#qi-hardware IRC log for Thursday, 2010-11-18

wpwrakhehe, antenna fixed ;-) it was a bad solder joint. ant-120B4 now looks like its "good" twin, ant-120A. (both suck, but that's okay) http://downloads.qi-hardware.com/people/werner/tmp/ant120-fixed.png00:14
wolfspra1lnice work00:18
wpwraktrial and error at its best ;-)00:18
wolfspra1lI'll add a thumbnail of this pic to the wiki homepage slider - we don't have any representation of antenna work yet, and this one looks nice :-)00:21
wpwrakhmm, so i better find an antenna that actually looks good quickly :)00:22
wpwrakwhat i'm doing is that i'm making a series of antennas with the size varying in steps of 10% from 90 to 120% of the reference design. i hope there'll be a good one among them, or at least a clear trend where i should look for the right size..00:24
wpwrakrafa: hmm, how does jlime-pkg like packages for which multiple versions exist ? e.g., man-pages*, matchbox-common*, rxvt, xserver-kdrive-fbdev00:35
wolfspraulwpwrak: do you have any tool recommendation off the top of your mind to convert dxf to an 'easily viewable' 2D representation like png/jpeg or so?00:38
rafawolfspraul: http://www.vialibre.org.ar/wp-content/uploads/2010/07/guia.dominio.publico.pdf00:39
rafawolfspraul: sorry the link is a spanish document..00:39
rafawolfspraul: we trust (arg linux peoñple like wer or me) trust at our politics00:39
rafafrom fundacion via libre00:39
wpwrakwolfspraul: hmm, qcad should be able to read DXF00:39
rafawolfspraul: they work for free content and free technologies on many areas for the argentine legislation00:40
wpwrakwolfspraul: and it can export to PNG (just looked it up :)00:40
rafawolfspraul: that documents is from beatriz busaniche.. an expert on thos matter00:40
rafawolfspraul: the document says that we are okey to use public domain stuff without to ask to use it, but it also says that00:41
rafawolfspraul: if we use it for commercial we need to pay something. I am not going to use any stuff of that because I still do not do business with this kind of stuff00:41
rafawolfspraul: just checking to learn a bit about argentina crap laws00:42
wolfspraulrafa: just make sure this really applies to you00:42
wolfspraulit's so easy to get distracted, scarily easy almost00:43
rafawpwrak: I guess that there is just one version of every package.. and packages with different versions should be called different.. something like vim should be a a package and vim3 should be a vim package of another version.00:43
wpwrakrafa: political idiocy at its finest00:43
rafawpwrak: but i am not sure if that is like I am telling you, maybe i am wrong..00:43
wolfsprauland sometimes we may well be adding more oil into the fire that we claim to want to put out00:43
wpwrakrafa: (packages) the examples above are from the jlime installation for qi-hw00:43
rafawolfspraul: no.. for me is okey.. I do not do any business at all00:43
wolfspraulthe question is even if you would, who would really come forward, and what the consequences and alternatives would be00:44
wolfspraulbut even that can be investigated once the situation becomes real00:44
wolfspraulno time for this :-)00:44
rafawolfspraul: yes, of course.. but then I could apply the same ideas to use another problematic stuff until some real problem appears.. better if we do not use any of them00:46
rafawolfspraul: that is a real thing (the law or argentina resolution)00:46
rafawolfspraul: we=me.. argentine00:46
rafain argentina00:46
rafawolfspraul: i mean.. sorry, my english right now sucks... another example.. I could do business with mpeg* without pay anything.. until some real problem appears.. Well, no I do not want to do that00:48
rafawolfspraul: and the public domain thing looks similar from beatriz document00:48
rafawolfspraul: I can ask her anyway00:48
wpwrakrafa: raises the interesting question of who would have to pay when you order a device containing public domain material from abroad for your own use00:49
wolfspraulmy first suspicion would be it's yet another case of engineers reading legal texts :-)00:49
rafawpwrak: no for your own use.. I mean.. if I sell stuff with public domain inside for example00:49
wolfspraulwhich cases was this law meant for? and where and how is it currently enforced?00:50
wolfspraulI would focus on that, if I had to...00:50
wpwrakwolfspraul: it looks like a pretty neat concept. they apparently replaced public domain with a transfer of the rights to a state entity.00:51
rafawolfspraul: anyway.. that was (the document about public domain stuff for business in argentina) just a comment.. it does not scare me or something :) ... and like that maybe there are other crap things in the nice argentine lays00:51
wolfspraulsee that's what I'm saying - we read things into this. but if we are interested, we should be interested in the practical consequences.00:51
wolfspraulnot in how much we can read into those texts...00:51
wolfspraulwhen you see a device (say the Ben NanoNote), the #1 question is whether you are actually selling the content on it, or the device itself?00:52
wolfspraulwho determines the 'value percentage'?00:52
wpwrakwolfspraul: it may be a problem for a reseller. not sure if it's a problem for a private importer.00:52
wolfspraulmaybe 90% is for the physical hardware, 10% for the content it's shipped with?00:52
rafawolfspraul: if I do some business with open source or CC technology we know where to ask for consequences ;) (fundacion via libre).. so I am fine until I need to do business.. and I will ask them :)00:53
wpwrakwolfspraul: the exact figures would be argued over in court for the next few decades :)00:53
wolfspraulthis is also why it is still questionable whether/how the cc-nc clause is enforceable at all, afaik00:53
wolfspraulif you sell a usb stick with 3gb of -nc content on it, what are you selling?00:53
wolfspraulthe usb stick or the content?00:53
wolfsprauland so on00:53
rafame.. tired.. I need to sleep. have a nice day.. cya00:54
wolfspraulI'm not paid for this kind of stuff, and not interesed. Lawyers get paid for every sentence they write, so I understand they are chatty :-)00:54
wolfspraulyeah :-)00:54
wolfspraulthat's better!00:54
wolfspraulget some new energy00:54
wpwrakwolfspraul: you just have to admire the idea - it's a complete perversion of the concept that intellectual property rights eventually expire !00:55
wolfspraullike I said, my first suspicion is we are reading things into it00:55
wolfspraulask someone working in the field what is really applied, and how00:55
wolfspraulif human being A has money, human being B will start to develop an interest in that money :-)00:56
wolfspraulI'd say those are some super basic 'eternal' laws and patterns...00:57
wpwraksure. just never heard of anyone going after the public domain by just installing a catchall. that's pretty clever, in a perverted way.01:00
kyakkristianpaul: quite strange...01:53
kyakkristianpaul: the problem is, i can't reproduce this error anywhere01:53
kyakxiangfu: regarding the latest image.. there are several things:01:55
kyak1) trigger happy won't build01:56
kyak2) the gmenu2x brightness can't be adjusted01:56
kyak3) nanoterm is not usable01:56
wpwrakwolfspraul, rafa: the origin of this appropriation of the public domain seems to be the protection of folklore. while, according to google, virtually unknown in the english-speaking world, the concept enjoys some popularity in latin america: http://www.educ.ar/educar/lm/1188223038496/kbee:/educar/content/portal-content/taxonomia-recursos/recurso/a2073f97-25e4-4cf6-b3a8-b390707f1655.recurso/6fd44608-dcbf-4210-974e-3155ec5bdd47/derechos_01:57
wpwrak(lovely URL :)01:57
wolfspraulwpwrak: so how is it applied? who pays? who determines how much? which 'public domain' content exactly is affected?02:01
wolfspraulmaybe if we would make a collection of traditional (public domain) Argentine folklore songs, and publish a CD, there would be some kind of tax on that CD?02:02
wolfspraulbtw, I could definitely imagine similar legislation elsehwere, there is a lot of debate in some countries about 'culture flat rate' etc.02:02
wpwrakwolfspraul: e.g., things where the copyright has expired. not sure if it would also apply to things that are put into the public domain abroad, e.g. public works in the US.02:02
wpwrakwolfspraul: it does apply to everything they can get their hands on, irrespective of origin02:03
wolfsprauland storage media are taxed in many countries now, empty CDs, empty USB sticks, etc.02:03
wolfspraulso one could definitely imagine that someone would take the Ben NanoNote because it is a 2GB storage medium02:03
wolfsprauls/would take/would tax/02:03
wpwrakthe text i referenced mentions that they're thinking about such things also in europe02:03
wolfspraullike I said, taxation for storage media already exists in many countries02:04
wpwrakyeah, it could reproduce radio -> GEZ :)02:04
wolfspraulempty CD-Rs are expensive in some countries :-)02:04
wolfspraulI'm not sure any of this affects what 'we' (copyleft hardware) should deem 'inclusive' for our hardware02:04
wpwrak(different sort of tax but still weird)02:04
wolfspraulI feel pretty good following the wmf, in general02:05
wolfspraulthere are special taxes for many categories of devices now, in many countries02:05
wolfspraulprinters are taxed as a 'flat rate copyright' fee, because they are assumed to be printing copyrighted material02:05
wolfspraulxerox as well02:06
wolfspraulstorage media are taxed under 'culture flat rate' regimes, who knows where the money goes :-)02:06
wpwrak(who would have to pay) they mention the importer. not sure if it would apply to everyone in the chain (e.g., if you have importer, distributor, and maybe some public performance) or just the last element.02:06
wolfspraulsometimes things are just taxed because there is or was a strong lobby to do so, for example the famous 'video cameras' in Europe, because the (now defunct) European video camera makers were hoping in the 60's it would help them stop the Japanese02:07
wolfspraulthen when the digital still cameras started to have video features, everybody was wondering whether they would be taxed this nice 20+% or whatever as well02:07
wolfspraulthen the smartphones02:07
wolfsprauletc. etc.02:07
wpwrakyeah, there are plenty of "because we can" taxes and para-taxes in that sector ...02:07
wolfspraulI'm not following all of this.02:07
wolfspraulI think there is a big difference between that kind of tax and a patent owned by someone like Sisvel.02:08
wpwrakyou're not an importer :)02:08
wolfspraulmassive difference :-)02:08
wpwrakthe taxes tend to be less intrusive02:08
wolfspraulyes sure, the importer pays the import duties, although selection of the right tax category is a joint duty, I am very well part of that02:08
wolfspraulthey are applied equally02:08
wolfspraulpatents can be withheld, and (some) patent owners do not refrain at all from the most draconian measures they can get away with under the law02:09
wpwrakyup. same thing for "regular" copyright.02:10
wolfspraulso I don't think any of this type of public domain taxation, culture flat rate, storage capacity taxation etc. influences our decisions or selection of content or software much02:11
wolfspraulwhereas patent, copyright or trademark infringement would02:11
wpwrakparaguay has an interesting model. there, the "public domain payant" must not cost more than the private royalties. i wonder what that would mean for things under the GPL, CC-BY-SA, etc. would they apply that work's historical record or invent some "standard value" ? well, should be fun for the courts in a few decades.02:12
wpwrakuruguay seems to have the oldest and most draconian model. there, the state basically becomes the author and can actively "defend" the author's rights.02:14
wpwrakwell, maybe they'll introduce something similar in germany soonish. after all, there's a lot of nazi material they want to keep locked away, and they presently use copyright for some of that.02:15
wolfspraulwpwrak: looking at qcad... how useful do you think it is?02:19
wolfspraulwhere do you see it in the bigger free software toolchain and for what type of work?02:20
wpwrakqcad .. it seems to be the principal free 2D cad02:20
wpwraki don't particularly like it, though. seems to be very clumsy to use.02:21
wpwrakwhen i had to do some drawings for my apartment, i tried to do them with qcad. after a day of fumbling, i gave up in frustration and did them with fped.02:22
wpwrakof course, fped has the home advantage in this case ...02:22
wpwrakah yes, qcad isn't parametric. so change one value and you get to redo everything that depends on it.02:23
wpwrakalso, there's a commercial and the free version. the latter lacks some features. (don't remember which)02:24
rohwpwrak not for long anymore02:24
wolfspraulI'm wondering which tool roh used to create the DXF files that he then used for laser-cutting the Milkymist One case02:24
wolfspraulahh, there he is :-)02:24
rohthere is somebody working on qcad. search for caduntu02:25
rohfirst step was making it qt402:25
wolfspraulin the .dxf files, I see some references to c:/program files/autocad 2002/ etc. but that may be just there for compatibility...02:25
rohi used qcad in the free version.02:25
wpwrakroh: "not for long" what ? :)02:25
rohwhat came packages with ubuntu02:25
wolfspraulroh: btw, excellent work on that case!02:25
rohwpwrak qt4 is so much better than qt3, alone from the looks02:25
wolfspraulI cleaned up one picture a little, included it in the picture slider on the qi homepage, the 12/1 community news, etc.02:26
rohn.p. i was quite tired and just wanted some shots02:26
rohand the light was bad02:26
wolfspraulindeed :-)02:26
wolfspraulwould you be interested to do a bit more work on this? I could pay for the costs.02:27
wolfspraulSebastien said he (or you) wanted to try to make one case with acrylic material.02:27
wolfspraulalso there is the issue of the microphone02:27
wolfsprauland the 3 buttons02:28
wpwrakroh: (qt) okay, but will also something in the functionality of qcad change ? i see better CAM support being mentioned. that's nice ... but we can do that with heekscad, too.02:28
wolfspraulif we improve this a little more, maybe we have a good enough case to make 30-40 for our upcoming run02:28
wpwrakacrylic tends to be fairly expensive. but perhaps that doesn't matter so much in this case.02:30
wolfspraulI have one question for the current wood case, I can't really tell from the pictures. is the top wood part only lying down on the side walls? or are there screws to hold it down?02:30
wolfspraulfor cost, this case took 12 minutes on the cutter, that's 12 * 1 EUR = 12 EUR02:30
wolfspraulplus material (wood), plus roh's work (priceless :-))02:30
wpwrakthere was also the manual groove-making :)02:31
wolfspraulI'm far away from raumfahrtagentur, so I cannot do much more right now than to say this is awesome work, offer some money, and encourage people to continue.02:31
wolfspraulwhen I'm in berlin, maybe I can dive into this myself as well, if raumfahrtagentur would be in Beijing I would be there already now :-)02:31
wpwrak(groove-making) but i'm sure that can be automated with the CNC mill.02:32
wolfspraulyes true, that's another thing that could be improved [groove]02:32
wolfspraulso roh - what do you think?02:34
wpwraki think the structure is just a sandwich with long screws pushing the top/bottom plates onto the sidewalls. maybe there are spacers to carry the load, in which case the sidewalls would "float" (more or less)02:34
wpwraki think that case design is pretty ideal for this sort of board. it has lots of connectors, which makes it hard if not impossible to find a pre-made box of the right size02:36
wolfspraulsebastien thinks (if I remember correctly) that the whole thing is not stable enough02:36
wolfspraulso that was one big reason why he wanted to try acrylic02:37
wolfspraulmaybe we can get roh onto this a bit more, like I said I need cases anyway and could pay for the costs etc.02:37
wpwrakhow big is the board ?02:37
wolfspraul10x15 cm02:37
wpwrakplywood should be okay for something that small02:38
wpwrakbut if there are no spacers (yet), then it may be a bit fragile02:38
wolfspraulwrong, sorry. 12.5x15cm02:39
wpwrakroh: so, are there spacers ? :)02:39
rohnope. but its intended to fit with 5mm spacers02:40
rohbelow the pcb02:41
wpwraknone on top ?02:41
rohnot yet. one could use these metal spacers with machine screws02:42
wolfspraulroh: are you planning any more work on this?02:43
wpwraknot sure what you mean. but spacers come in all sizes, so there ought to be something that fits. just as long as the thread is long enough02:43
rohwolfspraul i could. but i'd need a pcb ;)02:44
rohwpwrak my initial idea was 'spaces and nice imbus-screws which also work as 'feet'02:47
wpwrakroh: yup, that would be nice and tidy02:49
wpwrakroh: if you can't find the right spacers, you could also use a piece of pipe (of whatever material you have around) cut to the right length and terminate it with washers02:51
wpwrak(and a long screw through all this)02:51
wolfspraulkyak: hetzner introduced a new virtual root server, going down to 7.90 EUR / month02:55
wolfspraulfor running an irc server, that should be perfectly fine02:56
xiangfukyak, about gmenu2x brightness, forget git push. sorry03:03
xiangfuforget to update the package makefile in openwrt-xburst.git03:04
qi-bot[commit] Xiangfu Liu: update gmenu2x http://qi-hw.com/p/openwrt-packages/0a2704c03:04
rohwolfspraul i think irc ports were filtered or so at hetzner.. need to check03:06
kyakwolfspraul: hey, thanks for information, i remember seeing it last time. Will ask them if the allow IRC03:07
kyakxiangfu: np! thanks you03:07
xiangfukyak: the triggerhappy package error is strange. :(03:08
kyakit's upstream already, so we can't fix it right away..03:08
kyakxiangfu: i liked fn-handle :) small and nice, doing it's job03:09
xiangfukyak :) but it's way too simple :)03:10
kyakxiangfu: what use cases for triggerhappy do you have in mind?03:10
xiangfukyak: my package version is triggerhappy-0.1.3. and md5 is 274807068ffa73166c7024f6334ac45a. maybe we compare the package03:10
kyak$ md5sum dl/triggerhappy-0.1.3.tar.gz03:11
kyak19bc56dfff83e41431af1063f7cda926  dl/triggerhappy-0.1.3.tar.gz03:11
xiangfukyak: we can set some hotkey.03:11
kyakxiangfu: you need to rm your triggerhappy-0.1.3.tar.gz -\03:11
xiangfukyak: then it's different.03:11
kyakthe same old problem - updated git commit, nto updated packages version03:11
kyaktherefore, openwrt won't checkout from git but wil use dl/03:12
kyakxiangfu: i know about the hotkey, what i can't think of any use for it :)03:13
xiangfukyak: 1. delete   dl/triggerhappy-0.1.3.tar.gz , 2, make package/triggerhapply/compile V=99, compile fine.  can you make a diff, this is mine. (just download )http://downloads.qi-hardware.com/people/xiangfu/tmp/triggerhappy-0.1.3.tar.gz03:15
kyakone sec03:15
kyak$ md5sum dl/triggerhappy-0.1.3.tar.gz03:17
kyake1e4b403ec594e39dca649ce19b6f230  dl/triggerhappy-0.1.3.tar.gz03:17
kyakwhat the hell03:17
kyakdifferenet md5sums every time :)03:17
kyaklet me nwo get your triggerhappy-0.1.3.tar.gz03:17
kyak$ md5sum dl/triggerhappy-0.1.3.tar.gz                   52b2f58ddb4d543ad3bc4731d284faf1  dl/triggerhappy-0.1.3.tar.gz03:18
kyakxiangfu: your pacakge..03:18
kyaksame error03:18
kyakoh, wait a second..03:19
wolfspraulroh: why would hetzner filter irc ports? I mean I believe you but that sounds crazy to me.03:21
kyakxiangfu: not working... i suggest a clean test:03:22
kyakrm -rf dl/triggerhappy-0.1.3.tar.gz && make package/triggerhappy/{clean,compile} V=9903:22
kyakthen if it works for you, we next assumption is missing dependency03:23
kyakevtable_KEY.h:244: error: 'KEY_RFKILL' undeclared here (not in a function)03:23
kyakthis is what i get.. and following errors03:23
xiangfukyak: oh. look into the triggerhappy makefile there is one line: evtable_%.h: /usr/include/linux/input.h03:30
kyakxiangfu: diffing your and mine triggerhappy-0.1.3.tar.gz unpacked, shows nothing03:30
kyakwe have identical archives03:30
xiangfukyak: I think there is something different with your HOST input.h and mine.03:31
xiangfuthe evtable_KEY.h is create by HOST head file.03:31
kyakwhy would it want to look into my host input.h? weird03:31
kyakmaybe i don't even have the lindex-headers.. is it an openwrt dependancy?03:31
xiangfukyak: I don't think it's an openwrt dependancy. just the wrong path in triggerhappy makefile03:32
xiangfuwe should path it to like $(KERNEL_DIR)/usr/include/linux/input.h in openwrt.03:33
kyak$ grep KEY_RFKILL /usr/include/linux/input.h03:33
kyak#define KEY_RFKILL              247     /* Key that controls all radios */03:33
kyakdoes your /usr/include/linux/input.h any diferernt from my?03:33
kyakyou are right..03:34
xiangfukyak: there is no KEY_RFKILL in my .../input.h03:34
kyakok.. so what should we do now?03:34
xiangfukyak: send patch to Mirko and cc to upstream mailing list.03:35
kyakxiangfu: ok. thanks!03:37
xiangfukyak: btw what is your HOST kernel version?03:38
kyakLinux bas #1 SMP Mon Sep 20 19:00:25 UTC 2010 i686 i686 i386 GNU/Linux03:38
xiangfuseems this KEY is only show up after 2.6.3303:39
xiangfuyes. there is no such key in 2.6.3203:39
kyakxiangfu: wait a second.. since the change should be done in triggerhappy makefile (not the openwrt makefile)03:40
kyakwhy should i send a patch to Mrko?..03:40
kyakoh ok03:41
kyakit would be an openwrt specific patch03:41
kyakthe one in patches/03:41
xiangfukyak: yes.03:43
kyakxiangfu: http://pastebin.mandriva.com/2128004:16
kyakworks for me now.. could you please check?04:16
kyakxiangfu: this is an svn diff output in feeds/packages directory04:18
kristianpaulkyak: is  clean debian squeeze installation05:55
kristianpauli also stared over again with openwrtr repo according wiki doc05:55
kristianpaultell me what look for on my system i will give you the information05:58
kyakkristianpaul: do you have locales installed in your system?05:59
kristianpaulkyak: yes05:59
kyakkristianpaul: can you compile this program?06:03
kyakgcc prog.c -o prog && ./prog06:04
kristianpaulpaul@micro:~$ gcc locale.c -o locale06:04
kristianpaulpaul@micro:~$ ./locale06:04
kristianpauldo i need en-utf8 locale en my system?06:05
methril_workmorning all!06:06
kyaklocale -a |grep en_US.UTF-806:06
kyakkristianpaul: what it says?06:06
kristianpaulsetlocale(LC_CTYPE,"en_US.UTF-8") failed!06:07
kristianpaulmy fault06:07
kyakwhat are your avaialble locales?06:08
kristianpaulbut i wonder why it compiled bwfore06:08
methril_workhere is where you miss dpkg (dpkg-reconfigure locale)06:08
kyaklet's install that locale for starters06:08
kristianpaulkyak: that locale never was in my system before and openwrt compiled ok..06:08
kyakkristianpaul: openwrt now has wide character support06:09
kristianpauli never activate that local, i dont live in the US or english is my native language ;-)06:09
kristianpaulkyak: ha !, that changes everything06:09
kristianpaulgood :=)06:09
kyakkristianpaul: just intersting, what locales you have? (locale -a)06:10
kristianpaulcan i tell use my es_CO.utf8 instead of en_US.utf8?06:10
kristianpaulkyak: es_CO.utf806:10
kyakjust install those locales06:11
kyaki guess this is the reason why uClibc developers used to distribute pregenerated locales06:11
kyakso that people won't have to install it into their host systems06:12
kyakbut the download of pregeranted locales is broken and old06:12
kristianpauli see06:14
kristianpaulok compiling now06:14
kyakmirko_: hi!06:42
mirko_kyak: hey06:42
kyakmirko_: triggerhappy won't compile if host's kernel is > 2.6.3306:42
kyaki've a patch for it..06:43
kyakthereis a hardcoded "/usr/include/linux/input.h" in triggerhappy's makefile, as noticed by xiangfu06:44
mirko_kyak: ui.., ok06:45
mirko_compield fine for me - thought it was kernel version > 2.6.33 as well06:46
kyakso it's just xiangfu's quess.. he doesn't have KEY_RFKILL in his /usr/include/linux/input.h06:46
mirko_who's mark tuson btw? ..06:47
mirko_kyak: will check it later - wanna complete some other tasks before06:48
kyaksounds like "mike tyson"06:48
kyakmirko_: np, take your time06:48
mirko_everybody should use what he/she wants - but trying to tell, why you should/have to use this and this and _not_ to use this.. *sigh* :)06:48
kyakmirko_: the patch that can be applied cleanly: http://downloads.qi-hardware.com/people/kyak/tmp/trigger.patch06:52
kyakpastebin breaks something06:53
mirko_kyak: btw - if you're sending patches to the openwrt-dev list, please attach them as file06:53
mirko_they get screwed up, especially in regard of tabs <-> spaces, newlines, etc.06:54
kyakyeah, that's probably what happens06:54
mirko_tested once again06:55
mirko_triggerhappy compiles fine with kernel _without_ this patch06:56
mirko_will get a shower, be back for a couple of minutes and then need to go (jfyi, when i'm not answering (anymore))06:57
kyakwith this patch, however, it builds fine on my machine ( and also on buildhost06:57
kyakso it does no harm :)06:58
kyakand it doesn't build on _without_ this patch07:00
CavyShould i buy a notebook or stay with the desktop ?07:08
CavyI have no experience with longevity of notebooks.07:09
xiangfukyak: hi07:19
kyakxiangfu: hi07:20
xiangfuwe will try to create a package name: triggerhappy-fixed in openwrt-package.git , apply your patch07:20
xiangfuthis hardcode is definitely a bug. the openwrt package should not read the HOST file07:20
kyakxiangfu: i just asked mirko to have a look at that patch (http://downloads.qi-hardware.com/people/kyak/tmp/trigger.patch)07:20
kyakso maybe we won't have to create additional packages07:21
kyakthough mirko says that it builds fine on
xiangfukyak: need look into why it's builds fine in Mirko's , but according to this HARDCODE, the patch is good for openwrt. :)07:23
kyakxiangfu: i also tested both on my machine and build host with this patch, build fine07:23
larsckyak: in my opinion a better fix would be to use something like `echo "#include <linux/input.h>" | $(CC) $(CFLAGS) -E - | awk ...`07:25
xiangfukyak: ok. let's just create a package triggerhappy-fixed, when this patch goto upstream we remove the triggerhappy-fixed. :)07:26
kyaklarsc: oh! in fact, is a good idea07:28
kyaklarsc: but then, it should better be fixed upstream in triggerhappy07:28
kyakand we don't have to change the openwrt makefile07:28
mirko_kyak: well, the git rev you have to ;)07:29
lekernel"Read the white paper: Building an IP Surveillance Camera System with a Low-Cost FPGA"07:29
mirko_inside the openwrt makefile07:29
lekernelwell, I guess that's where the money is :)07:29
mirko_need to go07:29
kyakmirko_: right..07:29
xiangfularsc: great.07:30
kyakxiangfu: do i recall correctly that it is a pain to make triggerhappy developers accept a patch?07:31
kyakit's very tempting to create just triggerhappy-fixed :)07:32
xiangfukyak: sorry. what you mean "do I recall correctly " (sorry for bad English)07:33
kyakinstead of asking both openwrt and triggerhappy developers to do something07:33
kyakdo i remember it right07:34
kyakthat you were discussing with wolfspraul about it..07:34
mirko_need to go now to univeristy07:34
mirko_see you later07:34
xiangfukyak: i just take a look of triggerhappy. it's already go to 0.1.407:39
kyakdid they change that hardcoded path?07:39
xiangfukyak: no07:40
kyakxiangfu: then i will prepare the patch for triggerhappy makefile like larsc suggested07:42
kyaklet's see if they accept it fast07:43
larsckyak: one minor change, use $(CPPFLAGS) instead of $(CFLAGS)07:43
kyakafterwards, we will ask someone to update the triggerhappy git revision in openwrt :)07:43
rafawpwrak: (packages) from the examples what you gave me I have two categories: 1) there as you said some packages with different versions. One live under all/ repo dir, and another version under mipsel/ repo dir. I do not know exactly why are them there.. and how to manage that yet. ON this category07:43
kyaklarsc: thanks for your hints.. i will test it07:44
rafawpwrak: the packages are man-pages, *common*.. I would say that all/ repo dir has packages which you can install on any arch.. in this case our repo is just for mipsel, but OE put those there if these are packaeges wich dont depend of any arch.. (like manpages.. or stuff like *common*).07:45
rafawpwrak: but, I need to learn about why there are some of them with different version and how to manage them. I am not an oe expert anyway :).. Also, under all/ there are few packages..and them are not criticail I would say.. It just have stuff like documentation or maybe icons.. and these are not many packages there either (in comparison with mipsel/ repo dir). Now, the second category:07:46
xiangfukyak: I already fort the github of triggerhappy.07:46
rafawpwrak: 2) the second category are packages you exactly found :) I created an extra-packages/ repo dir which has few packages which I built with toolchain. I put those packages there to not confuse them with packages built from OE. But then you found that I built rxvt with toolchain and OE already found another rxvt with another version (one live07:48
rafawpwrak: under mipsel/ repo dir and another under extra-packages/). The same happens xserver-kdriver for example. I built another xserver-kdriver version and put that under extra-packages because the version buit by OE had keyboard problems with nn.07:49
rafawpwrak: now.. how to manage them?.. the jlime-pkg is so simple, so it is not smart to choose or decide what to use.. it just find the name of package and whatever version it is it keeps that version to install07:50
rafawpwrak: but again.. these are few cases I would say.. and that will not break anything anyway.. I mean.. most of them are standalone and for the category 1) I would say that these do not have critical packages.. just documentation and stuff like that (and no many packages either)07:51
viricwhat is that trigger happy?07:58
xiangfukyak: https://github.com/xiangfu/triggerhappy/commit/307612f57533c7d56cf82a03a34cbcfec908584707:59
kyaka daemon to produce events triggered by keys08:00
xiangfuviric: A lightweight hotkey daemon08:00
kyakxiangfu: btw, would be great to fork it as "triggersad" :)08:00
wpwrakrafa: okay. i just wanted to check that jlime-pkg indeed doesn't do anything special about them (explicitly or implicitly). what i'm after is a faster version of jlime-pkg. but i need to understand exactly what it does first.08:00
wpwrakrafa: so i guess it can't hurt to be able to also consider versions. not quite sure how to compare version strings, though.08:01
rafawpwrak: I found another bug or missing feature in jlime-pkg if you are doing some smarter thing :)08:01
kyakxiangfu: awesome. id' leave it like that :) have you tested?08:01
xiangfukyak: tested in my laptop.08:01
viricxiangfu: ah great. I can't find the url of it..08:01
wpwrakrafa: what would that be ?08:03
larscxiangfu: '$<' should be removed from awks arguments08:04
xiangfularsc: thanks.08:04
rafawpwrak: the jlime-pkg does not use the "Provides:" field from Packages files.. For example. Package: alsa-utils-locale-ja .. it has an extra "Provides: alsa-utils-locale, ja-translation" field, so if another package to be installed needs "alsa-utils-locale" packages installed08:06
rafawpwrak: some of alsa-utils-locale-* shoulb be installed, or some of them shoulb be installed.08:07
rafawpwrak: no idea how to manage that.. because many packages provides the same.. If some package needs alsa-utils-locale installed and it is not then which version of alsa-utils-locale-* packages jlime-pkg should choose to install ?08:08
rafawpwrak: but, also.. I have not had that problem yet installing packages.. No sure how to test that to find the problem. I tried a couple of them which needs some package which just appears under Provides: but it happened that all my tries were already installed08:09
rafawpwrak: Do you follow me? :).. maybe I need to explain better :P .. I mean.. if you install (on the example above) alsa-utils-locale-ja.. you are installing that package and also the sytem registers that alsa-utils-locale is installed as well.08:10
kyakxiangfu: builds fine here with and without $< arguments to awk (dunno what they mean)08:11
wpwrakare "alsa-utils-locale" and "ja-translation" two separate "packages" in this case ? or is "ja-translation" a qualifier of "alsa-utils-locale" ?08:12
kyakxiangfu: i think it's good to git poke the author :)08:14
rafawpwrak: no.. those packages dont exist as packages.. those are just virtual names For example.. if you need to be installed alsa-utils-locale you need to install some alsa-utils-locale-*. Any of them provides alsa-utils-locale, but alsa-utils-locale is not a real package on the repo08:15
rafawpwrak: ah no.. you are asking me the relation between  "alsa-utils-locale" and "ja-translation" .. Mmh.. no sure if there is a relation between them08:16
wpwrakrafa: (virtual packages) yes, i understand that part08:17
xiangfukyak: send out to upstream author.08:17
wpwrakdo you know if there's any formal specification of this file format ?08:17
kyakxiangfu: thanks!08:17
rafawpwrak: file format?.. you mean the package file?08:19
rafawpwrak: the files are debian format.. those are "ar" and inside you have two files.. a data.tar.gz and a control.tar.gz.. the data contains the files for / .. the control the metadata of the package.08:21
wpwrakno, i mean files like packages-all08:23
xiangfukyak: let's wait until tomorrow. see if the triggerhappy author accept the patch.08:27
kyakxiangfu: sure08:28
wpwrakrafa: also, do you know the semantics of the stuff in the Status: lines of /usr/lib/opkg/status ?08:33
kyakxiangfu: i see the authoer has commented08:33
xiangfukyak: oh. really. he is fast08:33
kyaki can't say if he' right or not08:34
rafawpwrak: no sorry, anything of those08:34
wpwrakrafa: ah well, i'll use the "best guess" strategy then :)08:35
kyakxiangfu: i think he's right.. evtable_*.h files are empty now..08:37
rafawpwrak: :)08:38
xiangfukyak: ok. yes. I tested it before I remove the "$<"08:38
kyakwhat's "$<" doing? i tested it also after i removed "$<", those files are still empty08:39
xiangfukyak: the "$<" is "/usr/include/linux/input.h" in that makefile.08:43
xiangfukyak: we are not well test it. let's just create a triggerhappy-fixed for now. it's fast and well tested. :)08:44
kyaki suggest the triggersad name :)08:45
kyakecho "#include <linux/input.h>" | gcc -E -dD - | grep define08:58
kyakxiangfu: works liek this08:58
kyak-dD makes preprocessor leave the defines08:59
kyakxiangfu: you might want to suggest this as an option to author09:03
kyakxiangfu: tested, works fine09:06
kyak        echo "#include <linux/input.h>" | $(CC) $(CPPFLAGS) -E -dD - | \09:06
kyak        awk '/^#define $*_/ && $$2 !~ /_(MAX|CNT|VERSION)$$/ {print "EV_MAP("$$2"),"}' > $@09:06
xiangfukyak: take a look: https://github.com/wertarbyte/triggerhappy/commit/d5a041b18f010d4b1f4a33cfed00fcbe604e3a9109:12
kyakseen it...09:13
kyakhe's in fact just grabbing the header file name from preprocessed output09:14
kyakmight work, too09:14
kyaki just think -dD is better09:14
kyaksince we are interested in those defines09:14
kyaknot files name09:14
kyakah see his comment09:16
kyakmakes sense; though i doubt someone wants to define a custom input.h09:17
kyakwell, at least we've drawn his attention to this09:20
kyakand now it will maybe work :)09:20
kyakmake[3]: *** No rule to make target `evtable_EV.h', needed by `eventnames.o'.  Stop.09:24
kyakwith the authors's commit -\09:24
xiangfukyak: have you edit the openwrt package Makefile??09:25
kyakno, i just create the patch for TH makefile09:25
kristianpaulunbelievable!!, GPIF designer a sofware from Cypress http://www.cypress.com/?rID=1444809:40
wpwrakkristianpaul: *.zip ... looks windows-ish09:41
kristianpaulnow the posible usable code i may use is generated from a non-free tool :*09:42
kristianpauli wonder is this thing is used in usrp as well...09:42
kristianpaulas i can read this code for the gn3s dongle is reused from usrp09:42
wpwrakperhaps the other way around ?09:43
xiangfukyak: http://pastebin.ca/199548409:43
kristianpaulwpwrak: universal driver or ehat u mean?09:44
kyakxiangfu: should work like this.. whatever :)09:44
kyakxiangfu: thanks!09:44
kyakthough the authors fix is not complete.09:44
xiangfukyak: also sended this patch to Mirko.09:45
wpwrakkristianpaul: that ursrp code got merged somewhere along the way (not quite sure what you're dealing with here, though)09:45
kyakxiangfu: great!09:45
xiangfukyak: now I need working on the "netsurf" :)09:45
kristianpaulwpwrak: wait a min i-ll paste some funy code i think09:45
kyakxiangfu: good luck to you :) see you later09:45
xiangfukyak: see you.09:46
kristianpaulfirmware code is not so big, i was readina bit what i can read about sige chip ans is just initialziation and basic setup09:47
kristianpaulwpwrak: http://paste.debian.net/100123/09:47
wpwrakthat's "not so big" indeed :)09:49
wpwrakrafa: hehe, there are HomePage and Homepage tags :)09:53
wpwrakrafa: are you sure the entry of gmu in packages-extra-packages is correctly formatted ? it's the only one what has lines that don't begin with  <tag>:09:56
kristianpaulwpwrak: http://paste.debian.net/100125/  DO NOT EDIT ;)09:57
kristianpaulwpwrak:  well the main code is more lines, if can pastebin if you want, but is just initilization and then jump to the loop09:58
wpwrakkristianpaul: aaah ! now i have the connection.09:58
kristianpaulhttp://gnuradio.org/redmine/wiki/1/UsrpFAQFX2 <- dig GPIF10:00
wpwrakyeah, got it. you're right. that stuff comes from the evil proprietary tool10:01
wpwrakluckily, the USRP2 doens't have USB :)10:01
kristianpaulgood for you10:01
kristianpaulnow i'm warned in case i decide some day buy usrpv110:02
wpwrakmandatory seppuku narrowly avoided :)10:02
rafawpwrak: no.. all under extra-packages could not be well formatted.. It is just an extra packages which could not be present.. The main repos are the other dirs.. If we remove the whole extra-packages (rm -rf extra-packages/ :) ) all the repository is okey10:05
rafawpwrak: (okey=sane)10:05
wpwrakrafa: so you say extra-packages shouldn't be there ?10:11
kristianpaulNo. The formatting of the data is all handled in the FPGA. The FX2 firmware doesn't even see the bulk data going in and out. The data is transferred by DMA to/from the GPIF from/to the FPGA USB buffers.10:11
kristianpaulDoes FX2 reformat incoming/outgoing FPGA data?10:12
wpwrakkristianpaul: i don't think that's its purpose10:12
wpwrakkristianpaul: i mean, beyond making it fit into USB packets and such10:13
kristianpaulwpwrak: what u think then?10:13
wpwrakwell it just passes the data stream between FPGA and USB, no ?10:13
kristianpaulbueno me avisa cuando para cambiar la config d cleinte10:16
rafawpwrak: it could or not.. I just added that manually to upload packages built using toolchain. But if it is not there all the repo is sane. I would like to have extra-packages there to continue building more nice stuff which OE does not bring or not builds.. (like emacs for example).10:18
rafawpwrak: if it is a problem because you see that things do not work under extra packages then we can remove it10:19
rafawpwrak: I have my own local copy anyway :)10:20
wpwraknaw, i'm just curious what it does10:20
wpwraki'm a bit surprised that OE wouldn't have emacs, though10:20
rafawpwrak: OE has10:20
rafawpwrak: do you remember the story I was telling you about what OE tried to do to build emacs?10:21
rafawpwrak: (I named it "if you know how OE tries to build emacs you would like to leave the it world"10:22
wpwrakyeah, dump/undump :) so it tries to build it but it takes very long ?10:23
rafawpwrak: it took long and it failed.. but the fun part was seeing OE trying to have a huge qemu rootfs to run a whole mips system to build emacs..10:25
wpwraki guess nobody thought of cross-compilation when they did that dump/undump things in emacs :)10:27
qi-bot[commit] Xiangfu Liu: add libparserutils http://qi-hw.com/p/openwrt-packages/2b8e30410:57
qi-bot[commit] Xiangfu Liu: add netsurf, direct fb web browser http://qi-hw.com/p/openwrt-packages/22bdc3510:57
qi-bot[commit] Xiangfu Liu: add libhubbub, an HTML5 compliant parsing library http://qi-hw.com/p/openwrt-packages/5ab222610:57
qi-bot[commit] Xiangfu Liu: add libcss, a CSS parser and selection engine, written in C http://qi-hw.com/p/openwrt-packages/1d2f2bc10:57
qi-bot[commit] Xiangfu Liu: mark libcss libhubbub libparserutils netsurf as BROKEN, not finish yet http://qi-hw.com/p/openwrt-packages/1f8d6bd10:57
aisaI'm looking at the config.full_system file, and it contains this line:11:33
aisa"CONFIG_PACKAGE_tune2fs is=y"11:33
aisais that space in there a bug?11:33
aisagood morning everyone.11:39
aisaor good evening ;-)11:39
Action: kristianpaul lunch hour12:03
kristianpaul:O  The Milkymist GUI toolkit, because X11 sucks!13:32
LunaVoraxkristianpaul, Oh I didn't knew there already was alternative to X1114:05
mindw0rkHey guys, I've heard that it was announced here about the gamepad usb-hid support being added to the kernel14:07
mindw0rkWho might be the person behind it?14:08
rafalarsc: in current kernel, drivers/mfd/jz4740-adc.c.. there is a struct platform_driver jz4740_adc_driver = {..} which has .remove = __devexit_p(jz4740_adc_remove). Should not be it __exit_p(..) ?19:03
kristianpaulohh sie projects are in final week20:35
kristianpauloh the paintbrush got interesting20:47
kristianpaulohh i founded an fft implementation in verilog.. good20:50
kristianpaulok time to get free of patent jlime :)21:21
kristianpaulalso the boostraping system21:21
kristianpaulsi es cierto !!21:28
kristianpauloh nanoterm font is tooo small21:43
kristianpaulwpwrak: can you tell me how are you using jlime boostraping thing..21:57
kristianpaulrafa: como suena "hardware al reves" ?22:05
wpwrakkristianpaul: i didn't use the bootstrap. i just installed the tar.gz on an uSD card: boot openwrt from nand, insert card, umount card, mke2fs /dev/mm..p1, mount /dev/mm..p1 /mnt; cd /mnt; bzcat ...bz2 | tar xf -; umount /mnt; sync; press M and reset22:13
wpwrak(not quite sure if it was .bz2. or .gz)22:13
kristianpaulwpwrak: ah you are using bootstraping image on uSD. right? :)22:14
kristianpauli still not undetstand that method22:16
wpwraki didn't use the bootstrap image. i used the other one, ~80 MB22:16
kristianpaulah ok22:16
kristianpaulyeah i have that too22:16
kristianpaulbut i tought you'were crosscopiling apps fot jlime on your pc, right?22:16
wpwraksure, i do that too :)22:17
kristianpaulwhen i saw that 9~ Mb image for boostrap i tought, ok wait a min where is the toolchain..22:18
wpwraki untarred jlime-2010.1-2010.1-mipsel-linux-toolchain.tar.bz2 on my pc (from /)22:20
wpwraknow .. the url ...22:20
kristianpaulfrom / .. oh22:21
wpwrakall the paths begin with ./usr/local/jlime-2010.1/mipsel/22:22
wpwrakso it's safe22:22
wpwrakthen  . /usr/local/jlime-2010.1/mipsel/environment-setup22:23
wpwraknote: ^-- the dot is important !22:23
kristianpaulahh i got the http://downloads.qi-hardware.com/jlime/images/Jlime-muffinman-bootstrap.ben-nanonote.16nov2010.tar.gz22:23
kristianpaulwait  a amin22:23
kristianpaul45Mb~ good22:24
kristianpaulwpwrak: are you sure is not danegours, just descopmresed ona separte folder and i'm seeeing  (bin  boot  dev  etc  home  lib  linuxrc  media  mnt  proc  sbin  sys  tmp  usr  var)22:47
qi-bot[commit] Xiangfu Liu: fix libhubbub, libparserutils compile and install http://qi-hw.com/p/openwrt-packages/59dacaa22:53
qi-bot[commit] Xiangfu Liu: add libwapcaplet, a string internment library http://qi-hw.com/p/openwrt-packages/da7e7d122:53
qi-bot[commit] Xiangfu Liu: fix libcss compile and install http://qi-hw.com/p/openwrt-packages/72bfce522:53
--- Fri Nov 19 201000:00

Generated by irclog2html.py 2.9.2 by Marius Gedminas - find it at mg.pov.lt!