**** BEGIN LOGGING AT Tue May 28 03:00:00 2013 May 28 09:53:12 word. May 28 10:09:06 o/ May 28 10:09:13 hi May 28 10:28:24 damn, I have a package here (librlog5) from Debian that cannot be uploaded via the Extras assistant May 28 10:28:30 the assistant does not ask for one of the two source files May 28 10:28:42 had hoped it'll ask later, but it did not May 28 10:28:56 waiting for my ssh key to get added then so I can dput May 28 10:34:27 would BTW guess this build had the very same issue: https://garage.maemo.org/pipermail/extras-cauldron-builds/2013-May/108901.html May 28 10:39:54 hahaa, MWN: "talk.maemo.org reorginaziation moves Jolla/Sailfish into a primary subforum" May 28 10:40:02 not sure if accidental, or perfect troll May 28 10:41:02 we need to reinvent MS comic chat... so it finds fitting meme images for lines where applicable... e.g. in your case it should have shown the futurama guy pic May 28 10:41:25 if there is no match at all, it'll screw up spelling and use a cat May 28 10:48:59 A wonderful Pungenday to You all. May 28 10:54:57 tadzik: help me out plz dont see the pun May 28 10:58:53 DocScrutinizer05: reorgiNAZIation May 28 10:59:41 oh May 28 10:59:48 not sure if that even qualifies as a pun, just a Fruedian slip of some sort, probably ;) May 28 11:01:25 as the author ;-) May 28 11:03:26 friggin DHL, once again "wasn't able to deliver" a package to me, so I now need to figure where to pick up my MX5500 May 28 11:05:37 asK* the author ;-) May 28 11:09:33 DocScrutinizer05: I have a friend with that mouse not keen on the scroll wheel. May 28 11:11:02 what is the cause of this: May 28 11:11:04 [2013-05-28 11:07:01] REJECTED: Can't find 'architecture' field in dsc file May 28 11:11:08 the .dsc contains the field May 28 11:11:35 I suggest compare to a working .dsc May 28 11:12:09 yes, this one has linux-any kfreebsd-any May 28 11:12:12 a working one has just any May 28 11:12:18 so that's what I am trying now May 28 11:13:49 oh lol, even better: DHL tracking has "wasn't able to deliver" but the package was in front of my door May 28 11:14:16 sixwheeledbeast: I'd be willing to buy the mouse May 28 11:14:33 He likes it I hate it :) May 28 11:14:39 sixwheeledbeast: what's his objections against the wheel? May 28 11:14:45 ooh May 28 11:15:31 DocScrutinizer05: that is nice... so, send that to the vendor and get a second item :) May 28 11:15:36 j/k May 28 11:16:02 tempting May 28 11:16:04 smooth mode has no feedback and bumpy mode isn't bumpy enough. but I am being fussy TBH. May 28 11:16:42 I like the rechargeable battery and darklaser thing works very well on any surface. May 28 11:25:01 how ould a feedback for smoothmode look like? May 28 11:25:56 lol. I mean you feel like you are not in control of it. :P May 28 11:26:28 on the too soft bumpy mode, I can completely understand. That's why in my MX Revolution during 3rd fixing I increased the load of the spring doing the bumps May 28 11:27:09 give it an aggressive swipe and you could be in Africa... May 28 11:27:44 the whole point of smooth mode is that the wheel runs at 10 RPM for 5 or even 10s and you can give it further kicks to extend the spinning period May 28 11:28:33 and you can stop the whell any time, absolutely natural May 28 11:28:50 the bumps are too far apart IMO. but as I say I am fussy about my mouse, hence why I have a M$ one. May 28 11:28:58 and it should switch back to bumpy mode as soon as you stop it May 28 11:30:51 a pity is that linux doesn't support mousebuttons >5 that reasonably May 28 11:31:27 so I have 5 buttons on that mouse which usually don't doo anything useful May 28 11:33:15 DocScrutinizer05: can a package be marked so it only builds for armel but not i386? May 28 11:33:31 encfs's dependencies are fulfilled for armel but not i386... some moron[tm] forgot to upload libfuse-dev for i386 May 28 11:34:47 probably this is the Architecture: field too, just wonder which value to use for armel... just armel, or linux-armel May 28 11:36:03 well, look for a pkg that has only one arch, check the .dsc May 28 11:36:11 ;-) May 28 11:36:18 yes, that far I already know May 28 11:36:23 I just know no package with just one arch :) May 28 11:36:43 err >>some moron[tm] forgot to upload libfuse-dev for i386<< May 28 11:37:11 and yes, I could try doing that myself, but I do not want to touch a library package others use too that early :) May 28 11:37:28 you still miss my point May 28 11:37:32 OH, I see what you were getting at May 28 11:37:45 and I thought that means the some moron[tm] just did a binary-upload May 28 11:38:12 usually there's no such thing like binary uploads May 28 11:38:20 yay, encfs is now OK May 28 11:38:25 so Architecture: armel was a good guess May 28 11:38:43 except that I am now such a "some moron" too... May 28 11:38:53 lol May 28 11:39:13 so maybe I should soon try figuring out WHY fuse-dev is architecture restricted, and possibly fix it May 28 11:39:19 lpt: check git-blame before you say "what idiot wrote that" ;) May 28 11:39:38 I mean May 28 11:39:53 I wonder how big the chain of these "some morons" is... like, if FUSE depends on something else that Nokia never provided for i386 May 28 11:40:14 why do we have a Maemo i386 repo anyway? Is this just for the SDK? May 28 11:40:44 yes May 28 11:40:50 :nond: May 28 11:41:06 s/:nond:/:nod:/ May 28 11:41:06 sixwheeledbeast meant: :nod: May 28 11:41:36 DocScrutinizer05: ok, then maybe it was just arch-restricted because FUSE stuff isn't really very useful on the SDK May 28 11:41:54 quite possible May 28 11:42:21 ~pkg May 28 11:42:21 pkg is, like, http://maemo.org/packages/ May 28 11:42:30 though I could think of useful usecases May 28 11:43:30 I see, sshfs is already in the repos... that means now with encfs, all two properly working FUSE FSes are there :) May 28 11:44:26 BTW, I still have to admit I did all this without having the SDK installed anywehre... just from my Debian server. And did notice some differences... May 28 11:44:35 why the .... did Maemo have to rename the boost packages? :) May 28 11:45:06 or rather, the unversioned metapackages are missing May 28 11:52:45 haha, another difference: for some reason, "fuse" exists on the build server but not in the extras repo. Fix: use fuse-utils as dependency May 28 11:57:04 DocScrutinizer05: damn, I broke something... https://garage.maemo.org/pipermail/extras-cauldron-builds/2013-May/108939.html this repeats now "forever". Now I know to ALWAYS bump the version :( May 28 12:00:35 if you could find a *.maemo.org page where I could try to do sth via midgard, that could help May 28 12:01:27 ~tm0 May 28 12:01:27 well, tm0 is trolls, mooses and orangutans May 28 12:01:39 troll.maemo.org May 28 12:02:32 I _might_ have a way to fix it... but all I can do is scp May 28 12:04:17 https://garage.maemo.org/extras-assistant/index.php - anyway, the first of the two encfs packages is to be deleted May 28 12:05:57 exists(path) May 28 12:05:59 Test whether a path exists. Returns False for broken symbolic links May 28 12:06:01 wonder if I can fix it knowing this May 28 12:06:24 err, what? https://garage.maemo.org/extras-assistant/index.php doesn't yield anything useful for me May 28 12:06:56 no, can not May 28 12:07:11 fremantle build queue May 28 12:07:20 the top one of the two, encfs_1.7.4-2.4, is to be cancelled May 28 12:07:24 encfs_1.7.4-2.4maemo1 will work May 28 12:08:02 alternatively, deleting /mnt/incoming-builder/fremantle/rejected/encfs* on the build server will also fix the loop May 28 12:08:35 i can't do that (because scp cannot copy symlinks :P) May 28 12:08:46 sorry for that May 28 12:10:32 same happened on Apr 28 before :) May 28 12:15:00 http://privatepaste.com/0ab75a7383 have fun! May 28 12:16:35 what does this tell me? May 28 12:16:44 I hope more than me May 28 12:17:34 it just doesn't match the paths I dput to May 28 12:18:14 deleting scratchbox/users/builder1/home/builder1/maemo-fremantle-armel-extras-devel/work/encfs_1.7.4-2.4* should be safe though May 28 12:18:27 but I can't say this will fix it for good... is that the only machine involved with building? May 28 12:19:56 what looks especially odd is that none of this found the files with maemo1 in the name (the new attempt I uploaded with non conflicting version) May 28 12:20:03 so you may be on the wrong server May 28 12:21:17 if this helps, /mnt/incoming-builder/fremantle/rejected/encfs_1.7.4-2.4.dsc is the path that conflicts May 28 12:31:16 you're dputting to drop.m.o which has a NFS mount to bulder.m.o aiui May 28 12:31:51 sounds right May 28 12:31:53 that far May 28 12:32:11 but if you hit the right mount, you also should see files named after encfs_1.7.4-2.4maemo1 May 28 12:32:15 which your "find" output did not show May 28 12:33:15 http://wiki.maemo.org/Migrating_to_Community-driven_Infrastructure#Autobuilder_and_friends if you bother to investigate/learn. I'm out, got pressing stuff to do and no idea about that builder stuff anyway. I'll just have a look tonight (now+6h+) and muse about the pieces of the explosion May 28 12:33:30 ok May 28 12:33:36 just one last thing to do... May 28 12:34:53 ah, so the solution is to find the files on garage, not builder1, and delete the encfs_1.7.4-2.4 files May 28 12:35:16 personally I can wait, but I feel a bit sorry for people who have the mailing list of cauldron subscribed... May 28 13:01:57 anyway, while I am breaking stuff: May 28 13:01:59 dbus-send --print-reply --system --dest=com.nokia.icd2 /com/nokia/icd2 com.nokia.icd2.disconnect_req uint32:0x8000 May 28 13:02:04 how to extend this to disconnect GPRS only, but not wifi? May 28 13:04:08 I tried adding the args string: uint32:0 string: string:GPRS uint32:0 array:byte: May 28 13:05:21 the method does exist with the signature, from dbus-send's reply, but the command does nothing then May 28 13:07:31 have you tried: string:" " uint32:0 May 28 13:09:25 a space? funny May 28 13:10:36 no, does not help May 28 13:11:24 I thought IAP_ID for GPRS is a space? May 28 13:12:13 it is? trying that then May 28 13:12:29 no, does not help either May 28 13:14:30 from the state_sig I get when connecting, the parameters are "", 0, "", "GPRS", 83886080, then "Sunrise Internet" encoded as array:byte: May 28 13:14:50 thing is, I'd prefer my code to be generic and not limited to this one connection name May 28 13:17:23 even with that field filled in propery, the disconnect_req does not work May 28 13:17:51 wait, no, it works then... May 28 13:17:57 I still had GPRS replaced by a space May 28 13:18:07 so it does work if I fill in the whole service name... which I don't want to if poissible May 28 13:20:11 the docs say "network ID or empty string", but neither array:byte:0 nor array:byte: work May 28 13:21:16 Also, dbus-send does not permit empty containers or nested containers May 28 13:21:18 (e.g. arrays of variants). May 28 13:21:21 argh, that may be it :( May 28 13:27:46 ok, have it working now... now screen locking disconnects GPRS, just like I want it May 28 13:27:53 only issue, I have to hardcode the name of the provider :( May 28 13:28:36 grep the IAP first? May 28 13:44:19 sixwheeledbeast: yes, that would be the idea May 28 13:44:31 only catcdh is, it is encoded as byte array, not string... not very convenient to grep from a shell script May 28 13:44:46 wonder if the old icd_ui interface can distinguish the two May 28 13:44:53 it is undocuemnted though **** BEGIN LOGGING AT Tue May 28 14:16:22 2013 May 28 14:21:23 What are you trying to do? May 28 15:01:15 DocScrutinizer05, freemangordon: ok, then back from the other channel - a .dsc may refer to "anything", but dpkg-source created packages (and all are, Debian guys say it's even impossible to create a valid .dsc without going through dpkg-source (some way or another) May 28 15:01:29 have a specific requirement on the file name of the source tarball May 28 15:01:49 divVerent: sure May 28 15:02:55 however, none of this has anything to do with the encfs package, which is by all standards a valid source package May 28 15:03:16 the only "weird" things I caused by not going through the Maemo SDK are wrong dependency package names in the previous version, causing builds to fail (as expected) May 28 15:03:37 and in one case, I depended on Debian's "fuse" instead of Maemo's "fuse-utils", whcih caused the build to succeed but the package to not be installable (makes no sense IMHO) May 28 15:04:13 I would even claim that the SDK would not have caught this error either, as the build DID succeed after all May 28 15:04:41 I suspect rather that the SDK's environment has a package called "fuse", while the extras repo does not+ May 28 15:06:06 divVerent: could be May 28 15:06:17 or there is some package which "provides:" May 28 15:06:23 yes, or that May 28 15:06:47 although shouldn't apt then have figured that out on apt-get install encfs (with the previous broken version)? May 28 15:07:13 only if you have that repo enabled on the device, sdk and tools repos are not even installed by default May 28 15:07:17 yo fmg. May 28 15:07:20 vi__: yo May 28 15:07:23 of course... May 28 15:07:32 and AFAIK the sdk repo doesn't belon gon the device at all May 28 15:07:38 divVerent: try to install that package in SB May 28 15:07:48 in a few weeks, when I will actually be able to run SB May 28 15:07:52 divVerent: oh, there is no problem to enable it ;) May 28 15:07:54 I currently am on my eeepc 701 May 28 15:07:57 still? May 28 15:08:00 oh, shit May 28 15:08:01 yes, still May 28 15:08:25 I wonder how hard it is to install SB on a server BTW... and if there is any documentation on that May 28 15:08:30 I do have a server running Debian with free space May 28 15:08:59 If you want to experience a new kind of hell in setting up cross compilation environments, have a go at bitbake. May 28 15:09:02 ~bitbake May 28 15:09:03 extra, extra, read all about it, bitbake is a simple tool for the execution of tasks. It is derived from Portage, which is the package management system used by the Gentoo Linux distribution. It is most commonly used to build packages, and is used as the basis of the OpenEmbedded project. See http://developer.berlios.de/projects/bitbake/ and http://openembedded.org/. May 28 15:09:31 well, I'd rather want the official SDK running May 28 15:09:40 berlios... reeks of shily May 28 15:09:42 *schily May 28 15:10:34 divVerent: hmm, i am not sure SB cares whether it is server or not :) May 28 15:10:46 oh, it cares a lot :) May 28 15:10:48 vi__: what's up, besides oil prices? May 28 15:10:51 if it's a GUI machine, I can just use the VM imaghe May 28 15:10:54 that's how it cares :) May 28 15:11:07 divVerent: you don;t need vmware to install sb May 28 15:11:31 Getting a new job. May 28 15:11:32 not REALLY May 28 15:11:39 bjut I haven't figured out how to install sb normally May 28 15:11:41 http://wiki.maemo.org/Documentation/Maemo_5_Final_SDK_Installation#Installing_Maemo_5_SDK_using_Text-based_Installer May 28 15:11:44 is kinda useless there May 28 15:11:48 divVerent: why? May 28 15:12:00 for starters, http://www.scratchbox.org/download/scratchbox-hathor/ is a dead link May 28 15:12:05 I have passed telephone interview, first stage interview, technical questions test. May 28 15:12:06 i used that to install under ubuntu 64bit a few weeks ago May 28 15:12:17 vi__: nice, wish you luck May 28 15:12:23 This is enough to earn me the right to send them my references. May 28 15:12:36 This will perhaps lead to the 2nd interview. May 28 15:12:47 ~rolleyes May 28 15:12:52 divVerent: comeon, those are about 10 .debs, wget them from skeiron May 28 15:13:09 ~mirrors May 28 15:13:09 it has been said that mirror is http://maemo-archive.wedrop.it/ http://talk.maemo.org/showthread.php?p=1315143#post1315143 or extras-devel.merlin1991.at - for fighting hashsum error, or see ~rmo-new May 28 15:13:09 freemangordon: hehe, if they're new enough that'll do then May 28 15:13:46 divVerent: http://skeiron.org/scratchbox/ May 28 15:14:05 To be fair the technical questions was a joke. Stupid stuff like 'what is the difference between a linear and a switching regulator. WHat are the pros/cons of each?'# May 28 15:14:37 freemangordon: admittedly, this doesn't tell MUCH... like, what is a "hathor" May 28 15:14:50 meh, sb.m.o is supposed to work May 28 15:14:57 vi__: it is my english asking, what "linear" regulator is supposed to be? May 28 15:15:10 regulating what? current? May 28 15:15:22 think lm7805. May 28 15:15:25 divVerent: hathor is a version May 28 15:15:27 ooh May 28 15:15:31 Generally a voltage regulator. May 28 15:15:32 vi__: ^^ May 28 15:15:36 got it :D May 28 15:16:10 thedead1440 is scratchbox.m.o maintainer May 28 15:16:24 if there's sth missing or not working, ask him May 28 15:16:50 What are you working on now FMG? May 28 15:17:01 maybe booting that VM broke (or rather not established) some stuff that's needed May 28 15:17:02 damn robots.txt :( May 28 15:17:47 ah, -e "Robots = off" did the trick May 28 15:17:57 vi__: relaxing on embedlite/gecko, still waiting for the new CSSU-x updates to come out so we can continue the work on kernel-cssu May 28 15:17:58 to be able to download all the .debs, as it is not an apt source May 28 15:18:10 divVerent: yep May 28 15:18:30 thedead1440: seems like scratchbox.m.o is down May 28 15:18:52 down?? May 28 15:18:57 404 May 28 15:19:27 /var/lib/dpkg/info/scratchbox-core.postinst: 13: [: configure: unexpected operator May 28 15:19:29 not sure if I like that May 28 15:19:32 well, it lacks a proper frontpage May 28 15:19:47 "Firefox can't find the server at scratchbox.maemo.org." May 28 15:19:49 argh, it's a bash script using #!/bin/sh May 28 15:20:37 meh, sorry, http://www.scratchbox.org/ May 28 15:20:43 oh :) May 28 15:20:50 ~scratchbox May 28 15:20:50 i heard scratchbox is a cross-compiling system that uses binfmt_misc, rpc calls, and an nfs mount to make a cross-build appear to be 100% native, and is found at http://www.scratchbox.org/, or hosted by maemo.org now, maintainer thedead1440 May 28 15:21:43 divVerent: add this site to sources.list or wherever and follow the instructions May 28 15:21:57 thedead1440: false alarm, sorry :) May 28 15:22:03 lol loving the text based home page. May 28 15:22:14 http://www.scratchbox.org/debian/dists/hathor/ May 28 15:22:26 Another question they asked was: You need a 300MHz clock. Of the oscillators available you can choose 1: 100pS rise and fall. 2: 30pS rise, 30pS fall. 3: 20pS rise, 400pS fall. May 28 15:22:30 forget it May 28 15:22:32 Which do you choose and why? May 28 15:22:32 root@v220120211684756499:~# sudo /scratchbox/sbin/sbox_adduser rpolzer May 28 15:22:34 /scratchbox/sbin/sbox_adduser: 4: /scratchbox/etc/question_func.sh: Syntax error: "(" unexpected May 28 15:22:47 clearly not made for wheezy May 28 15:22:51 maybe we need a symlink or a rewrite rule there? May 28 15:22:56 (as you see, AGAIN a wrong #!/bin/sh) May 28 15:23:43 vi__: 30ps one? May 28 15:23:52 no. May 28 15:24:00 hmm, why not? May 28 15:24:04 You should not use a clock that is faster than you need. May 28 15:24:29 Fast edges mean interference. May 28 15:24:36 yep May 28 15:24:58 ~vi_ May 28 15:25:00 :( May 28 15:25:07 but that you usually handle by other means May 28 15:25:08 * freemangordon feels stupid May 28 15:25:13 ~vi__ May 28 15:25:13 it has been said that vi__ is 13:21 < DocScrutinizer05> vi__: you're absolutely right May 28 15:25:20 :D May 28 15:26:08 however faster slopes usually also mean stronger driver stage that uses more power May 28 15:26:19 and that's the true criterion May 28 15:26:37 Quick question - I was reading some threads about cssu thumb2; and it also mentioned a newer pulse version May 28 15:26:40 rise/fall time itself can get handled by RC-filters etc May 28 15:26:42 no, it doesn't work on wheezy... I'll use the VM May 28 15:26:45 I did not know that, however it make sense. May 28 15:26:49 Does it mean that I can use the old n900 pulse with my new pulse? May 28 15:27:02 Could find the exact details of the change May 28 15:27:13 Wizzup: it is "new" as "some stuff changed and recompiled" May 28 15:27:23 well, I think the ``fix'' only needed a small change May 28 15:27:40 (literally 16 -> 15) May 28 15:28:28 Wizzup: http://gitorious.org/community-ssu/pulseaudio/commits/master May 28 15:28:44 also, vm.mmap_min_addr = 0 is really something I refuse to set on a server connected to the internet... more reason to only use a VM for SB May 28 15:28:51 found it May 28 15:28:59 https://gitorious.org/community-ssu/pulseaudio/commit/5b6b4843defa6e7778fd2d73653ec7e54410a207 May 28 15:29:02 freemangordon+= May 28 15:29:03 ++ * :-) May 28 15:29:33 FYI, what failed just now was that it didn't detect the armel support somehow installed via qemu when installing the SDK May 28 15:29:47 even though this binfmt stuff was mentioned before without error during the installation of scratchbox May 28 15:30:11 now I will need to install pulse on my netbook again ;-) May 28 15:30:27 Wizzup: there are more changes, but yes, this fiz allows you to use n900 pulse with your desktop pulse May 28 15:30:37 *fix May 28 15:30:44 very nice May 28 15:30:48 PulseRoll'd - like RickRoll'd, but with total silence May 28 15:30:58 or rather ... excellent *rubs hands* May 28 15:31:38 Wizzup: but this is not specific to cssu-thumb, mind you, cssu-t has (or at least is supposed to) the same pulse May 28 15:31:46 yeah, I figured May 28 15:32:00 I also don't think I'll try thumb2 just yet - I already have the other cssu installed :-) May 28 15:32:14 now to figure out how to use garage to make my first maemo .deb May 28 15:32:18 * Wizzup back to reading wiki May 28 15:32:24 have to admit, the N900 is the first and only device I have on which pulseaudio actually works :) May 28 15:36:23 N900: One of God's own prototypes. A high-powered mutant of some kind never even considered for mass production. Too weird to live, and too rare to die. May 28 15:37:24 BTW, I recently saw in a shop a cheap Nokia supporting "Ovi Store" and having an actual keyboard; case design was quite Blackberry-cloned, but white May 28 15:37:30 what does run on it? May 28 15:38:40 Probably S40. May 28 15:38:47 A stripped out symbian variant. May 28 15:39:03 Sorry, castrated. Not stripped out. May 28 15:39:38 Symbian. No thanks :) May 28 15:39:51 Symbian was the best mobile phone OS ever made. May 28 15:40:15 focus on "phone", then yes May 28 15:40:19 No other phone OS was as feature complete. May 28 15:40:26 Yes, phone OS. May 28 15:40:42 (as opposed to smartphone) May 28 15:40:46 ... May 28 15:40:50 it is not a S40, though May 28 15:40:52 WHat makes a phone a smartphone? May 28 15:40:54 as it has a qwertz keyboard May 28 15:41:02 good question :) May 28 15:41:38 EPOC had *REAL* multitasking 15 years ago. May 28 15:41:49 oh, S40 is a quite broad series May 28 15:41:50 The modern devices STILL do not have it. May 28 15:42:07 N900 has it. Android has it. Even iOS has it. May 28 15:42:14 divVerent: C3? E5? May 28 15:42:25 Symbian handled SIP, VPN etc beautifully. May 28 15:42:38 could do the same on iOS, though May 28 15:42:46 using extra apps, of course May 28 15:42:47 E71? May 28 15:43:12 divVerent: I respectfully disagree. Neither IOS nor droid have multi-tasking. May 28 15:43:36 so Linux has no multitasking? that's news to me May 28 15:43:42 eh? May 28 15:43:55 Android has no trouble at all running a process in background May 28 15:43:58 divVerent: have you used andriod? May 28 15:44:06 yes, and? May 28 15:44:13 okay, I have no questions then May 28 15:44:19 how long an app "survives" when inactive depends on said app May 28 15:44:23 iOS has a quite hard 10 minutes limit May 28 15:44:34 I didn't know what this "real multitasking" circlejerk is about until I moved to android for a few months May 28 15:44:36 but even that's only enforced by policy May 28 15:44:55 or what do you MEAN by multitasking? May 28 15:45:08 running two apps at the same time on the same screen? THAT is what iOS/Android don't do May 28 15:45:18 they both can be active, but only one can be displayed at the same time May 28 15:45:21 from my point of view, it makes a big difference whether I close an app or put it into background May 28 15:45:29 if the system is trying to guess for me, that's a no-go May 28 15:45:30 same for me May 28 15:45:36 ah, I see May 28 15:45:42 in case of Android, it's all up to the app May 28 15:45:42 that's why using android was fucking annoying May 28 15:45:46 how it behaves when "backgrounded" May 28 15:45:57 and on iOS, there is this "you can not live in background for more than 10 minutes" policy May 28 15:45:59 hate that May 28 15:46:03 divVerent: please tell me what you're missing exactly on scratchbox.org May 28 15:46:05 well, the system may kill it or not, depending on memory usage (I think) May 28 15:46:05 when not following it, app is kicked from the store May 28 15:46:17 DocScrutinizer05: on scratchbox.org domain name, nothing May 28 15:46:25 eh? May 28 15:46:39 on android, most apps cannot be backgrounded because google says they should not go to background but rather exit May 28 15:46:43 to save battery life[tm] May 28 15:46:52 but an app can still do whatever it wants to, if it makes sense May 28 15:46:53 ™ May 28 15:46:59 so, summing it up: real multitasking is whether I can jump from one app to another without having to worry if one will mysteriously disappear May 28 15:47:08 (for me, that is) May 28 15:47:09 perhaps supporting ARM power domains would do better for battery life? May 28 15:47:12 iOS's solution to that is actually somewhat neat May 28 15:47:24 an app SHOULD not behave any different whether it was closed or just backgrounded May 28 15:47:32 i.e. it is supposed to save full state when exiting May 28 15:47:39 . o O ( perhaps not having an OS check for updates few times during the night would do even better ) May 28 15:47:48 ok then, when you don't miss anything on http://www.scratchbox.org/download/files/ then why did I bother? May 28 15:47:57 But what if I want to not be looking at a program while it does something? May 28 15:48:05 oh yes May 28 15:48:17 like, I want to lock my phone when it's still looking for a GPS signal May 28 15:48:19 DocScrutinizer05: there is a dead link in the wiki, but it was not to scratchbox.org May 28 15:48:31 tadzik: I fully agree there May 28 15:48:32 android then thinks "oh, so you're not going to use it then..." and stops trying May 28 15:48:34 this iOS policy is nonsense May 28 15:48:38 but it is NOT an OS restriction... May 28 15:48:44 it is merely policy in case of iOS May 28 15:48:52 I never used iOS, I'm ranting about android because I used it, and it's fucking awful May 28 15:48:52 and in Android, there isn#t even that policy, there is just apps not doing what you want May 28 15:48:59 divVerent: was not to scratchbox.org??? May 28 15:49:09 http://www.scratchbox.org/download/scratchbox-hathor/ ??? May 28 15:49:16 DocScrutinizer05: oh, it was/is May 28 15:49:18 http://www.scratchbox.org/download/scratchbox-hathor/ is dead, yes May 28 15:49:30 :-S May 28 15:49:30 http://www.scratchbox.org/documentation/user/scratchbox-1.0/html/installdoc.html May 28 15:49:33 these two are dead May 28 15:50:37 tadzik: apparently, the real problem is that both Android and iOS lack a GOOD way for the GUI to handle closed vs backgrounded apps May 28 15:50:54 in iOS it's meant so you never notice whether an app is still running or not May 28 15:50:59 if you ever figure out, I might create symlinks so you find the stuff wehere you expect it. Or you fix the wiki. Or best you check what the jobs do May 28 15:51:06 #even when it's been closed, it still appears in the list of "previously running apps" May 28 15:51:08 afk, bbl May 28 15:51:18 tadzik: I want to have a music player paused, a pdf open, my email checking for updates, a web page open with a half filled in form WHILE I am on a phone call that is being recorded by some program. Can android do this? May 28 15:51:29 technically yes May 28 15:51:36 it's up to the apps you use for the other things May 28 15:51:39 vi__: I don't suppose so. There's a reason I ditched it and bought another n900 :) May 28 15:51:53 music player - why do you even care whether it runs or not when it's paused? May 28 15:52:00 all you really care for is that you can "unpause" at any time May 28 15:52:14 maybe it's downloading lyrics? May 28 15:52:16 divVerent: if you know what pages on the wiki either fix or let me know so we can repair it. May 28 15:52:22 sixwheeledbeast: yes May 28 15:52:28 for that all I'd need is the proper URLs May 28 15:52:35 if I find them, I can update the wiki May 28 15:52:38 Because it needs to be open and waiting for me to unpause it. I do not want to 'open' it again. May 28 15:52:48 vi__: you have to switch to it though May 28 15:52:56 Nope. May 28 15:52:56 and the OS could then use that to transparently open it again May 28 15:52:59 this is actually how iOS works May 28 15:53:08 how do you unpause without first activating the app? May 28 15:53:24 The app is monitoring dbus or some such for a keypress. May 28 15:53:25 widget, keyboard shortcut, whatever means May 28 15:53:27 on iOS, there is "global media controls". And Apple did this bad indeed. May 28 15:53:42 vi__: ok, then the OS should then monitor dbus in the app's place, and start and control the app again if it was closed :) May 28 15:53:48 this is how Apple WANTED to do it on iOS, but didn't finish it May 28 15:53:50 On the n900, I enjoy using the headset button to play/pause. May 28 15:54:26 web page with half filled form - iOS does this part right May 28 15:54:35 anyway. I do not want this discussion to turn into an OS bashing. May 28 15:54:35 if iOS does decide to close the browser, the state of the forms is saved with it May 28 15:54:48 just saying... yes, we all want REAL multitasking May 28 15:54:54 and all mobile OSes have it May 28 15:55:06 but they "prefer" closing apps after inactivity to save memory/CPU/battery May 28 15:55:14 we have to agree that we don't agree with each other May 28 15:55:19 and Android is indeed prone to do this wrong in many places May 28 15:55:20 I disagree! May 28 15:55:25 when an app is waiting for IO it's not wasting cpu and battery anyway May 28 15:55:31 tadzik: I agree there May 28 15:55:34 it still wastes RAM May 28 15:55:35 Yup. Thread is sleep. May 28 15:55:40 it can be swapped May 28 15:55:44 Nope, it is swapped out! May 28 15:55:48 tadzik: you know what we both would REALLY want? May 28 15:55:56 yes. I want N901 May 28 15:55:56 user controls for controlling auto close inactive app behaviour May 28 15:55:59 that too May 28 15:56:00 fuck no May 28 15:56:12 you can turn it off entirely May 28 15:56:13 I don't want the system to guess and close apps when the system thinks it's right May 28 15:56:15 and I can set it per app May 28 15:56:20 that's why I said user controls May 28 15:56:25 because it's always wrong and confusing to me May 28 15:56:28 maybe it's just me May 28 15:56:33 sure, then you can turn it off May 28 15:56:36 in an "ideal OS" May 28 15:56:56 thing is, if the OS and the apps were done right, you'd never know if apps were closed or not May 28 15:57:04 only difference would be longer startup time when it WAS closed May 28 15:57:10 well, I don't care for any auto-close behaviour, I don't want it anywhere near me. That's it May 28 15:57:15 sure May 28 15:57:19 and you should also have that option May 28 15:57:25 I cannot think of any application I would want to die in the background without my request. May 28 15:57:34 what makes the situation on both iOS and Android bad is that it's enforced not by the OS, but by policy May 28 15:57:45 so the OS can't offer user controls for it if it wanted to May 28 15:57:55 each app has its own logic for it May 28 15:58:00 Except nokia cherry. I would like that to die, alone, in the backround with no eulogy. May 28 15:58:19 when OS decides when to close apps, the thing which shows me "running applications" stops making sens May 28 15:58:36 "what apps are running" is a valueable information to me May 28 15:58:44 "oh, I kept this one open, there's probably something I should do about it" May 28 15:59:07 "recently ran apps" carries no valuable information May 28 15:59:38 of course, you can argue that the OS could distinguish between the apps I closed and the apps OS closed May 28 16:00:02 (( and I can set it per app)) false approach: the app eventually goes on select(), wait() or whatever. All it needs is the app doing this in a reasonable way. Alas many apps user usleep() or timers to wake up every 100ms, for doing nothing May 28 16:00:17 but this is more and more becoming an OS where it doesn't close anything on itself because that'll confuse the user May 28 16:00:19 divVerent: http://wiki.maemo.org/Maemo_training_material/Getting_Started/Installing_the_SDK This page? May 28 16:01:11 weren't there ubuntu VMs with scratchbox already set up? May 28 16:01:22 on a properly designed system wirh proper apps, usually everything is sleeping, and thus you don't need to stop any task and make OS monitor keypresses to wake the task again, since that's exactly what select() *does* May 28 16:01:34 *nod* May 28 16:01:50 homework: open htop, see how many tasks are there and how many of them are running May 28 16:02:21 ~1d6 May 28 16:02:22 You roll a 5 May 28 16:02:22 here I have 144(360 threads), 1 running (probably htop) May 28 16:02:30 Critical hit!1 May 28 16:02:37 the rest of them is wasting neither cpu, nor ram, nor battery May 28 16:03:21 ~dice 1d5 May 28 16:03:23 rolled 4 = 4 May 28 16:04:15 and this page http://wiki.maemo.org/Documentation/Maemo_5_Final_SDK_Installation May 28 16:04:21 tadzik: on a one-core system, usually it's always only top that's running May 28 16:04:37 since no other task can be in state running while top runs May 28 16:05:20 "running" means: has CPU and executes commands on it May 28 16:05:27 oh, true. It's running, not ready May 28 16:05:39 or whatever's the name for ready May 28 16:06:05 NFC what the links should be changed to but that's all I can locate with the broken URL's May 28 16:08:35 I think http://www.scratchbox.org/download/ might need to point to http://www.scratchbox.org/download/files/sbox-repository/debian/dists/ May 28 16:08:48 not sure though May 28 16:09:20 thedead1440:^^^^? May 28 16:10:40 check http://www.scratchbox.org/download/files/ which is about all we have May 28 16:12:11 http://privatepaste.com/852f21565f is what we inherited May 28 16:16:18 we have an incredible mess of data from a backup on /mnt/store: May 28 16:16:44 http://privatepaste.com/981650afd6 May 28 16:17:24 /var/lib/files is a symlink to /mnt/store/var/lib/file May 28 16:17:52 lrwxrwxrwx 1 root root 24 Apr 5 15:21 /var/lib/files -> /mnt/store/var/lib/files/ May 28 16:18:12 wiki, mail, etc is not set up May 28 16:20:11 http://privatepaste.com/e3443fe624 May 28 16:22:05 http://privatepaste.com/ab3def2be2 May 28 16:28:16 sixwheeledbeast: if you want to search for a certain file: http://www.scratchbox.org/download/files/ls.txt May 28 16:28:58 tadzik: auto closing apps can still save a lotm of ram May 28 16:29:09 arguably I would rather just have enough ram May 28 16:44:03 root@scratchbox:/mnt/store/usr/share/moin# du -hs . May 28 16:44:04 279M . May 28 16:44:17 could copy those to the real system May 28 16:52:59 done May 28 18:14:38 DocScrutinizer05: all ok with sb now? another alternative is http://www.scratchbox.org/debian/dists/ May 28 18:14:52 they are all symlinks May 28 18:54:22 thedead1440: I didn't change any symlinks so far May 28 18:54:31 ok May 28 18:54:51 anyway you have posted the file list; it should be enough for anyone who requires further digging May 28 18:55:21 in apache conf in etc (vhost file) there's a lot of aliases May 28 18:55:39 I think this needs a major review May 28 18:56:30 the problem is we need to investigate all the domains all the different scratchboxes use May 28 20:01:27 thedead1440: divVerent: /var/lib/repos/www/_darcs/current/www/download/scratchbox-hathor/index.html /var/lib/repos/www/www/download/scratchbox-hathor/index.html May 28 20:07:01 thedead1440: Alias /debian /var/lib/files/sbox-repository/debian; Alias /download/files /var/lib/files May 28 20:09:57 http://www.scratchbox.org/repos/www/www/ May 28 20:11:30 http://www.scratchbox.org/repos/www/www/download/ May 28 20:15:28 ok, I think we need an alias (http://www.scratchbox.org/)download/ -> /mnt/store/var/lib/repos/www/www/download/ May 28 20:33:01 I changed document root dir /var/www/ to: May 28 20:33:04 root@scratchbox:/var# ll -d www May 28 20:33:05 lrwxrwxrwx 1 root root 32 May 28 22:28 www -> /mnt/store/var/lib/repos/www/www/ May 28 20:34:37 so far to me it seems as if this has done "the trick" :-D May 28 20:36:48 divVerent: sixwheeled: thedead1440: http://www.scratchbox.org/download/scratchbox-hathor/ and http://www.scratchbox.org/documentation/user/scratchbox-1.0/html/installdoc.html work, as does every other link I checked so far May 28 20:40:36 hello, I want buy an USB (TV) adapter with composite video input for my N900 to see the video signal of some cameras with broken displays... have anyone an idea for this? maybe I can install tvtime on maemo or use mplayer... May 28 20:41:30 the player should be the smaller problem, the driver would be more interesting May 28 20:42:05 note that many cheap usb video input devices are not supported under linux AT ALL May 28 20:42:50 I had a pinnacle tv-card and watched tv on linux... May 28 20:43:03 but it was PCI... May 28 20:43:18 NeutrinoPower: In principle, you can use a USB video capture card. May 28 20:43:26 NeutrinoPower: I'm not aware that anyones done it. May 28 20:43:56 xawtv and mplayer can play /dev/video* devices May 28 20:43:58 i said _cheap_ _usb_ May 28 20:44:08 you can buy video dongles for $5 on ebay, those don't work May 28 20:44:15 I failed to get any such USB TV stick to work on linux so far May 28 20:44:19 There are ones with source drivers. May 28 20:44:41 ok, general notice: scratchbox.org got fixed May 28 20:45:11 please check what (if anything, except wiki and bugzilla) might be missing May 28 20:58:27 http://linuxtv.org/wiki/index.php/Hauppauge_USB-Live-2 May 28 21:19:03 Easy CAP Video-Grapper are supperted by kernel 2.6.38 is it an problem to run this kernel version on N900? May 28 21:25:58 NeutrinoPower: i'd consider it non-trivial to build a custom kernel for n900, specifically a different version than already supported... but i haven't tried, and others might know better May 28 21:35:39 almost impossible, since several drivers like the graphics driver/lib are closed blobs that won't work under newer kernels, unless you port the old API to the new kernel May 29 00:00:58 freemangordon_: just tried with stable cssu; pulse still gives protocol errors on the other side May 29 00:01:13 I think... May 29 00:01:18 Well, off to bed first May 29 01:15:04 ah May 29 01:15:10 I'm on testing even **** ENDING LOGGING AT Wed May 29 02:59:58 2013