**** BEGIN LOGGING AT Thu Mar 14 02:59:58 2013 Mar 14 04:36:44 hi! Mar 14 05:02:49 http://talk.maemo.org/showthread.php?p=1328897#post1328897 Mar 14 05:03:11 nice Mar 14 05:11:51 * DocScrutinizer05 moos at freemangordon Mar 14 05:12:04 morning doc Mar 14 08:58:10 ~seen pali Mar 14 08:58:15 pali <~pali@Maemo/community/contributor/Pali> was last seen on IRC in channel #maemo, 14h 54m 53s ago, saying: 'I remember that there was big problems about it...'. Mar 14 08:58:58 2) Delete the expired Release, Release.gpg, Packages and Packages.gz Mar 14 08:58:59 from the local replica of the repo Mar 14 08:59:01 3) Use atpftparchive and gzip to create these files anew Mar 14 08:59:02 4) Sign the whole thing using our renewed 4v1 key Mar 14 08:59:04 5) Upload the new signed repo back to SSU into a temp folder. Mar 14 08:59:31 renewed 4v1 key??? how is this gonna work? Mar 14 08:59:55 no cert on any device for such a "renewed 4v1 key" Mar 14 09:01:19 I told Pali they wouldn't really grok the ramifications of that 4v1 exoired Mar 14 09:01:35 kerio: freemangordon: ^^^ Mar 14 09:01:53 (from Matti/ext-Nokia) Mar 14 09:02:11 yeah, they don't understand Mar 14 09:02:33 there's no mechanism for automatic key renewal in apt Mar 14 09:02:50 that's the whole point Mar 14 09:02:59 you have packages that hold keys that are updated quite some time before the current keys expire Mar 14 09:03:44 so rolling out what's basically an update, based on a "renewed" 4v1-key, without shipping new cert to install to devices same time, is futile Mar 14 09:04:19 aiui Mar 14 09:04:53 given the expiration date is in the cert on device, not in the signing key used on server side Mar 14 09:05:16 unless I'm *completely* mistaken Mar 14 09:07:41 we need to *somehow* deploy a *new* standard key cert to devices Mar 14 09:08:18 and/or sign the repo with a key that has a valid cert on all devices still Mar 14 09:08:43 such key would be that sw-admin key Mar 14 09:09:10 yep Mar 14 09:10:02 any such repo"-update" SHOULD deploy a renewed cert for a 4v9 key, which community could inherit Mar 14 09:10:06 and to deploy the key, the only option at this point is to poison packages in extras Mar 14 09:10:10 but fuck that Mar 14 09:10:27 :nod: and that's exactly what we should do Mar 14 09:10:42 no we shouldn't Mar 14 09:11:40 especially because we should poison *everything* to be sure Mar 14 09:11:51 and even that it's not guaranteed to work Mar 14 09:12:09 though I think we could also provide a new certs-pkg for SSU that includes a cert to a key we created and that gets included to SSU as an update to the existing certs-pkg and then gets signed (like whole repo"-update" by that SW-adm-key so devices would accept it Mar 14 09:12:31 s/te"/te")/ Mar 14 09:12:32 DocScrutinizer05 meant: though I think we could also provide a new certs-pkg for SSU that includes a cert to a key we created and that gets included to SSU as an update to the existing certs-pkg and then gets signed (like whole repo"-update") by that SW-adm-key so devices would ... Mar 14 09:14:09 from that moment on we could sign new SSU updates with that new community-borne secret key Mar 14 09:14:43 while the then-current SSU would be signed by SW-adm key and probably would stay like that for quite some time Mar 14 09:14:58 hm, i don't think that there's a way of doing that cleanly Mar 14 09:15:05 for *next* update however we wouldn't need the SW-adm key anymore Mar 14 09:15:07 you'd have to add a key to a domain Mar 14 09:15:35 and iirc that's dealt with by the ham package Mar 14 09:16:06 oh c'mon there *is* *some* package in core Nokia repos that ships the certs, no? Mar 14 09:17:03 yeah, quite likely *ham*.dpk Mar 14 09:17:28 pali elaborated on all that Mar 14 09:18:24 however the suggested solution been not clear enough to Nokia it seems, or maybe not even been properly thought Mar 14 09:20:47 I asked pali to send a postscriptum mail to matti, explaining clearly what needs to be done. He preferred to wait for Matti's answer first. Now e have that answer. I think it's about time to send that postscriptum mail now, that clearly puts straight what are the steps to get done, to reclaim control over the SSU repo Mar 14 09:24:42 i'm off since some months, are still fightin with nokia for SSU repo? Mar 14 09:24:47 DocScrutinizer05: one part of it is the keys... ah ok, we can just ship the updated key Mar 14 09:25:13 the thing is, HAM wants a valid key and wants that key to be listed in its weird domain thing Mar 14 09:25:42 kerio: there *is* *one* valid key: SW-adm-key Mar 14 09:25:47 indeed Mar 14 09:26:18 but a package to add more keys to the current domains will kinda conflict with the existing one Mar 14 09:26:37 use this key to deploy a renewed cert for 4v1.1 key Mar 14 09:26:47 that would work Mar 14 09:27:09 thought as much ;-D Mar 14 09:53:33 it's still kind of ugly to make though Mar 14 09:53:51 because iirc variant-keys.gpg is in the HAM package Mar 14 09:54:51 oh no, it's a separate package Mar 14 09:55:28 DocScrutinizer05: so, we need to test what happens if we make hildon-application-manager-settings-standard a user package and update that Mar 14 09:55:44 like, what happens if a user tries to remove it Mar 14 09:56:17 why do we need to make it a user package? Mar 14 09:56:27 because otherwise HAM won't even prompt for its update Mar 14 09:56:32 it's either that, or pr1.3.2 Mar 14 09:56:34 why not simply edit maemo-MP-PR? Mar 14 09:56:51 call it PR1.3.2 and good Mar 14 09:56:56 i'm not entirely sure nokia will like that Mar 14 09:57:06 also, if "we" do pr 1.3.2 we should also add the turktrust fix Mar 14 09:57:07 Nokia doesn't give a shit Mar 14 09:57:17 that's the idea Mar 14 09:57:23 yeah but we're asking nokia to do those modifications for us Mar 14 09:57:52 Matti probably asked because they planned to add turktrust fix Mar 14 09:58:22 no, Nokia is doing those for them Mar 14 09:58:31 since they want to fix that Mar 14 09:58:39 it's not "for us" Mar 14 09:59:33 it's ext-nokia/matti who's willing to fix fremantle repos Mar 14 09:59:39 i see Mar 14 10:00:07 when he needs to edit mp-pr for that, so what? Mar 14 10:00:57 when he declares that PR1.3.2, even fine!! "nokia still cares" who's to be blamed for that? Mar 14 10:01:09 yay nokia :3 Mar 14 10:01:35 but i thought that to release a new pr they needed extensive testing Mar 14 10:01:56 and actually Nokia-the-inc kinda actually still cares, since they pay Matti and colleagues to maintain fremantle Mar 14 10:02:21 c'mon extensive testing for *one* updated package? Mar 14 10:03:06 it would seem like you can scientifically prove what impact it can have, at worst case Mar 14 10:04:10 just calling it pr1.3.2 and shipping a new cert-pkg doesn't mean we need to re-test dialer, modest, hildon-desktop, whatnot else Mar 14 10:04:13 also, aren't you supposed to fix things such as CA problems really, really quickly? :) Mar 14 10:04:54 meh, OT Mar 14 10:08:20 how do I check if a BL-5J is genuine? Mar 14 10:08:46 ok, so here we go: create a new signature key 4v99. add the cert of that key to the certs.dpk. add that updated certs.dpk to repo, edit mp-pr to depend on that new certs.dpk. Add turktrust fixes to the browser certificates (!!unrelated!!). Sign the whole stuff with SW-adm key, call it "PR1.3.2 Maemo Fremantle security system-update" and deloy via Nokia SSU Mar 14 10:09:11 have a toy (MP3 player) that contains a BL-5J, would know if I can safely use it as 2nd battery for the N900 Mar 14 10:10:07 yes, you can Mar 14 10:10:30 the toy came from the Ukraine, it may be a fake ;) Mar 14 10:10:38 doesn't matter Mar 14 10:11:20 ah, it IS a fake, as I see now, it does not have nokia written on it... ok, then it probably is "honest rebuilt" ;) Mar 14 10:11:24 like the ebay stuff Mar 14 10:11:26 you're 99.999999999% *safe* to try it anyway Mar 14 10:11:31 good :P Mar 14 10:11:46 worst thing that could happen: it performs poorly or doesn't work at all Mar 14 10:11:51 ok Mar 14 10:12:14 it quite sure will not explode ;) Mar 14 10:12:23 it worked in the player after all Mar 14 10:27:33 ok, my current state of the disk encryption stuff AND the gconf moving: MOST settings survive, oddly, the screen vibration setting gets reset on reboot Mar 14 10:27:41 it seems to be the only one yet Mar 14 10:28:07 and for some reason (I DID make homedir and gconf empty for normal bootup) the clock is in 12 hour mode despite the language scheme set to Europe/Berlin Mar 14 10:28:46 (ok, just found that setting now) Mar 14 10:30:44 also, WLAN settings seem to get lost at bootup, but THAT probably is fixable by a SIGHUP or SIGTERM to the right process Mar 14 10:31:17 (but that's unconfirmed, I just had a bad power down due to removing the battery, fsck then wanted some stuff to be manually fixed, so it MIGHT also have been that) Mar 14 10:31:25 I had thought I can safely pull the battery when on USB... I could not Mar 14 11:03:09 yes, I do lose the WLAN settings on every bootup with my current setup... apparently these are started by upstart before the mounts Mar 14 11:17:39 divVerent: well, you basically can't pull battery any time unless device powered down. (there are tricks to do so, see my tmo "hotswap battery" thread. that however is not exactly interesting for your usecase I guess) Mar 14 11:21:03 haha, so basically this is yet another thing this evil bme does... Mar 14 11:23:21 BTW, how critical are the operations of bme anyway? Does bme actually control the charging (so a misbehaving bme could actually cause one of those LiIon fires)? Mar 14 11:23:53 Or is it like "usual" and the actual charging circuit is in the battery, and bme only controls how much current it can crain from USB? Mar 14 11:25:46 divVerent: the worst you can do is charge the battery too quickly and wear it down Mar 14 11:26:02 but if the battery is good, that won't be dangerous Mar 14 11:26:11 there's two chips regulating battery and charge Mar 14 11:26:18 ah, ok Mar 14 11:26:31 one of them is supposed to be mounted to the battery, but it's in the n900 instead Mar 14 11:26:50 which leads to loss of calibration when you change batteries :( Mar 14 11:26:59 apart from that, it's Safe™ Mar 14 11:27:11 DocScrutinizer05: correct? Mar 14 11:27:20 BTW, is it safe to do the calibration script when just being connected to a notebook?+ Mar 14 11:27:20 PROBABLY yes, in worst case it'll power off at the end Mar 14 11:27:22 is my guess Mar 14 11:27:34 but that's already AFTER the mAh value got updated Mar 14 11:27:46 i think you can reach VDQ with a usb host charger Mar 14 11:27:54 but i'm not sure Mar 14 11:28:03 because musb_core takes lots of power to keep a usb host connection active Mar 14 11:28:10 so you have... hm, 420mA for charging? Mar 14 11:28:12 ah, I see Mar 14 11:28:15 yeah, that should be enough Mar 14 11:28:19 but basically, the script will tell me if it won't work Mar 14 11:28:40 the script does nothing you couldn't do with your hands Mar 14 11:28:44 as VDQ is that state the battery needs to be in to start learning Mar 14 11:28:45 sure Mar 14 11:28:48 well, it does ONE thing Mar 14 11:28:51 resume charging when done Mar 14 11:29:03 you could do that with your hands! Mar 14 11:29:09 if I watch it all the time, sure ;) Mar 14 11:29:11 /leave #maemo Mar 14 11:29:14 urgh. Mar 14 11:29:19 don't leave #maemo! Mar 14 11:29:20 :( Mar 14 11:30:29 all the script really seems to do, is: stopping BME, checking VDQ ("full") state first, i2c hacks to turn off any charging despite being connected, stopping BME based charging, and waiting till battery is empty enough to have updated its data Mar 14 11:30:42 but most of the hacks are only needed BECAUSE of being plugged in while doing it Mar 14 11:31:01 so I would THINK a recalibration can also be done by first ensuring VDQ, then just unplugging and waiting for it to turn itself off Mar 14 11:31:55 unless BME then would power down too early, in which case shutting down BME would also be necessary ;) also, LiIons don't like being empty for a while, which is why the method to instantly turn on charging again when "empty" is better to the battery Mar 14 11:32:04 than e.g. letting it run out over night then sitting in that state for hours Mar 14 11:33:07 I do wonder though why a "recalibration script" even was necessary... I mean, notebooks don't need that Mar 14 11:34:00 BTW, does BME really FULLY ignore the calibration data? Mar 14 11:36:24 as in, does a miscalibration on that chip lead to too early power down? Mar 14 11:41:10 kerio: yep, correct Mar 14 11:42:11 calibration will work from USB as well Mar 14 11:42:19 USB host that is Mar 14 11:42:57 and what is the impact of wrong calibration data, when using BME? none? Mar 14 11:42:59 it's using normal bme controlled charging, just stops bme to discharge battery til learning cycle completed, then restarts bme Mar 14 11:43:10 divVerent: none Mar 14 11:43:12 good Mar 14 11:43:20 just asking, as I am now calibrating for the other battery Mar 14 11:43:28 and when that is done, I probabyl won't recalibrate for the main one again Mar 14 11:43:31 as I know its capacity now Mar 14 11:44:14 it's 1158mAh, quite the expected value given initial capacity and age of the battery Mar 14 11:44:38 calibration will usually not work the manual way you suggested. BME / bat-applet will shutdown before chip completed learning cycle Mar 14 11:44:44 ok Mar 14 11:44:51 so for the "manual way" one would have to stop BME anyway Mar 14 11:44:55 and thus risk deep discharge Mar 14 11:45:22 which probably the security chip on the battery will prevent, but it lowers battery life anyway Mar 14 11:45:42 yes Mar 14 11:46:02 also, I remmeber Nokias don't like to wake up immediately from a totally discharged battery even when on the charger Mar 14 11:46:08 the N900 probably has the same issue :P Mar 14 11:46:22 and then take up to half an hour to be able to be powered on again Mar 14 11:46:49 ~flatbatrecover Mar 14 11:46:51 Remove battery for 1 minute. Insert battery. Plug powered Nokia wallcharger to device. Watch steady amber. Let sit and charge. Do NOT try to boot. After 30 min, you got either a) a booted up N900, b) flashing amber which means you can boot, c) steady amber going off - in this case start over again with ~flatbatrecover Mar 14 11:47:12 yes, same procedure as with other nokias :P Mar 14 11:47:22 divVerent: the n900 is similar to every other nokia, yes Mar 14 11:47:28 that's why ACT_DEAD exists, afaik Mar 14 11:47:31 well, that and alarms Mar 14 11:47:37 AH, that's what it does :P Mar 14 11:47:49 nope, ACT_DEAD is unrelated to flatbatrecover Mar 14 11:47:49 so even in this state the linux system is actually booted... Mar 14 11:47:51 kinda weird design Mar 14 11:47:53 oh, ok then Mar 14 11:47:59 flatbatrecover is a NOLO thing Mar 14 11:48:06 no Mar 14 11:48:06 charging while "turned off" still boots the kernel Mar 14 11:48:11 no? Mar 14 11:48:23 12:48:06 kerio | charging while "turned off" still boots the kernel Mar 14 11:48:33 which explains the stupid annoying "cannot power off when on USB" issue ;) Mar 14 11:48:34 emergency charging is a mere hw thing, done because CPU can't boot yet Mar 14 11:48:35 that's the normal charge Mar 14 11:48:41 probably has to be done so that BME can run Mar 14 11:48:50 DocScrutinizer05: oh, so it's even before ROMBL? Mar 14 11:48:55 yes Mar 14 11:48:59 i see Mar 14 11:49:18 what checks VBAT before allowing you to power on? Mar 14 11:49:53 SoC needs a certain minimum voltage of some 3300mV or whatever, to boot into cpu working state Mar 14 11:49:55 ah... now I see what emergency charging is REALLY for Mar 14 11:50:02 the dead battery state... NOT the case that BME doesn't work :P Mar 14 11:50:03 emergency charging is to allow you to boot Mar 14 11:50:12 yes Mar 14 11:50:44 divVerent: this design has a critical flaw that leads to a "brick" if you don't have another way to charge the battery: Mar 14 11:50:45 it's a property of the bq24150 charger chip Mar 14 11:50:46 hm... I don't trust the calibration value... 1158mAh sounds good, it has 1320 printed on it... but that's 88% of original Mar 14 11:50:52 but aren't they supposed to lose 20% per yearß⎈ Mar 14 11:50:55 to bootstrap from flat battery Mar 14 11:50:57 if you bork your rootfs, and lower the battery below the flashing threshold, you're boned Mar 14 11:51:24 boned-ish, because you can still boot rescueOS and charge from that, aiui Mar 14 11:51:31 that's another unrelated issue Mar 14 11:51:36 no, wait, can you? Mar 14 11:51:40 and that's actually a NOLO thing Mar 14 11:52:04 you can, as long as NOLO allows it ;-) Mar 14 11:52:12 DocScrutinizer05: will NOLO allow you to load rescueOS with a low battery? Mar 14 11:52:17 kerio: but emergency charging still works... or does it not bring the battery up enough? Mar 14 11:52:28 emergency charging is enough to boot, not enough to flash Mar 14 11:52:28 I actually dunno if NOLO would allow RAMload rescueos Mar 14 11:52:44 DocScrutinizer05: NOLO isn't signed, right? Mar 14 11:52:53 NOLO isn't signed Mar 14 11:52:58 XLoader is Mar 14 11:52:59 can't we lower the value it checks for with hex editing? Mar 14 11:53:13 sure Mar 14 11:53:42 you 'just' need to disassemble and analyze/RE it Mar 14 11:53:53 ;-D Mar 14 11:54:12 just change every instance of the voltage it wants with a lower voltage Mar 14 11:54:15 what can possibly go wrong Mar 14 11:55:19 http://talk.maemo.org/showthread.php?t=83880 Custom NOLO Splash, USB and R&D icons Mar 14 11:55:39 let's get HiFo to pay for a Hex-Rays ARM license for jonwil! Mar 14 11:55:54 well, you have a few 100k of bytes you can try to change ;-P Mar 14 11:56:29 i can't believe nobody has put the arm disassembler on bittorrent :s Mar 14 11:56:30 kerio: now THAT would be a great idea, and actually inline with what HiFo is meant to do Mar 14 11:56:55 ah, here's an explanation: https://news.ycombinator.com/item?id=2611984 :( Mar 14 11:58:32 DocScrutinizer05: apparently hex-rays doesn't even want your money unless you're special or have a perfect track record of keeping it for yourself Mar 14 12:06:13 wow Mar 14 12:06:19 another stranded carnival cruise ship Mar 14 12:11:41 another cryptic comment that nobody can make sense out of Mar 14 12:12:28 lol Mar 14 12:14:19 kerio: when you paid a sufficiently high amount of money for a licence of a program that comes with dongle and watermark, then your motivation to crack and piratebay it are rather limited Mar 14 12:15:39 DocScrutinizer: the hex-rays arm decompiler has a 105GB bounty on what.cd Mar 14 12:15:40 particularly when a support account is a huge part of that deal, and you'd risk to see that account go poof and void when you'd dare to mess with proliferating your copy Mar 14 12:15:56 that's a lot! Mar 14 12:17:27 think of it like MSDN, kind of Mar 14 12:18:27 you're paying to be a member of a very exclusive club, where quite exclusive service is offered to those who are members Mar 14 12:19:20 and I guess all members contribute to that service by sharing their own results to others, indirectly via the company Mar 14 12:19:22 msdn is far more mundane Mar 14 12:19:27 hmm MSDN isn't that exclusive nowadays isn't it? Mar 14 12:19:27 indeed Mar 14 12:20:01 hex-rays sits in stratospheric ivory towers Mar 14 12:24:09 HEXALLLx86+ARM Decompiler Base License (Linux) 2699 EUR Mar 14 12:24:17 absolutely reasonable Mar 14 12:24:23 actually a bargain Mar 14 12:24:40 compare that e.g. to Lauterbach products Mar 14 12:24:45 don't you have to send in dna for a copy? Mar 14 12:25:01 as well as your first born's soul for collateral? Mar 14 12:26:13 >>This order form is for IDA Computer Licenses only. IDA Computer licenses can be used on a single computer b yany user, provided than the simultaneous seat count does not exceed the license seats. In order to benefit fro mthe discounted "additional user" prices for IDA, additional seats have to be purchased at the same time as th einitial license they complement.<< So no, I can't see any special conditions Mar 14 12:27:24 DocScrutinizer05: is that the disassembler or the decompiler? Mar 14 12:27:48 HEXARM LARM Decompiler Base License (Linux )1800 EUR Mar 14 12:27:54 hm Mar 14 12:28:16 are the licenses transferrable? Mar 14 12:28:23 :-/ Mar 14 12:28:24 like, if HiFo actually decided to buy it Mar 14 12:28:26 >>This order form is for IDA Computer Licenses only. IDA Computer licenses can be used on a single computer b yany user, provided than the simultaneous seat count does not exceed the license seats. In order to benefit fro mthe discounted "additional user" prices for IDA, additional seats have to be purchased at the same time as th einitial license they complement.<< So no, I can't see any special conditions Mar 14 12:28:32 it does say single computer any user Mar 14 12:28:39 hm Mar 14 12:29:34 well, we could install it on a private VM on maemo.org Mar 14 12:29:42 :-D Mar 14 12:29:46 and let just one person at a time access it Mar 14 12:29:54 we probably are allowed to do so, yes Mar 14 12:30:20 i mean, what do we need money for? Mar 14 12:30:25 given you can't log in with more than one VNC concurrently Mar 14 12:30:47 i was thinking x forwarding actually Mar 14 12:30:56 or that Mar 14 12:31:39 i wonder if we can batch process every ARM binary on dmnc :> Mar 14 12:32:08 DocScrutinizer05: where is that order form? Mar 14 12:33:12 oh nm Mar 14 12:33:21 https://www.hex-rays.com/products/decompiler/order.shtml Mar 14 12:33:35 you could have known, no? Mar 14 12:33:59 DocScrutinizer05: hehe, 1800 plus the cost of IDA Mar 14 12:34:30 huh? Mar 14 12:34:44 hex-rays runs on top of ida Mar 14 12:34:51 you think ALL doesn't mean all? Mar 14 12:35:03 hmm Mar 14 12:35:06 dunno Mar 14 12:37:41 IDAPROCL IDA Pro Base Computer License (Linux )1299 EUR Mar 14 12:38:20 k, that's 3k € then, still not too much for the amount of funds HiFo has collected ;-D Mar 14 12:42:25 UPDHEXARML Support renewal for ARM Decompiler License (Linux )600 EUR <- 1 year Mar 14 12:43:01 DocScrutinizer05: we have permission to redistribute (C)Nokia things, right? Mar 14 12:43:08 who has the right to the decompiled source? Mar 14 12:43:13 UPDPROCL support renewal for IDA Pro Computer License (Linux) 429 EUR Mar 14 12:43:46 interesting question Mar 14 12:44:19 non-free with source is an interesting concept Mar 14 12:44:31 yup Mar 14 12:47:01 another thing... often when I power up the N90, it instgantly shuts down after dsme startup Mar 14 12:47:14 by setting upstart's bootstate to SHUTDOWN Mar 14 12:47:36 any obvious reason for it? Mar 14 12:47:41 I see no errors in the backlight-less console Mar 14 12:48:03 which I see due to added "modprobe fbcon" early Mar 14 12:49:10 though not entirely unheard, it's not a really known problem Mar 14 13:04:43 kerio: i thought decompiled anything rewritten was fair game Mar 14 13:05:07 i think that binary->decompiler->upload is stretching it Mar 14 13:05:10 ie: manage to learn how they did it and use that in order to write your own that will function Mar 14 13:05:32 DocScrutinizer05: yes, it's normally invisible Mar 14 13:05:36 no i meant batch uploading the decompiled source Mar 14 13:05:41 and just shows up as power up not doing "the right thing", and a second attempt to work Mar 14 13:05:49 so I suppose most people don'Ät even care :P Mar 14 13:06:02 to me it SEEMS to happen especially when connected to USB Mar 14 13:06:19 not sure Mar 14 13:06:30 but.... Mar 14 13:06:36 i'm pretty sure i saw an article about this Mar 14 13:06:38 it's probably not going to be remotely good anyway Mar 14 13:06:46 you have to tweak a bunch of things when decompiling Mar 14 13:06:55 that you can take a decompiled program, rewrite, and distribute as you see fit Mar 14 13:08:11 In the United States, the copyright fair use defense has been successfully invoked in decompilation cases. For example, in Sega v. Accolade, the court held that Accolade could lawfully engage in decompilation in order to circumvent the software locking mechanism used by Sega's game consoles Mar 14 13:08:31 wow lol Mar 14 13:09:16 In Europe, the 1991 Software Directive explicitly provides for a right to decompile in order to achieve interoperability. Mar 14 13:09:23 wikipedia has everything Mar 14 13:10:11 so if you were to decompile and replace strictly using the decompiled program to achieve interoperability you are fine :) Mar 14 14:01:35 for those peeping: http://privatepaste.com/c6e7b160c7 Mar 14 14:02:03 t-180m Mar 14 14:03:03 based on http://pastebin.com/sCJ4uw27 Mar 14 14:04:28 [GENERAL NOTICE] in 3h from now, all maemo.org services except tmo are expected to go down for a few hours Mar 14 14:48:58 hi, is there a "screen" for maemo/n900? Mar 14 14:50:06 iirc tmux was in some repo Mar 14 14:59:44 hmm, seems I have screen from somewhere, user doesn't seem to have permission to /var/run/screen though but it starts as root Mar 14 15:00:13 haven't messed with the phone in a long time now since it's been working so well :p Mar 14 15:14:50 hello Mar 14 15:20:59 can anyone recommend a GOOD webkit browser for Maemo? As alternative to MicroB to try out. Mar 14 15:21:07 I think I have tried them all, and they all had annoying display issues... Mar 14 15:21:26 like old pixels not being repainted when scrolling around, much more annoying than using white for not yet rendered areas Mar 14 15:29:25 ew, why webkit? :( Mar 14 15:31:32 why not webkit? Mar 14 15:32:25 isn't webkit from Apple? Mar 14 15:32:43 therefore, eew :) Mar 14 15:33:26 Even better, it's also from Google Mar 14 15:33:59 double eew Mar 14 15:34:22 hard trying to escape from apple and google :) Mar 14 15:35:16 jon_y: KTHML ;) Mar 14 15:35:18 seriously, though Mar 14 15:35:25 because I already have a good Gecko browser (MicroB) Mar 14 15:35:36 so I want to find out if a Webkit browser can perform better on the N900 or not Mar 14 15:36:01 I would _also_ like a tuned-for-speed ultrafast featureless browser Mar 14 15:36:14 I tried dillo for it, but it doesn't support drag-scrolling on the N900 Mar 14 15:36:48 maybe I should patch dillo for better N900 support... not sure how easy that would be though Mar 14 15:37:00 the main things needed are dragscrolling, and "right click by click and hold" Mar 14 15:37:17 also, virtual keyboard entry should work on location bar and input fields Mar 14 15:37:19 that's basically it Mar 14 15:37:49 some guy actually did something like it by emulating the middle mouse button for dragscroll using an external script... but that's ugly and can break in many ways Mar 14 15:38:04 e.g. when switching out of dillo and into something else while dragscroll is active :P Mar 14 15:38:51 which is BTW another idea I may want to do anyway on the N900... there is this mostly unused distance sensor thing, which can be used to emulate ONE extra mouse button Mar 14 15:39:04 by remapping mouse buttons via xmodmap when it's touched Mar 14 15:39:14 do we have a SECOND mostly unused sensor/button available, for middle mouse? ;) Mar 14 15:41:24 hm... for right mouse button I could probably live with something like ctrl-click, and use sensor-click for middle Mar 14 15:42:00 given most context menu options in dillo require keyboard input anyway (e.g. save as) Mar 14 15:50:51 divVerent: there's opera, for fast browsing Mar 14 15:59:10 kerio: not really faster - more like equal to MicroB Mar 14 15:59:14 Opera was indeed not quite bad Mar 14 15:59:27 faster for some stuff, slower for other stuff... but in the end slower for me Mar 14 15:59:37 e.g. it's quite slow for large simple HTML files (ebooks) Mar 14 15:59:43 illegal ebooks, that is ;) Mar 14 16:00:00 but, it has that "Turbo" proxy server feature, which is great when in GPRS or 3G Mar 14 16:00:33 maybe should check if I can somehow set up bookmark sync betwen MicroB and Opera, and change the browser switcher so that its default depends on 3G/Wifi :P Mar 14 16:35:43 waiting for godot, err Santa, err.... *the* DNS switch Mar 14 16:43:17 what repos should one be pointing at if one wants to track development of new stuff? (I get the impression that very little gets updated in the repos I'm currently using, so imagine developers are using repos elsewhere, or is there just very little development happening these days) Mar 14 16:45:44 there's just very little development i'm afraid Mar 14 16:46:11 DocScrutinizer05: why wasn't the TTL gradually lowered in this past week? Mar 14 16:46:18 * kerio flees >:D Mar 14 16:46:40 I imagine bcause RTT to hostmaster exceeds a week ;D Mar 14 16:48:07 kerio: no idea, ask nokia DNS-authority/hostmaster Mar 14 16:49:32 ah -- looking around some recent forum posts, it looked like it might be the case that some parallel repos were being set up, with less insurmountable barriers to entry -- did I get that wrong, or was that done, but too late for all the potential developers who've already wandered off to pastures new? Mar 14 16:50:17 fil: the problem are not the repos but the autobuilder Mar 14 16:50:55 and that one been botked since I think november Mar 14 16:52:55 http://talk.maemo.org/showthread.php?p=1293858#post1293858 Mar 14 16:55:35 oh, ok -- so all packages are rebuilt from source, except when the autobuilder is broken? -- oops Mar 14 17:03:35 ah, I see -- there are a load of things that apt-get update would like to install, but HAM doesn't mention -- what's the canonical way of keeping up to date these days then? (is there anywhere that answers these dumb questions that I should be reading instead?) Mar 14 17:04:00 packages are /built/ from source Mar 14 17:04:23 what I'm after is something like Debian's testing, or if that's not available Debian's unstable Mar 14 17:05:30 kerio: so people just upload new source packages and let the autobuilder build them for the first time? (I was rather expecting people to do a local build before upload) Mar 14 17:05:39 well obviously they test them locally Mar 14 17:05:54 anyone have existing scripts-settings for emulating mittle and right mouse button (e.g. for X11 forwarding use)? Mar 14 17:05:57 but the integrity of the binaries is only ensured by the fact that maemo.org does the building and keeps the source Mar 14 17:06:10 like, does X11 actually provide a nice way to do ctrl-click -> right mouse, or the like? Mar 14 17:06:43 I noticed that dbus-scripts can catch the camera button, which combined with xmodmap commands would give me ONE extra button. but I need two, and can't get dbus-scripts to work with proximityd as second button... Mar 14 17:07:07 kerio: hence "rebuilt" ;-) Mar 14 17:08:31 I see that xbindkeys can be abused to support ctrl-click... together with xvkbd Mar 14 17:08:52 divVerent: search the wiki for vi's awesome keyboard or something silly like that Mar 14 17:09:05 ~factsearch vi Mar 14 17:09:09 hm Mar 14 17:09:11 ~factinfo vi Mar 14 17:09:11 error: you do not have enough flags for that. (o required) Mar 14 17:09:11 vi -- last modified at Mon Feb 16 23:23:30 2004 by George-!~Test@adsl-167-107.freeuk.com; it has been requested 9 times, last by Deaod at Thu Jan 24 09:40:35 2013; it has been locked by vi. Mar 14 17:09:20 DocScrutinizer05: how do i search factoids? Mar 14 17:09:51 ~listkeys kerio Mar 14 17:09:53 Factoid search of 'kerio' by key (3): kerio #DEL# ;; oldtest kerioth ;; kerio. Mar 14 17:09:55 ~listkeys vi Mar 14 17:09:57 Factoid search of 'vi' by key (13 of 1075): airport vix ;; beandog's favorite movie ;; because travis_143cw ;; cervisia ;; cmd: keighvin (.*?) ;; corewar virtual machine ;; eview ;; fbview ;; jargon virtual friday2 ;; linux loving sluts ;; media forte sf16-fmd2 v4l device driver ;; of course, i assume this device ;; oldtest mahavites. Mar 14 17:10:04 hm Mar 14 17:10:08 ~listvalues kerio Mar 14 17:10:10 Factoid search of 'kerio' by value (2): oldtest kirioth ;; kerio. Mar 14 17:10:51 https://github.com/vahit/N900-HKlayout I did find this, but it does not emulate mouse Mar 14 17:10:58 ~listvalues vi Mar 14 17:11:00 Factoid search of 'vi' by value (19 of 9810): asynchronous communication ;; cd32 ;; domain name service ;; gene ;; graphics device interface ;; id2 ;; libgraph ;; mandrake ;; no minipci ;; oldtest nicodemus ;; pkgview ;; quake3 voodoo3 ;; rfc 1675 ;; rfc 647 ;; rfc994 ;; rsts/e ;; tuxtime ;; weathercode mvengue ;; x-rar. Mar 14 17:11:07