "services", including kernel services etc Feb 27 21:11:15 if the phone vendor doesn't care about being compatible, it's open source , they can do whatever they want I suppose, but they lose the value of this compatibility Feb 27 21:11:25 Services are insufficient for portable apps Feb 27 21:11:38 qgilN900, important part is if i write some low-ish level app/library that i want all meego devices to be able to run, it should be runnable on all meego certified devices Feb 27 21:11:47 maybe one could take the default fedora (or debian) kernel, and just change the minimum amount required to make it run on a device :D Feb 27 21:12:25 why would you take fedora/ubuntu kernel over mainline? Feb 27 21:12:27 You need abstractions for inputs *and* abstractions for UI that are agnostic toward screen sizes and input paradigms Feb 27 21:12:28 arjan, can we influence the basic set of services chosen? Feb 27 21:12:34 ali1234, I mean regarding the .config Feb 27 21:12:55 Not exactly the easiest kind of design job Feb 27 21:13:01 zaheerm: kernel wise? I'm sure once we get a prerelease out you can see what we start with, and then the normal bug/feature request will work for suggesting changes Feb 27 21:13:02 Instead of, say, taking an allno config and enabling the minimum amount that makes the device and default meego userspace stack work Feb 27 21:13:22 with android, only google practically can influence it...hence the apalling multimedia on android Feb 27 21:13:30 zaheerm: I suspect that's the most efficient way of going about it; since I bet 90% or more of what you'd want will just be there already Feb 27 21:13:39 zahherm, fully agree there - I just wanted to point out that if a vendor wants to cut features A, B, C and that is more important for their business than an official MeeGo logo or fully compatible device, they can still get the whole stack, customize with the MeeGo infra a cranck a device with the resultin image Feb 27 21:13:54 but if you have something very specific concern, do speak up.. Feb 27 21:13:56 qgilN900, fair enough, that is the case with everything Feb 27 21:14:47 arjan, nothing specific sorry, i am just asking to see how open the process is Feb 27 21:15:00 * ShadowJK mentions cifs Feb 27 21:15:09 zaheerm: fair enough Feb 27 21:15:17 android, the code is free, the process is very closed Feb 27 21:15:18 zaheerm: it's going to be a rather tricky balancing act btw Feb 27 21:15:31 since one the one hand people will want every kernel thing under the sun Feb 27 21:15:41 and on the other, it needs to run on a wide range of devices Feb 27 21:15:58 but hey... if it were easy they'd not pay us to work on it Feb 27 21:16:20 arjan, understood.... Feb 27 21:16:50 arjan, qgilN900: will there be non-intel non-nokia non-LF people in any of the decision influencing/making roles for the platform? Feb 27 21:17:05 * qgilN900 starts thinking on the book he has next to his bed.... Feb 27 21:17:27 try send a patch and see if/why it gets rejected, I guess ;-9 Feb 27 21:17:28 ;-) Feb 27 21:17:32 * arjan hands qgilN900 a kindle instead ;-) Feb 27 21:17:44 * zaheerm starts thinking of porting some ugly asp code... Feb 27 21:18:59 zaheerm that depend on non-intel non-nokia non-LF people getting involved with responsibilities etc Feb 27 21:20:20 qgilN900: it would give extra credence to the platform if such people do get involved Feb 27 21:20:39 qgilN900, what book are you on atm, ive just finished another set of larry niven books and am eyeing up something different Feb 27 21:21:05 zaheerm sure, this will come mainly from 2 sides: Feb 27 21:21:20 corporate partners joining Feb 27 21:21:30 individuals joining Feb 27 21:21:33 qgilN900: in the way that it is not just corporate governance on the project Feb 27 21:21:59 for the latter I expect the usual way Feb 27 21:22:06 you start in some area Feb 27 21:22:26 or you area already The King of Mambo in certain area upstream Feb 27 21:22:44 you get a role, permissions, responsibilities... Feb 27 21:23:46 qgilN900, great, i look forward to the whole thing panning out Feb 27 21:24:07 me too, believe me :) Feb 27 21:25:01 :) the announcement excited me, now i have to keep patient to see it in action Feb 27 21:25:34 me too, believe me :)) :/ Feb 27 21:25:49 zaheerm, how is the scanning process coming, do you have the documents available from bcn long weekend? Feb 27 21:26:00 anyway, http://en.wikipedia.org/wiki/Don_Camillo 's book is waiting for me Feb 27 21:26:06 whilst you are being patient ;) Feb 27 21:26:14 lcuk, all done, i emailed you a while back with them! Feb 27 21:26:19 lcuk, all on wiki Feb 27 21:26:21 * arjan has the Don Camillo movies.. brilliant Feb 27 21:26:32 lcuk, let me dig up the link Feb 27 21:26:38 * lcuk has been run off feet, musta missedit, ill dig and look thanks Feb 27 21:26:50 Arjan you rock now even more :P good night! Feb 27 21:27:11 lcuk, did it before fosdem http://wiki.maemo.org/Video_editor Feb 27 21:27:28 heh cool stuff :D Feb 27 21:30:36 arjan, thanks for the clarifications btw Feb 27 22:24:12 wow you should check this http://bit.ly/bFi9I4 Feb 27 22:29:36 wow you should check this http://bit.ly/bFi9I4 Feb 27 22:31:32 wow you should check this http://bit.ly/bFi9I4 Feb 27 22:31:38 wow you should check this http://bit.ly/bFi9I4 Feb 27 23:11:40 wow! Feb 28 00:03:16 hm... testing btrfs on sd.. Feb 28 00:11:31 doesn't that have a SSD mode? Feb 28 00:16:03 yes Feb 28 00:16:09 I mounted it with -o ssd_spread Feb 28 00:16:18 the wiki said this is better for less expensive ssd.. Feb 28 00:17:11 I'm copying /var/spool/squid to it, it's a pretty good filesystem-on-flash torture test I've discovered Feb 28 00:17:46 ahh Feb 28 00:17:54 dmesg is full of kernel screaming Feb 28 00:18:12 hung task detection Feb 28 00:18:40 * ShadowJK ^C's the copy Feb 28 00:19:15 So, write speed is about 50 kilobytes/sec for minutes at a time, occasionally shooting up to 5 megabytes/sec for 10-20 secs Feb 28 00:20:46 hmm Feb 28 00:29:03 sounds like kernel cache/buffering issue Feb 28 00:29:24 What kind? Feb 28 00:29:28 Any hints on how to solve it? Feb 28 00:29:31 low on ram? Feb 28 00:29:52 and sd cards are not a real ssd, looking from the pc side Feb 28 00:29:59 sure Feb 28 00:30:07 But I'm looking at it from the meego side ;-) Feb 28 00:30:08 their internal controller messes everything up Feb 28 00:30:17 erm Feb 28 00:30:56 Except maybe on high end netbooks there wont be ssd Feb 28 00:31:11 well Feb 28 00:31:20 harddrives on netbooks probably, which is fine, but when you go smaller than that it's a tiny nand and a big mmc :) Feb 28 00:31:21 my laptop has a 64GB ssd Feb 28 00:31:43 but it's far from low-end Feb 28 00:32:15 damn you Feb 28 00:32:32 high end in the sense that the SSDs cost more than the netbook? :P Feb 28 00:32:45 certainly worth it though... but they don't fit in small devices Feb 28 00:32:59 if you really have a low low end ssd Feb 28 00:33:02 turn on btrfs compression mode Feb 28 00:33:13 that way it does data compression -> saves bandwidth Feb 28 00:33:27 you probably will want to balance the drive after turning on compression Feb 28 00:33:30 arjan, using a microSD actually Feb 28 00:33:35 funny thing: my laptop is 13". back when i bought it, it was considered just a normal laptop, just a bit small Feb 28 00:33:48 nowadays, people tend to call 13" laptops "netbooks" Feb 28 00:33:51 ShadowJK: well "any slow in bandwidth storage" :) Feb 28 00:34:07 arachnist: only if they also have a shitty touchpad and keyboard ;) Feb 28 00:34:34 I mounted with -o noatime,ssd_spread,compress Feb 28 00:34:39 arjan: it has a great keyboard, great trackpoint and a little bit small touchpad Feb 28 00:34:46 and yes, it's a thinkpad Feb 28 00:34:46 Running a sanity test now... Feb 28 00:35:06 18 megabyte/s sequential read, seq write varies between 4 and 8 Feb 28 00:35:24 seems typical for a sd card Feb 28 00:35:37 yes Feb 28 00:35:45 tried comparing it to ext2/journal-less ext4 or vfat? Feb 28 00:36:33 ext3 is on my list of things to try Feb 28 00:40:15 Oh this card's write speed... Kingston is wonderful.. Feb 28 00:40:22 it has an area of 600k write speed Feb 28 00:46:28 cyas Feb 28 01:59:47 oh, some of the btrfs processes use alot of PCU Feb 28 01:59:49 CPU* Feb 28 02:00:12 When write speed is at 50kbyte/s SYS cpu time is at 70% Feb 28 02:00:24 btrfs-delalloc- Feb 28 02:01:06 iowait is "only" 25% in that case :D Feb 28 02:01:27 100kb/sec with 99% iowait too though.. **** ENDING LOGGING AT Sun Feb 28 02:59:57 2010