**** BEGIN LOGGING AT Sat Mar 08 02:59:59 2014 Mar 08 09:55:00 What's the difference between hildon-initscripts 1.26-1+dp1 and 1.26-1+0m5? Mar 08 09:55:25 What does "dp1" mean? Mar 08 10:07:56 since it only available in SDK repository, probably "development something" Mar 08 10:09:07 You can download both of them and compare Mar 08 10:09:14 I hove no time currently Mar 08 10:09:19 have even Mar 08 10:34:49 Hmm .. wonder how MicroB gets reasonable performance. Mar 08 10:35:39 running conkeror with the same xulrunner (assuming it actually uses a xulrunner equivalent to the one in /usr/lib/microb-engine/xulrunner), doesn't scroll as smoothly etc Mar 08 10:35:51 maybe it's the default configuration. Mar 08 10:37:26 I think it's just that the usual MicroB removes images that are offscreen somehow. Mar 08 10:38:04 scrolling is actually relatively consistent (relative to Fennec) when all the images on the Youtube homepage are loaded, but it's just a low fps Mar 08 10:49:14 hm, bugzilla activit :) Mar 08 10:50:18 "microb.suspend_images" .. sounds useful, not in xulrunner though. Mar 08 10:50:33 so does MicroB just use a plain xulrunner? Mar 08 10:51:11 and implement the funny scroll mechanism (where you see the grid on unrendered parts) on top? maybe in XUL? Mar 08 13:51:53 See that guys http://goo.gl/zWa5yY thx you Mar 08 13:58:41 Location: http://www.youtube.com/watch?v=F-zIaOHs-30 Mar 08 13:58:52 wonder what the point of the URL shortener was. Mar 08 13:59:40 some gaming video. Mar 08 14:28:39 same URL spamming Mar 08 14:28:45 please stop that! Mar 08 14:30:25 as a general rule you shouldn't post any URL in channel without sufficient context explaining how it's relevant and on topic, and what it is about Mar 08 14:31:37 I guess mMrUtuber got k-lined by freenode meanwhile Mar 08 14:34:01 this is not one of the hardcore channels where posting any URL automatically causes a ban on you, but nevertheless posting of URLs shall be on topic, and users are strongly encouraged to ignore and NOT click any suspicious URL getting posted Mar 08 14:34:26 Sorry. Mar 08 14:35:00 I'll avoid unshortening probably spam URLs next time. Mar 08 14:35:03 np, I'm not even able to adhere to these rules myself all the time ;-) Just quoting them to let you all know Mar 08 14:35:28 though I only showed what it was because it obviously wasn't something malicious. Mar 08 14:35:44 well, probably a click harvester Mar 08 14:36:29 or simply ad spam Mar 08 14:37:00 though real exploits get spread by same means Mar 08 14:38:39 Maxdamantus: thanks for checking it anyway :-) Mar 08 14:42:57 Hm. Wikipedia works well with Conkeror/xulrunner on the N900 at least. Mar 08 14:43:11 * Maxdamantus added some crude touch-drag thing to it. Mar 08 14:56:22 Conkeror would probably be a good starting point for a completely chromeless mobile browser. Mar 08 14:56:30 for mobiles with keyboards. Mar 08 14:56:45 * Maxdamantus finds it more comfortable to hold with the keyboard open anyway. Mar 08 14:58:21 * Maxdamantus might try coming up with something nice tomorrow. Mar 08 17:32:41 :-D Mar 08 22:21:32 Just wonder. I tried to write as root something to /sys/devices/platform/gpio-switch/proximity/state since it's writeable, but I get permission denied. Mar 08 22:24:16 yes, well, it's a physical input Mar 08 22:24:44 But as it is gpio, general purpose input output Mar 08 22:25:05 you could change it to output and then echo Mar 08 22:25:09 BUT Mar 08 22:26:36 That would make the cpu drive the pin to your desired state. while proximity sensor attempts to drive the pin to what it thinks is true Mar 08 22:26:47 One of them will die, or both Mar 08 22:27:27 ShadowJK: I see. Mar 08 22:28:51 (or alternatively, both survive, burning enormous amounts of power) Mar 08 22:29:01 ShadowJK: the device connected to this (proximity sensor) is dead, and the kernel has set it's state to closed, and disable to 1. I'm thinking if I could make applications think it's in the opposite state w/o modyfing the kernel. Mar 08 22:29:01 ((constantly)) Mar 08 22:30:07 Proximity sensor is used by very few things Mar 08 22:30:10 Because I'd rather have them think it reports there's nothing in proximity and lock the screen as needed. Mar 08 22:30:44 Personally I'd either do LD_PRELOAD tricks to the affected apps, or hack kernel... Mar 08 22:31:23 ShadowJK: it's used by the phone app, so I get screen locked when someone's calling and I have to unlock and then answer. This would be OK in my pocket, but it's happening when the phone lays onobscured on my desk. Mar 08 22:32:59 As a new owner of a Jolla phone, I find your attention to detail amusing :D Mar 08 22:33:09 ShadowJK: in either case, I would have to setup a dev envirioment to compile something. Mar 08 22:34:07 yeah :( Mar 08 22:34:18 ShadowJK: not sure what does that have to do with Jolla. Anyway, I could simply flip .active_low in the kernel to the opposite, is that right? Mar 08 22:35:07 I would find the prxomity driver and make it return false all the tme Mar 08 22:35:25 but flipping active_low sounds as fine a hack as any Mar 08 22:36:20 I don't know which file is that unfortunatelly. I found in board-rx51-peripherals.c the set-up for GPIO keys. Mar 08 22:38:07 Oh damn. I'm watching the Android-N900 kernel. Mar 08 22:43:07 You probably want power52 Mar 08 22:43:35 maybe another alternative would be to just mount a tmpfs over /sys/devices/platform/gpio-switch/proximity/ and make a "state" file with the content "1" Mar 08 22:45:01 if it's handled by mce, I think it just opens the file each time and reads it. Mar 08 22:45:20 at least, when vibrating, it just opens a file in /sys and writes to it. Mar 08 22:46:04 er, not "1" .. "open" Mar 08 22:51:59 CR0W: you *might* try mounting a plain file directory over the /sys/devices/platform/gpio-switch/proximity/, and have plain files in there to which you can write the state you want. We tried that for iirc magnet sensor and it worked until next boot, then stuff gone terribly haywire iirc Mar 08 22:52:31 DocScrutinizer05: What do you mean by "then stuff gone terribly haywire iirc"? Mar 08 22:52:32 ooh Maxdamantusalready said exactly same Mar 08 22:53:08 iirc we seen some process who used toread from that sysnode spinning busy and eating 100% cpu Mar 08 22:53:26 Anyway, I'll try that. What is the name of the phone app exec so I can kill it? Mar 08 22:53:44 you'll notice via top ;-) Mar 08 22:53:51 Okay. Mar 08 22:54:22 or worst case, a reboot will safe you, until you make that mount in initscripts, which is what didn't work iirc Mar 08 22:54:31 save* Mar 08 22:54:58 most likely though, Maxdamantus been right with his guess about mce Mar 08 22:55:31 What's mce? Mar 08 22:55:38 in which case you could use your own version of mce instead of a custom patched kernel Mar 08 22:55:42 ~mce Mar 08 22:55:42 MCE = Machine Check Exception. Mode Control Entity Mar 08 22:55:52 the latter Mar 08 22:56:04 /etc/mce/mce.ini Mar 08 22:56:08 So it's userland? Mar 08 22:56:11 yep Mar 08 22:56:23 That would be safer to modify then. Mar 08 22:56:43 could also be kerecv Mar 08 22:58:09 /usr/sbin/ke-recv Mar 08 22:58:16 rtcomm-call-ui? Mar 08 22:58:27 maybe, but unlikely Mar 08 22:58:39 most maemo apps use some sort of middleware Mar 08 22:58:59 Damn, it's rerun after kill. Mar 08 22:59:53 what?mce? Mar 08 23:00:25 Hmm .. you tried it for the back cover magnet sensor? Mar 08 23:00:27 No, that process I mentioned. Mar 08 23:00:43 which process did you mention? Mar 08 23:00:49 Maxdamantus: yep Mar 08 23:00:50 DocScrutinizer05: the proximity sensor has open or closed in uncovered state? Mar 08 23:01:00 DocScrutinizer05: rtcomm-call-ui Mar 08 23:01:03 open afaik Mar 08 23:01:09 I think the SD device is made invisible by the kernel, so it would just prevent userspace from realising the device went missing. Mar 08 23:01:26 yeah, rtcom-call-ui gets preloaded by hildon Mar 08 23:01:52 it's nearly impossible to shut it down, you need to edit the .desktop file Mar 08 23:02:53 Maxdamantus: I can't follow Mar 08 23:03:04 DocScrutinizer05: I mounted a tmpfs over proximity and added state and disable files with the right contents. I'll kill the process and see. Mar 08 23:03:16 yep Mar 08 23:03:42 don't kill such processes too often, or you'll see a reboot Mar 08 23:04:12 Damn, it works! Mar 08 23:04:20 until next boot. toldya Mar 08 23:04:32 So now, I just need to add this to the init system. Mar 08 23:04:56 I don't know how it works on debian, can I get a quick guide? :) Mar 08 23:06:01 last time adding it to sysinit caused massive trouble Mar 08 23:06:02 DocScrutinizer05: in the omap_hsmmc driver, there's a omap_hsmmc_protect_card function that checks whether the cover is open. Mar 08 23:06:13 I dunno why Mar 08 23:06:38 which obviously doesn't work through sysfs Mar 08 23:06:55 err yup, that too Mar 08 23:06:56 Maxdamantus: hey, thanks. So I need to change this function and rebuild the kernel to "fix" the hall sensor. Mar 08 23:07:33 there are patched kernel modules anyway Mar 08 23:09:03 CR0W: I don't know. I'm just pretty sure you need to do something like that to fix the effects of the back cover sensor. Mar 08 23:09:20 because the effects are in the kernel, rather than in userspace. Mar 08 23:09:34 I'm guessing that sensor being broken or magnet missing is a common issue. Mar 08 23:10:15 DocScrutinizer05: is there a source for that? Mar 08 23:10:36 I think there's a common fix to the magnet missing thing though: add another magnet. Mar 08 23:10:45 dunno what the stand feels like without the magnet. Mar 08 23:10:53 feels nice with it, because it snaps onto the back. Mar 08 23:11:43 Maxdamantus: (efects in kernel) both kernel and userland effects Mar 08 23:11:46 * Maxdamantus would disable the sensor anyway if he ends up using a newer kernel, because he'd also use btrfs across the SD card and eMMC. Mar 08 23:12:19 CR0W: I'm not sure about source available Mar 08 23:12:40 DocScrutinizer05: I'll figure it out, thanks. Mar 08 23:12:57 Now I know that that proxmity sensor "fix" won't do. Mar 08 23:13:32 Because when the screen is locked now, the phone app unlocks it, so I could receive a call in a pocket. Mar 08 23:13:51 well, possibly the card had io errors while userland thought it should be available Mar 08 23:13:56 Does that app have source available? Mar 08 23:14:04 according to Maxdamantus reasoning Mar 08 23:15:00 which app? Mar 08 23:16:22 The phone call one. Mar 08 23:16:48 I would make it ignore the proximity sensor. Mar 08 23:17:08 rtcom-call-ui. nope. and there's more than just rtcomm-call-ui to it when doing a phonecall. Not trivial to even spot what needed "fixing" Mar 08 23:18:27 I'd honestly rather fix the hw Mar 08 23:18:51 I'm working on that. Mar 08 23:19:07 a flex cable is pretty cheap Mar 08 23:19:30 If the charge indicator, front camera, proximity sensor are all dead, it's most likely flex, right? Mar 08 23:19:39 sure Mar 08 23:19:54 It's weird the LCD works, I guess the flex is damaged in that left part only. Mar 08 23:20:03 possible Mar 08 23:20:48 I found a "factory new original" one, but a few persons had trouble with the seller and I'm worried I'll join them. Mar 08 23:21:04 one broken VDD or GND line may suffice to cause all those errors Mar 08 23:21:07 So now my strategy is to buy a more broken N900 for spare parts. Mar 08 23:21:32 that's nonsense particularly for flex cable Mar 08 23:21:33 I wonder if it's something that could be fixed by tape. Mar 08 23:21:45 no Mar 08 23:22:26 flex cables for N900 sources are legion on fleabay Mar 08 23:22:56 and for a 7 bucks, your risk is manageable ;-) Mar 08 23:22:57 I'd rather not buy something from abroad, but I'll think about it. Mar 08 23:24:54 I'd send you one of my spares, but I'm afraid I can't afford to package and pay the porto to compete with what it costs in china Mar 08 23:26:23 the thing the flex cable connects to on the motherboard is they bit you were unsure about obtaining for a while, isn't it? Mar 08 23:27:06 so there are people actually manufacturing the female side, but not the male side? Mar 08 23:27:23 maybe other way around Mar 08 23:27:29 yes Mar 08 23:27:55 but we found that connector Mar 08 23:28:48 also I doubt anybody is building new flex cables now Mar 08 23:28:59 Ah. Mar 08 23:29:00 might even be original Nokia spare parts Mar 08 23:29:43 though some flex come with camera fake, those for sure are mo original Nokia parts ;-) Mar 08 23:29:59 no* Mar 08 23:30:45 we should've started that Neo900 project a year earlier, would've made a lot of things easier for us Mar 08 23:31:29 the connector got EOL'ed a mere 2 months before we started searching for it Mar 08 23:32:51 but we found a source that's supposed to have another 10k or maybe more on stock Mar 08 23:33:10 so if our 1000 don't suffice... Mar 08 23:33:32 ...we will order more of them Mar 08 23:46:22 Anybody knows the most up-to-date OS version number for an N900? Mar 08 23:53:05 ~pr131 Mar 08 23:53:06 [combined] the rootfs fiasco image of maemo. For N900 latest (PR1.3.1) see http://nds2.fds-fire.nokia.com/fdp/interface/FiRe/2011/9/--FID--A0A22MVWFVFAM/--LID--FiRe1317015685654/RX-51_2009SE_21.2011.38-1_PR_COMBINED_MR0_ARM.bin (or http://maemo.cloud-7.de/maemo5/fiasco+co/ ) Mar 08 23:53:37 aka 21.2011.38-1 Mar 08 23:54:22 or simply cssu-testing-9? Mar 08 23:56:34 ~cssu Mar 08 23:56:34 rumour has it, cssu is http://wiki.maemo.org/Community_SSU Mar 08 23:56:58 Latest version: Testing(2014-01-01): 21.2011.38-1Tmaemo9.1 Mar 08 23:57:38 which probably is as recent as it gets Mar 08 23:57:57 err, as up-to-date as it gets Mar 08 23:58:10 Bernte: ^^^ Mar 08 23:58:56 infobot: literal pr131 Mar 08 23:58:57 "pr131" is "see combined" Mar 08 23:59:09 infobot: literal combined Mar 08 23:59:10 "combined" is "the rootfs fiasco image of maemo. For N900 latest (PR1.3.1) see http://nds2.fds-fire.nokia.com/fdp/interface/FiRe/2011/9/--FID--A0A22MVWFVFAM/--LID--FiRe1317015685654/RX-51_2009SE_21.2011.38-1_PR_COMBINED_MR0_ARM.bin (or http://maemo.cloud-7.de/maemo5/fiasco+co/ )" Mar 08 23:59:48 infobot: no, combined is the rootfs fiasco image of maemo. For N900 latest (PR1.3.1) see http://nds2.fds-fire.nokia.com/fdp/interface/FiRe/2011/9/--FID--A0A22MVWFVFAM/--LID--FiRe1317015685654/RX-51_2009SE_21.2011.38-1_PR_COMBINED_MR0_ARM.bin Mar 08 23:59:48 DocScrutinizer05: okay Mar 09 00:01:46 infobot: combined is also http://galif.eu/nokia/ Mar 09 00:01:46 okay, DocScrutinizer05 Mar 09 00:02:19 * CR0W good night Mar 09 00:02:26 n8 CR0W **** ENDING LOGGING AT Sun Mar 09 02:59:58 2014