**** BEGIN LOGGING AT Wed Mar 18 02:59:57 2009 Mar 18 03:00:34 use uSD-image! Mar 18 03:00:47 Neo1973 Mar 18 03:00:56 ooops Mar 18 03:02:03 DukeOfURL: do you have a microSD card? Mar 18 03:02:23 yes, and I successfully downloaded to it. thanks. Mar 18 03:15:46 w00t Mar 18 03:15:52 managed to aquire me some wifi Mar 18 04:05:07 Hi. Which is the best list to put up suggestions/questions for Google Summer Of Code? Mar 18 04:30:50 need answer quick Mar 18 04:30:50 does gpl allow for commercial use Mar 18 04:31:03 I think so. Mar 18 04:31:54 But there are restrictions, of course. You can sell gpl'd code, but you must make the source available for free. Mar 18 04:32:37 You can obviously use it in a commercial enterprise. Mar 18 04:32:48 kk Mar 18 04:32:55 You can sell support for gpl'd code. Mar 18 04:39:07 u can sell gpl'd code too Mar 18 04:39:13 nothng wrong with that Mar 18 04:39:29 ooh wait Mar 18 04:39:33 u said that already Mar 18 04:39:54 Yes, but it means more coming from you. Mar 18 04:40:03 ahhaha no no Mar 18 04:40:08 i just saw your last line Mar 18 04:40:11 i didnt scroll back yet Mar 18 04:40:33 you are under no requirement to public gpl code or patches Mar 18 04:40:55 you ONLY need to provide them to your customer (or well offer them to the customer for at most the cost of media and shipping) Mar 18 04:41:05 the customer is free to do what they like with the code tho Mar 18 04:41:09 after that Mar 18 04:41:36 you are not allowed to restrict what the customer can do with the code after that beyond what the gpl already does Mar 18 04:44:19 What's the status with Tivo? Is the consensus that they are following the letter of gpl, while obviously violating the spirit? Mar 18 04:44:43 yes. Mar 18 04:44:55 (to be specific, GPLv2) Mar 18 04:45:04 yup Mar 18 04:45:05 they're obviously not following the letter of GPLv3 Mar 18 04:45:22 the letter of the gpl really isnt that "open" Mar 18 04:45:38 it doesnt require you give up that much - if you structure things correctly Mar 18 04:45:56 this is where we get to raster's crazy ideas on linking? :) Mar 18 04:46:12 ie put everything into closed kernel moduels that dont use any other kernel infra that gpl Mar 18 04:46:19 (basically nvidia-style) Mar 18 04:46:29 or put it into userspace and only release small kernel bits as gpl Mar 18 04:46:30 judges look at intent, so if it's obvious that you did something *just* to get around the GPL, you lose Mar 18 04:46:33 and avoid gpl libs Mar 18 04:46:34 etc. Mar 18 04:46:59 cjb: depends on jurisdiction Mar 18 04:47:09 there is a grey area Mar 18 04:47:21 there is a very EXPLICIT line between kernel and userspace Mar 18 04:47:31 and between glibc etc. Mar 18 04:47:36 and apps Mar 18 04:47:47 those are codified in the license Mar 18 04:47:53 yes Mar 18 04:47:56 there is an exception for system software Mar 18 04:47:57 thus the "explicit line" Mar 18 04:48:02 ah, right Mar 18 04:48:05 that is a problem there Mar 18 04:48:12 what is "system software" Mar 18 04:48:13 Can new releases of code that was origionally gplv2 be released under glpv3, thereby preventing bad actors from being able to keep up with the current community version? Mar 18 04:48:17 that is fuzzy grey land Mar 18 04:48:29 KenYoung: yup, usually Mar 18 04:48:30 as is shared memory Mar 18 04:48:31 eg Mar 18 04:48:43 KenYoung: but there's nothing stopping the bad actors from just using the old release Mar 18 04:48:44 if u have an external gpl process u talk to Mar 18 04:49:00 KenYoung: so, most of the GNU projects (emacs, gdb, maybe even gcc?) are now GPLv3 Mar 18 04:49:08 if you share "significant complex data structures" Mar 18 04:49:13 then you mayaswell be linking Mar 18 04:49:16 thus gpl applies Mar 18 04:49:24 raster: sure. that's where I think intent comes in. Mar 18 04:49:27 but wehat is "significant complex...." Mar 18 04:49:37 i can find the exact words int he gpl - but its something like that Mar 18 04:49:48 if you can't look a judge in the eye and say you weren't trying to cheat the GPL, you're probably going to lose a lawsuit Mar 18 04:49:56 raster: I think you're probably being overly legalistic Mar 18 04:50:05 i am Mar 18 04:50:07 hell Mar 18 04:50:12 i believe in the spirit of the gpl Mar 18 04:50:17 and i am all for followin g it Mar 18 04:50:21 but maybe that's just my experience with lawyers. very few of them just look at a license text and make decisions in a vacuum. Mar 18 04:50:30 screw arguing over meanings of words Mar 18 04:50:52 but - the problem is that in the end it comes down to a court and legalese Mar 18 04:50:57 they're going to say "and what were you thinking when you decided to make a shim to allow GPL access without linking explicitly?" and you're going to say "um." and they're going to say "you lose." :) Mar 18 04:51:10 true enough. Mar 18 04:51:30 well thats where i'd be citing examples Mar 18 04:51:46 "nvidia has done it - and despite controversy - no one has demanded they open up" Mar 18 04:51:48 blah blah Mar 18 04:51:52 not just intent Mar 18 04:52:03 but the other side of that Mar 18 04:52:04 too Mar 18 04:52:08 its a nasty world Mar 18 04:52:14 i'd rather never even talk to a lawyer Mar 18 04:52:29 just get the hell on with writing cool stuff Mar 18 04:58:52 raster: while it's true that nvidia has never been sued, the people who would sue them claim that it's just because they aren't interested in doing that, and Greg K-H says that every lawyer he's talked to thinks that nVidia is violating the license. Mar 18 04:59:01 the absence of a lawsuit does not prove compliance Mar 18 04:59:14 sure Mar 18 04:59:18 dont say it does Mar 18 04:59:22 if they were sued Mar 18 04:59:24 and nvidia won Mar 18 04:59:30 then it'd definitely prove Mar 18 04:59:33 * Sargun needs to upgrade to 2008.12... Mar 18 04:59:39 as an actual test case would show it Mar 18 04:59:48 but in the absence of one... Mar 18 05:00:10 one can only judge intent of the license holders by their actions Mar 18 05:00:17 in the absence, we have no evidence either way whether what they're doing is legal Mar 18 05:00:30 being fully aware of nvidia's efforts and not doing anything implies intent on their part Mar 18 05:00:33 intent goes both ways Mar 18 05:00:34 :) Mar 18 05:00:39 I think Linus has gone on record saying that he thinks nVidia is violating the license. That's his action. Mar 18 05:00:49 hmm Mar 18 05:00:54 i swear he went on record the other way Mar 18 05:01:03 He's also gone on record saying he doesn't believe in lawsuits, but that just means he doesn't see the point in wasting lots of money. Mar 18 05:01:04 with "binary kernel modules are ok" Mar 18 05:01:05 oh, huh Mar 18 05:01:13 that doesn't seem like something he'd say to me Mar 18 05:01:20 but I haven't talked to him about it Mar 18 05:01:24 i could have sworn he has Mar 18 05:01:40 why bother implementing EXPORT_SYMBOL_GPL(), if the license is irrelevant? Mar 18 05:01:42 which is why i'm just taking the position of "intent" there Mar 18 05:01:52 because by DEFAULT he thinks its ok Mar 18 05:01:53 BUT Mar 18 05:02:14 there are those that disagree and want to be very explicit about it and keep the likes of nvidia out of using their bits of kernel code Mar 18 05:02:25 so that is a technical solution to a disagreement on policy problem Mar 18 05:02:25 mmm Mar 18 05:02:28 as best i know Mar 18 05:02:50 ie the default is "you can use the symbol in a binary module... UNLESS its explicitly marked GPL" Mar 18 05:02:59 Why is it that these graphics hardware companies are so especially secretive? Intel doesn't try to hide the x86 architecture, as far as I am aware. Ditto every other general processor manufacturer I know of. Mar 18 05:03:04 and that got added aft6er there was much kerfuffle about it Mar 18 05:03:10 KenYoung: there are several theories: Mar 18 05:03:11 with linux disagreeing with others etc. Mar 18 05:03:31 1. we'd see how they cheat on artificial benchmarks with special cases in their drivers, and then they'd look dumber than usual Mar 18 05:03:34 i personalyl think nvidia should just have open drivers - all closed drivers do is create pain Mar 18 05:03:40 i am not into "open for open sake" Mar 18 05:03:45 2. they license code in their drivers from other companies who are more secretive than they are Mar 18 05:03:52 i'm into open because its frankly just a better way of "doing business" Mar 18 05:03:55 less hassle Mar 18 05:03:56 less pain Mar 18 05:03:59 less overhead Mar 18 05:04:06 raster: yeah, with AMD going the open way, nVidia's excuses are looking pretty tired Mar 18 05:04:13 yup Mar 18 05:04:20 so dont get me wrong Mar 18 05:04:25 i'm not "Defending" closed stuff Mar 18 05:04:30 yeah, I understand Mar 18 05:04:52 the other complication with nVidia is that it's not like Linus owns the copyright, the contributors do Mar 18 05:04:53 i just think that the line in the sand - given intent, license terms, fuzzy bits of licenses etc. lies in a certain spot Mar 18 05:05:04 so Greg K-H says he thinks it's illegal, and he owns a bunch of driver core Mar 18 05:05:08 i dont think people should totally take advantage of that though Mar 18 05:05:11 and push it to its limits Mar 18 05:05:20 there is fuzzy land and there is a spirit to uphold Mar 18 05:05:23 so uphold it Mar 18 05:05:36 though one thing i do disagree with is GPL libs Mar 18 05:05:48 here is a shared lib offering a public api someone is WANTING you to use Mar 18 05:05:49 cjb, So it's sortof like battery lifetimes, diaggonal measures of screens etc - the whole industry lies about what they provide, and all are afraid of being the first to market truthfully. Mar 18 05:05:56 (eg qt before it was LGPL) Mar 18 05:06:00 KenYoung: yes, this is one element Mar 18 05:06:05 thats a bit where i think there is a very clear dividing line Mar 18 05:06:14 and a documented API to divide it with Mar 18 05:06:18 KenYoung: note that this is not an argument against open specs, just open drivers Mar 18 05:06:28 so AMD has been publishing open specs and letting Xorg people write open drivers Mar 18 05:06:31 and that's working out fairly well Mar 18 05:06:48 cjb: well therte was a fair bit of lag to begin with Mar 18 05:06:50 I don't actually care if their driver is full of secret sauce if they'll tell me how to program the device Mar 18 05:06:52 that was not good Mar 18 05:06:55 yeah Mar 18 05:07:02 ie from psecs being open to them actually having real impact on drivers Mar 18 05:07:22 but given the driver core is now up to date - it may be better in future Mar 18 05:07:30 ie new chips, new specs, and support goes in much faster Mar 18 05:08:22 yeah, Intel usually pushes driver support for new chips before the chips are released lately Mar 18 05:08:33 Via is doing similarly Mar 18 05:08:37 (I mean video drivers) Mar 18 05:09:28 thats all good Mar 18 05:09:38 nvidia are just not playing ball Mar 18 05:09:47 though there are some problems with open specs Mar 18 05:09:54 some chips - implement ALL features Mar 18 05:09:58 patentend and more Mar 18 05:10:10 BUT the vendor only pays for the patents on the "higher end cards" Mar 18 05:10:36 and the driver is what polices the enabling/disabling of them Mar 18 05:10:38 not silicon Mar 18 05:11:07 and thus opening up specs fully would mean violating patent agreements and costing possibly a large chunk of cash to the vendor Mar 18 05:11:22 oh, yes, indeed! Mar 18 05:11:28 there is a good example of this in nVidia Mar 18 05:11:44 thus i can understand why a vendor might go "open specs/drivers/ hell no!" Mar 18 05:11:53 they designed their way into a corner and just cant be open Mar 18 05:11:59 the nVidia Quadro cards have features that are available on Geforce cards, but locked in the driver software when it finds a Geforce PCI ID Mar 18 05:12:18 so the nouveau driver, today, can turn a Geforce card into a Quadro, for at least some of the features Mar 18 05:12:27 this is another reason they avoid opening up drivers, yet :) Mar 18 05:12:30 s/yet/yes/ Mar 18 05:12:30 cjb meant: this is another reason they avoid opening up drivers, yes :) Mar 18 05:12:39 aah yes Mar 18 05:12:40 raster: here's the Linus quote: http://lkml.indiana.edu/hypermail/linux/kernel/0312.0/0670.html Mar 18 05:12:46 just ot keep their premium cards.. premium Mar 18 05:12:53 but thats something solvable Mar 18 05:12:59 in short, he thinks that binary kernel modules must be GPL *unless* they were written without Linux in mind, and then bolted on Mar 18 05:13:03 the problem is when its someone elses patents they have had to license... Mar 18 05:13:10 when they no longer are the ones in control Mar 18 05:13:31 that's an easy excuse to give, but the fact that AMD is opening up so much shows that it's not a big deal in reality Mar 18 05:13:49 are we really to believe that nVidia licensed huge amounts of code, but Intel and AMD/ATI didn't? Mar 18 05:15:56 hmm Mar 18 05:16:03 so actually linux does say binary moduels are ok Mar 18 05:16:19 it's an interesting argument, though Mar 18 05:16:21 its just required to be clearly not aderivative Mar 18 05:16:28 he says that if you aren't a derived work, the GPL is irrelevant Mar 18 05:16:29 in is view Mar 18 05:16:32 that shows his intent Mar 18 05:16:39 right Mar 18 05:16:44 or documents his intent and helps clarify some gray Mar 18 05:16:53 sure Mar 18 05:17:00 thats the real problem - the gray bits Mar 18 05:17:11 his "intent" can't affect what it and isn't a derived work, though Mar 18 05:17:23 the GPL doesn't get to make a claim about that; you have to go back to copyright law in general Mar 18 05:17:25 oh sure Mar 18 05:17:26 and yeah, it's very fuzzy Mar 18 05:17:41 but his intent as to what constitutes dervied Mar 18 05:17:50 he doesn't get to have an intent on that Mar 18 05:17:59 one could argue the nvidia drivers are shared between osx, windows and linux Mar 18 05:18:00 it is outside the scope of his power as someone licensing software Mar 18 05:18:13 right. I'm pretty sure that's how he feels about nvidia. Mar 18 05:18:17 they designed them originally for other os's but ported them with a open "shim" to interface to the kernel Mar 18 05:18:20 but its a gray land Mar 18 05:18:24 did they really do that? Mar 18 05:18:41 they certainly had a windows driver which they later ported to linux Mar 18 05:18:52 until it actually goes to court and is decided one way or another Mar 18 05:18:57 we will never know for sure Mar 18 05:19:11 yup Mar 18 05:19:11 also the question is Mar 18 05:19:16 if it starts as a windows driver Mar 18 05:19:19 This area of law will eventually make the Talmud look like a quick read. Mar 18 05:19:27 but over time "assumes" more about linux when in "linux mode" Mar 18 05:19:40 and begins to look like it is designed "for linux" Mar 18 05:19:53 just by virtue of growing into that position... does it then become a derivative? Mar 18 05:20:17 raster: yes, that's the question Mar 18 05:20:25 i'd just rather we never had the whole problem Mar 18 05:20:31 I suspect one judge would rule one way and a different judge would rule the other way Mar 18 05:20:31 and just had bloddy open drivers Mar 18 05:20:37 this tells me that it's a dangerous business to be in Mar 18 05:20:38 and everyone got on with doing cool stuff Mar 18 05:20:39 :) Mar 18 05:20:51 yup Mar 18 05:20:51 and that you should really be on the side of not probably breaking the law Mar 18 05:20:53 :) Mar 18 05:21:07 dangerous. its a risk. the question is - how do you bet your money on the odds Mar 18 05:21:08 really Mar 18 05:21:17 what are nvidia protecting Mar 18 05:21:28 is what they are protecting worth the risk? Mar 18 05:21:49 and the lost sales for Linux users Mar 18 05:22:07 at one of the Red Hat Summits they had Intel, AMD, and nVidia reps in the same place Mar 18 05:22:12 and the nVidia dude got booed out of the rooom Mar 18 05:22:46 and the people who go to those conferences tend to be CIOs, not just college kids. Mar 18 05:23:09 yeah Mar 18 05:23:42 but maybe nvidia go "really - we dont care about linux. the only peolpe we do care about are already happy in the linux world - and until they say otherwise.. we just dont care" Mar 18 05:23:48 and thats their business decision Mar 18 05:24:09 i do know that... for me. nvidia are on a blacklist for making hardware Mar 18 05:24:15 nvidia have an SoC of their own Mar 18 05:24:25 arm core with nvidia 3d core Mar 18 05:24:30 and i wont touch it with a barge pole Mar 18 05:24:37 not even going to read its specs. Mar 18 05:24:44 that SoC has OpenGL-ES, too, which is weird Mar 18 05:25:04 if you're going to put an nVidia core on a chip, you might as well make it do actual GL Mar 18 05:25:06 nvidia's policy of pretty much being a closed block puts them on my "dont bother" list Mar 18 05:25:13 yup Mar 18 05:25:16 gles is indeed a bit bizarre Mar 18 05:25:22 i'm not totally happy with it Mar 18 05:26:07 but nvidia's policies on driver openness influence potential customers Mar 18 05:26:29 mind u.. i'm not happy with imgtec either. Mar 18 05:26:35 yeah, I think so too. but they probably don't realize what they're missing. Mar 18 05:27:55 that's their problem Mar 18 05:28:03 their bottom line will be the one suffering in the end Mar 18 07:30:00 cjb: Hey :) Mar 18 07:30:12 cjb: i have something to say about openness of AMD/ATI. Mar 18 07:35:17 cjb: This mailing list post by one of the core radeonhd developers show how "open" AMD/ATI are now (and that's more than a year and a half since they decided to provide the specs): http://lists.opensuse.org/radeonhd/2009-03/msg00112.html Mar 18 07:38:00 Looks like stupid USA software patent legislation contributes more than anything else to the shit going on with "opening" the specs. Mar 18 08:17:35 good morning Mar 18 08:31:53 good morning Mar 18 08:36:47 eh, nitght here Mar 18 09:35:47 how can i append to QI when booting from nand? Mar 18 09:37:43 Zoup: append to qi? Mar 18 09:41:04 lindi-: as in " glamo_mci.sd_max_clk=15000000 " issue Mar 18 09:41:30 Zoup: so you want to pass parameters to linux? Mar 18 09:41:43 lindi-: yep, thats right way to say that Mar 18 09:42:14 lindi-: its passing to linux indeed :) any idea how its possible using QI? Mar 18 09:42:30 lindi-: since its not SD boot, /boot/append ignored Mar 18 09:42:33 Zoup: i haven't read qi source yet, got alink? Mar 18 09:43:00 lindi-: sure Mar 18 09:43:17 lindi-: http://wiki.openmoko.org/wiki/Qi, http://git.openmoko.org/?p=qi.git;a=summary Mar 18 09:44:01 git clone git://git.openmoko.org/git/qi.git Mar 18 09:44:04 yeah.. Mar 18 09:44:04 Zoup: only by modifying the Qi source i guess Mar 18 09:44:52 PaulFertser: i see, thanks :) Mar 18 09:45:00 Zoup: or you can boot the kernel from sd, but pass root= to use nand rootfs. Mar 18 09:45:15 i wonder why glamo_mci.sd_max_clk issues happens on some SDs Mar 18 09:45:50 Zoup: some people reported that sd issues is a result of gsm interference. Does it look like your case? Mar 18 09:46:35 PaulFertser: its always working with some Sandisk 4GB Mar 18 09:48:42 Zoup: there's another idea which was unconfirmed but looks like a possible explanation: laptop cardreaders use higher voltages than FR. The FR's driver asks the card about the minimal voltage and uses that, unlike sdhci. Mar 18 09:49:30 PaulFertser: can you see any software fix to this issue ? Mar 18 09:49:44 PaulFertser: Zoup: i read the source and it seems to be the case Mar 18 09:49:52 that you need to modify source Mar 18 09:50:18 Zoup: if it was really the case, it could be easily fixed to use a higher voltage (add another kernel parameter). But so far nobody has confirmed it can actually help. Mar 18 09:50:44 Zoup: moreover when GPS is turned on, the maximum clock is lowered drastically, by 8 times iirc. Mar 18 09:50:57 PaulFertser: i see... Mar 18 09:54:07 isn't it just the slope of the clock? Mar 18 09:54:14 the drive strength Mar 18 09:56:32 shellevil: it's not, read the source Mar 18 10:13:57 what's the current status of openmoko usability? Mar 18 10:15:00 proyvind: openmoko? do you mean the OM-distribution? Mar 18 10:15:10 proyvind: toy for hackers, imho Mar 18 10:16:39 k, I've not been reading FR source - not having one PaulFertse Mar 18 10:16:45 well, I was first of all thinking more like if I'll order a neo freerunner now, will it satisfy my primary need for calendar, alarm clock, calling and sms? :o) Mar 18 10:16:49 no Mar 18 10:17:05 well - maybe - if you pick _just_ the rright image, and play witth it a lot Mar 18 10:17:15 and if you don't have buzz problems Mar 18 10:17:19 but basically - no Mar 18 10:17:23 shellevil: you can use http://git.openmoko.org web interface to see glamo-mci driver source, it's quite easy to understand. Mar 18 10:17:26 proyvind: calendar and alarm clock are not the problem, calling and sms are. Mar 18 10:17:33 PaulFertser: thanks Mar 18 10:17:42 so it's not usable for sms and calling¿? Mar 18 10:18:00 proyvind: not reliably. and not in an acceptable quality. Mar 18 10:18:15 proyvind: for calling you'll need a hardware buzz fix sooner or later. I guess sms works mostly too. Mar 18 10:18:36 proyvind: that said, i use it as my only cell phone from November. Mar 18 10:18:42 k Mar 18 10:18:52 proyvind: but i don't care much if it doesn't work. Mar 18 10:18:58 hardware buzz fix? Mar 18 10:19:05 what do you mean by that? Mar 18 10:19:07 Oh come on. Mar 18 10:19:26 jesperj: change one 0402 R and solder a big cap near it. Mar 18 10:20:00 jesperj: http://people.openmoko.org/joerg/GSM_EMI_noise/ Mar 18 10:20:07 so I will be able to communicate with the rest of the world then Mar 18 10:20:25 proyvind: not reliably he just said Mar 18 10:20:39 thanks PaulFertser and henk Mar 18 10:21:13 reliability is for pussies Mar 18 10:22:18 proyvind: ack. Or for those addicted to things, not understanding that things doesn't matter. Mar 18 10:23:10 (in the context of regular phone usage, of course; i'm not talking about life-critical applications, where reliablity is a necessety) Mar 18 10:24:24 yeah Mar 18 10:28:10 hm Mar 18 10:28:29 but I can use the qtopia stack on the freerunner to have a fully working cell phone, right? Mar 18 10:29:24 you will still have the hardware shortcomings... Mar 18 10:29:49 and qtopia is not a "good choice" imho Mar 18 10:30:04 proyvind: afaik it does not use x11, so most applications won't run Mar 18 10:30:26 henk: yeah, but it would still be working pretty much as a "regular phone" I mean Mar 18 10:30:33 hardware shortcomings other than buzz? Mar 18 10:30:59 i'm not sure if there are others... Mar 18 10:32:18 bass is poor with external headphones Mar 18 10:32:44 proyvind: the case that requires enabling quite harsh echo suppression. Mar 18 10:33:31 proyvind: very slow SD _and_ video memory bus. Coupled with VGA screen the performance plain sucks. Mar 18 10:33:32 I use the Freerunner with the OM2008.12 platform. I haven't been bothered from buzz, and has reliably sent sms and placed and received calls, without any problems since last september. Mar 18 10:33:59 Daramarak: you're not supposed to hear buzz. It's the other side who usually complains. Mar 18 10:34:07 Daramarak: and it depends on the band and transmit power. Mar 18 10:34:19 Daramarak: how many times that needs to be told? Mar 18 10:35:17 proyvind: oh, and wifi firmware really sucks. Not reliable and incompatible with many APs. Mar 18 10:35:32 PaulFertser: hav e you confiirmed that buzz is a problem on all bands? Mar 18 10:36:08 proyvind: calypso gsm modem has a "recamping" bug, that is workarounded by disabling deep sleep mode (hurts battery life). Mar 18 10:36:12 PaulFertser: If I am not bothered by it, (which means that I haven't had any complains). It goes to show that some use it without any problems. But you MIGHT experience some issues. Mar 18 10:37:16 shellevil: ask DocScrutinizer or write an e-mail to Dieter. I'm almost sure that it's possible to have buzz on any band given you're in a bad enough conditions where transmit power has to be really high. Mar 18 10:38:09 Daramarak: you can't rely on the phone if it will make communication hard or impossible in some not-that-ideal conditions (read far away from towers, or shielded etc). Mar 18 10:38:35 PaulFertser: just wondering - as some carriers use only one band, and that'd be a tidy explanation as to why no buzz. Mar 18 10:39:25 shellevil: i prefer another explanation: good coverage for some/most users. Mar 18 10:42:01 that works too Mar 18 10:42:02 shellevil: i think it's easy enough to test: get an unfixed freerunner, establish a call, make sure sound is transfered, start to shield FR thus forcing it to ramp up power. Monitor %EM output via ssh (to know which band is used). Monitor sound on the other side. So far i heard no reports of this simple experiment, so i guess it's reasonable to ignore those who said that buzz doesn't exist on some carriers. Mar 18 10:42:31 PaulFertser: Well, it isn't that black and white, given poor conditions any phone might misbehave. I live in a country where coverage might be very poor due to the terrain, and my FR performs better than some commercial phones, while some outperforms mine. It is a matter of degrees of reliability. Mar 18 10:43:37 Daramarak: it's a matter of stupidity of TPE EEs, who failed to test it in the first place and failed to properly fix it (i very much hope that Doc's fix of adding 3 beads in A8 will work) promptly. Mar 18 10:45:03 PaulFertser: I agree. Mar 18 10:46:30 shellevil: but really, why nobody's tried that so far? I think to try it myself i'll need to desolder wire from the cap but i can test only 2 bands and i don't know yet how to force it to use only one of those (disable switching temporarily). Mar 18 10:47:36 pau: I think you _drastically_ overestimate the number of people willing to try even simple things with hardware. Mar 18 10:47:42 shellevil: moreover i had complaints from the other side when i was outside of the city, so i see no point in proving it (except to prove that the experiment is a good test). Mar 18 10:47:47 For some reason, people seem unwilling to do sane stuff Mar 18 10:49:08 shellevil: getting some aluminium foil (used for cooking) is easy and using it to shield the whole device is not hard too. I guess i'll leave that to those who says he has no problems with buzz (as i clearly had). Mar 18 10:52:17 shellevil: i've read about so many cool hackers that i just get obsessed when i hear questionable information from people who can't even understand a simple sh script. Mar 18 10:52:50 shellevil: i'm not a cool hacker myself, but at least i usually can read! Mar 18 10:53:35 I can only explain it as people are unwilling to venture outside of their area of expertise. Mar 18 10:54:01 shellevil: and to me it'd be much better to communicate with those from whom i can learn. Mar 18 10:54:47 are the tx and rx buffers the same? Mar 18 10:54:48 oops Mar 18 10:54:50 SpeedEvil: i'm not an expert in anything. What should i do now, lay down and moan until i die of food/water deprivation? Mar 18 10:54:56 :) Mar 18 10:55:53 SpeedEvil: or probably i should stop reading and writing because my native language sucks and my english sucks even more? Mar 18 10:56:40 if so, the problem reduces to send 64 bytes from the serial interface - unless there is not 64 bytes within a timeout, then send that. Then monitor for a nak - if you get a nak, tell the hardware to resend the packet. If you get an OK, move on. Mar 18 10:56:42 meh Mar 18 10:56:55 * SpeedEvil hates doing that Mar 18 10:57:35 If you carry on learning stuff, it can't be bad I figure. Mar 18 10:58:09 SpeedEvil: i'll never be able to ride like a real sportsman, should i stop cycling then? Mar 18 10:58:27 i have a stupid question : http://wiki.openmoko.org/wiki/FreeRunner_unable_to_work_with_3G_SIM_cards : it means that with 3G non working cards you have nothing and with working cards you have GSM standard mode, or do you get 3G ? Mar 18 10:59:18 nullix: no 3G in fr. only "regular and slow" gprs Mar 18 10:59:35 SpeedEvil: i think that being an expert is not required for anything. You see something, you want to understand something, you learn something. Talking about that on irc is not enough, you have got to learn. An unwillingness to do any simple shielding test is ... I don't know how to explain it. Mar 18 10:59:43 nullix: my sim is 3g enabled and i can use gprs and gsm. Mar 18 11:00:39 but is is a hardware limitiation that 3g is different from gprs and can't be supported by freerunner or an operator issue ? Mar 18 11:00:50 nullix: hw limitation. Mar 18 11:01:13 then basically we can tell freerunner does not support 3g. Mar 18 11:01:24 nullix: the patent fees would probably have been totally overkill afaik Mar 18 11:01:27 nullix: but it supports 3G-enabled sim cards. Mar 18 11:01:27 nullix: afaik Mar 18 11:01:53 nullix: so if your operator has GSM/GPRS coverage, it will work. Mar 18 11:02:43 thanks, was no so clear in wiki, or i didn't read right page. Mar 18 11:02:51 nullix: so far there were no reports about really incompatible sim cards. If the card works in old GSM phone, it will work in freerunner too (provided you upgraded GSM firmware). Mar 18 11:05:07 nullix: i guess the page you mentioned contains outdated information that is no longer correct after we can easily reflash gsm firmware with the fixed version. Mar 18 11:19:02 i cant get original minisd card working even with glamo_mci.sd_max_clk trick Mar 18 11:21:29 Does anyone ever seen a credible report of _any_ good (read working in other GSM phones) simcard not working after calypso FW upgrade? Mar 18 11:28:38 nullix: Added this text to the page, thanks: "So far there were no credible reports of any issues left with any sim cards after upgrading GSM firmware. So, basically, if it works with any old GSM phone, it will work with GTA01 and GTA02 too after firmware upgrade." Mar 18 11:39:36 Paul: I tried the FW upgrade last night and it wouldn't go. The GSM still works, however. Mar 18 11:39:49 gta01b4 Mar 18 11:42:12 DukeOfURL: Details? Mar 18 11:44:36 didn't have enough free space to load calypso, so I put it on the SD card. Mar 18 11:44:54 followed the instructions exactly, but kept getting a timeout message Mar 18 11:45:48 DukeOfURL: timeout? Where? Mar 18 11:46:13 DukeOfURL: fluid.exe could connect to bootloader, but you got timeout afterwards? Mar 18 11:46:38 DukeOfURL: and you're sure you really disabled anything that could access /dev/ttySAC0 before starting fluid? Mar 18 11:46:54 no, not sure Mar 18 11:47:49 DukeOfURL: so, what did you do prior to starting fluid to ensure nothing will interfere? Mar 18 11:47:51 i got the 'Bootloader: (reset target)' message, sent the 's3c24xx-gpio b7=1', and then got the timeout Mar 18 11:49:25