#qi-hardware IRC log for Wednesday, 2012-10-10

kristianpaulhmm i tought they had give up with that *certification thing..01:39
kristianpauland now a hardware  first !01:41
kristianpauldid they certify software before? well i guess all gnu but not in that way.. :-/01:41
kristianpaulwell :-)01:41
kristianpaulgn8 ;)01:41
kristianpaulcat Arduino_MEGA_2560-Rev3.sch | grep eagle01:45
kristianpaul<!DOCTYPE eagle SYSTEM "eagle.dtd">01:45
kristianpaul<eagle version="6.0">01:45
kristianpaulthis sure respect my freedom of running gratis software :-)01:45
pabs3is that related to the LulzBot or an Arduino related comment?01:50
wolfspraulpabs3: unfortunately the fsf advocates the inclusion of proprietary software embedded in hardware (seriously), so if you care about freedom, you have to be careful about where they guide you02:00
pabs3well aware of that02:01
wolfspraulthere have been enough bizarre cases where between 2 solutions, one with some proprietary software open (so one could write a replacement), and a second one with the same proprietary software hidden (and unreplacable), the fsf advocates use of the latter02:02
wolfspraulalthough somehow those discussions feel like a thing of the past to me, I'm not sure who still cares about the fsf nowadays?02:03
pabs3I half-understand their reasons for that, but they seem to forget the existence of reverse engineering02:03
wolfsprauldefinitely in the past for me now, and I do find more and more people tell me "oh, I already stopped caring about that x years ago" :-)02:04
wolfspraulif apple would make the ios un-updatable, maybe the fsf could recommend/endorse the whole thing then :-)02:04
wolfspraulit's just good for a laugh nowadays...02:04
pabs3so are there no free tools to view/edit eagle files?02:05
wolfspraulthere was a chance for their leadership, but I think it passed many years ago02:05
wolfsprauldo you care about what hw they 'endorse' or 'certify'?02:06
wolfspraulI think there are maybe a few hundred people in the world that care02:07
wolfspraulmax 100002:07
pabs3me? no.02:08
wolfspraulmaybe a few of those rare species are on this channel though :-) quite possible...02:08
wolfsprauloh, not even you02:08
wolfspraulalright then02:08
wpwrakpabs3: haven't heard of any. there should be tools to convert footprint libraries, though. not sure about schematics symbols.02:19
pabs3oh, they have AutoCAD files too :/02:20
wpwrakit gets better and better ;-)02:22
wpwrakperhaps it's best to think of this as publicity. some people see "FSF" and that makes them feel good. few will dislike it because of the FSF. so it's a win for the project.02:24
wolfspraulwell I see it a little different. the fsf continues to harm itself with this.02:25
xiangfumth, just test your jz-3.6 branch. works fine on nanonote. (for work fine with OpenWrt rootfs. we still needs some patches from openwrt)02:25
wolfspraulthe positive impact for the project is minuscule, as I have seen on several occasions02:25
wpwrakmaybe it just makes the project members feel better ;-)02:25
wpwrak"the great gnu smiles upon us. now we're blessed."02:26
wolfspraulfsf is a software-only project and philosophy...02:26
wpwrakthat was good enough for many a warrior to travel to strange lands and get himself killed in a holy war, so it can't be all bad :)02:27
wolfspraultheir place in the software philosophy history books is secure, the rest is about securing a few lobbyist jobs some more years :-)02:27
wolfspraulautocad and eagle shouldn't surprise anybody02:27
wolfspraulanyway this is my perspective, funny to see that this hw endorsement stuff is still limping along :-)02:28
wpwrakwell, autocad surprises a bit. it's not a very common choice for such things. eagle is much less surprising.02:28
wpwrakwhat has never lived cannot die ;-)02:29
wolfspraulthat's a nice play on 'free software never dies'02:30
wpwraknaw, free sw resists death in other ways. i'd call that particular plan B "homeopathic survival"02:33
kyakmth: jz-3.6 and RNDIS still errors out with code 10 on Windows 7 -\04:51
kyakbut at least we know now that problem is not in RNDIS, but in USB stack or JZ4740 hardware, right?04:52
kyakdamnit, i really hoped it would work.04:56
larscThe FSF has become a institution and often institution care more about keeping themselfs running rather than their original purpose.07:23
kyaki think every non-commercial organization is more or less fighting to survive.. doesn't matter how it's called07:39
larscin this case it's not the struggle to survive but rather the inability to react to change07:44
larsce.g. one way a society can react to such a situation is to allow the organization to keep all its precious banners but any real power or saying is diverted from it07:47
larscand I think that fits quite nicely what's going on here07:47
larscone famous example is e.g. the british monarchy. The Queen is still allowed to call herself the Queen, but she has no other purpose but being the Queen.07:52
xiangfularsc, good point. 07:54
larscxiangfu: all stolen from "The evolution of Civilizations". Just saw that the book is on archive.org. The PDF starting from page 89 covers that topic.07:57
wpwraklarsc: hmm, i think the queen actually does have significant power. it's just that she's not obliged to exercise it. http://en.wikipedia.org/wiki/Reserve_powers#United_Kingdom08:24
wpwraki like the bit "The full extent of the Sovereign's prerogatives has never been fully disclosed,"08:25
wpwraki'll remember that when i establish my evil empire of world domination08:25
mthkyak: yes, the only relevant difference between the JZ4740 and JZ4770 situation is the hardware and USB driver08:28
mthlarsc said it might be possible to use the musb driver on 4740 as well, but I don't know how that would work exactly08:29
larschopefully it would work good ;)08:29
larscwpwrak: notice the "In practive however..."08:30
wpwrakyeah, she has to wait until there's a crisis in the government. e.g., in this case: http://en.wikipedia.org/wiki/1975_Australian_constitutional_crisis08:31
larscmaybe the FSF has to wait unitl there is a crisis in the opensource movement ;)08:32
wpwrakoh dear :)08:33
xiangfu"The power to declare War and Peace" how to declare a Peace?08:37
xiangfuwpwrak, there is no China in that wiki page [http://en.wikipedia.org/wiki/Reserve_powers]08:38
kyaklarsc, mth - how can use musb driver with JZ4740?08:44
kyakxiangfu: first you have to declare War :)08:45
larsckyak: it appears that the jz4740 uses the musb core.08:46
larscAnd there is a driver for the musb core08:46
larscwhat is missing is the glue code which hooks up some jz4740 specfic things to the musb driver. like clocks, interrupts, memory regions08:47
kyakoh, ok.. i thought it's a matter of configuration, but we need to code08:50
kyakdo you think using musb driver would help with RNDIS issue, or it is a hardware issue (like for other USB gadgets)?08:53
larscno idea08:55
kyakok, thanks anyway08:55
xiangfukyak, btw: I start to use tp-link + nanonote. that will fix most of the nanonote weakness.: http://www.openmobilefree.net/?p=180808:57
kyakxiangfu: looking good, kind of external WiFi/Ethernet card :)10:04
kyakjust wondering if tp-link is more powerfull than Nanonote?10:05
kyakthen it's a question where is a head and where is tail :)10:05
kyaki mean, if it's an external LCD for tp-link or external connectivity for Nanonote10:06
virichow do you power the tplink?10:09
kyakguess it's powered via USB10:31
kyakthough it's kinda strange.. can USB host be powered from USB clients? :)10:32
xiangfuviric, there is battery in that device.10:42
xiangfukyak, there are some kind of usb-screen. btw.10:45
viricah ok10:47
whitequarkkyak: usb host can, if it's an OTG host13:05
whitequarkcheck the OTG ACA specification13:05
whitequarkit's pretty simple, just a bunch of resistors13:05
Fallenouwin 3113:35
viric31, not bad13:37
kyakwhitequark: ah, ok13:52
whitequarkkyak: technically nothing prevents you from routing VBUS to +5V on the host and doing the reverse on the device13:53
whitequarkthrough USB spec requires you to do negotiation13:53
whitequarkvia ID pin13:53
kyaki guess i won't be able to charge Ben anymore after that, right? :)13:54
whitequarkkyak: yes14:19
kyakand then it will probably blow up another computer, which doesn't have that pin reversed :)14:58
kyaksomehow it reminds me about "cable of death" i received with milkymist :)15:01
kristianpaulxiangfu: usb-screen?15:44
xiangfukristianpaul, yes. I know there is such device. and someone have connect that to a OpenWrt router. he also connected usb-keyboard and usb-mouse15:45
xiangfukristianpaul, http://www.engadget.com/2009/05/18/displaylink-for-linux-turns-a-humble-wireless-router-into-a-beau/15:47
kristianpaulah yes now i remenber15:49
kristianpauli had seen that screen in some multi seat computers15:52
viricthat router is bigger than many computers :)16:12
kristianpaulguess 1Gigabit routers are not far big and powerfull than our currents netop computers 16:14
viricah yes, gigabit routers must be quite fast16:14
virichere we have a linksys with kamikaze...16:14
viricAnd it saturates at ~4MB/s between WAN and LAN16:15
viricthat's a bit annoying, given that we have a fast WAN link :)16:15
viricbut we can live with that bottleneck16:15
kristianpaulhope the router can live as well :)16:21
viricyes, quite well :)16:23
kyakviric: i guess you have NAT between WAN and LAN?16:23
kyaki had that too.. after i bridged WAN and LAN, problem went away16:24
qi-bot[commit] Xiangfu: uboot-xburst: update to v2012.10-rc2 (master) http://qi-hw.com/p/openwrt-xburst/107964716:24
virickyak: ha! well, I can't do that without trouble :)16:28
kyakyeah, this is not always possible :) in my configuration, i have a server before the router16:32
kyakwhich is doing NAT16:32
viricthis openwrt does vpn with openvpn too16:34
viricworks well enough :)16:35
virichave you seen http://www.wikispeed.com/ ? :) appeared on the press today16:38
kyakmine does two WiFi's :) i decided to set up a guest BSSID. this is insane how capable this 7 years device is (with openwrt, that is)16:39
kyakno, what's that?16:39
viriccollaborative open source car16:40
kyakfor "collaborative" there is AUTOSAR already. "open source" - what, you are going to modify your car's SW? ISO26262 will kick your ass really fast16:44
kyaki'm sorry, i'm just not too far from automotive subject16:44
kyakthis is going nowhere16:44
wpwrakkyak: if they can still catch you after the tuning ;-)16:44
viricI didn't pay much attention; I just saw it on the news16:45
kyakthey can always catch you dead, which they will if you tune wrong :)16:45
kyakthis is a car and human lifes were talking about, not some god damn linux16:46
kyakwpwrak: btw, they've fixed that problem with mkfs.ubifs: http://lists.infradead.org/pipermail/linux-mtd/2012-October/044488.html. The resulting patch contains 4 'goto' though :)16:51
wpwraki don't see any "goto" ?16:54
kyakthere is an updated patch in attachement16:56
kyakhttp://lists.infradead.org/pipermail/linux-mtd/attachments/20121010/b1472c7b/attachment.bin here 16:57
wpwraka download-only attachment ? lovely. i'll just believe you ;-)16:58
kyakis it still true that using goto's is a very bad coding practice? or something has changed?16:59
whitequarkwpwrak: consider using curl17:00
whitequarkkyak: kinda, except if you use it for error handling17:00
kyakmm.. might be the case here. The code handles errors17:00
whitequarke.g.: int error = 0; error = fun1(); if(error != 0) goto err; [repeat as needed] error: close(fd1); close(fd2);17:01
whitequarkah yes17:01
wpwrakin general, goto is okay if it leads to more understandable code. error handling is a situation where this is commonly the case17:01
whitequarkit's also used inside kernel in the same fashion17:01
whitequarkwpwrak: true.17:01
whitequarkalso, breaking from nested loops17:01
larsckyak: 'goto conisdered harmful' has been a bit harmful, because now everybody panics when he sees a goto.17:01
wpwrakyup :)17:01
whitequarkand in very very rare cases you need a computed goto17:01
wpwrakthat's called "switch" :)17:02
kyakwhat is a computed goto?17:02
whitequarka gcc extension: goto *addr17:02
whitequarkwhich performs what you think it does17:02
kyakah.. like, goto on the fly.. now this is definitely harmful17:02
whitequarkwpwrak: Safari's JavaScriptCore interpreter uses this17:03
larscbut it's not that harmful if you use it to create non spaghetti code. most other control-flow changing statements can also be considered harmful since they allow you to create bad code17:03
whitequarkit has bytecode where opcode id is machine word long and is exactly the same as the address of the label handling it in interpreter loop17:03
whitequarkwhich sounds pretty insane, but also allowed them to achieve pretty nice performance without sacrificing portability of C17:04
whitequarkit was also somehow related with fitting the entire interpreter loop in L1 cache17:04
whitequarkI'm not quite sure how switch prevented that, but it did17:04
whitequarkah, in fact computed goto is an extension specifically created for writing interpreter loops17:06
wpwrakcases where the address of a label affects the direct input of that computed goto may be the only situations where you need that sort of hack17:08
wpwrakas soon as there is some indirection, a switch () could be optimized to be equivalent17:08
whitequarkwpwrak: the problem is probably that adding another layer of indirection breaks branch predictor17:15
whitequarkthis comment is also pretty nice17:19
viricgoto isn't that bad; and compilers are quite good at giving warnings and errors for harmful gotos17:28
whitequarkeh? warnings and errors? can you provide an example?17:31
wpwrakwhitequark: what i mean is that another layer of indirection is likely to move the trick of using address as opcode beyond the optimization horizon of the compiler. it could still do it if it's able to determine that these values are used only as tokens and only inside that function, but it's much harder.17:32
wpwrakand if the values are passed outside the function, it gets even worse17:32
whitequarkwpwrak: yeah, exactly17:33
whitequarkno known compilers are able to determine that17:33
whitequarkyou'd need lots of visibility tricks and whatnot. maybe a very, very good LTO-capable compiler could do that.17:33
arossDOTmeare you there23:01
wolfspraulyes :-)23:04
wolfspraulgood morning23:04
DocScrutinizer05funny job right now: integrate a couple of "rollercoasters" and CT or MR to a smooth system23:08
arossDOTmearr thanks, my mailing list posts are being rejected. My email address works and I did a post with it but now  it's rejected.23:09
arossDOTmeits the aross.me one23:10
DocScrutinizer05http://www.amst.co.at/sites/products/humancentrifuges/centrifuge.html   http://www.swissray.com/formulaplus.html23:10
arossDOTmemaillist_qihardware@thatdomain.me or something like that23:10
wolfspraulthanks for the heads up!23:11
wolfspraullet me check this...23:11
wolfspraulhe :-)23:13
wolfspraulI set your account to 'mod' because I ran into your last mail which was so borderline 'spammish' that I thought it's just spam23:14
wolfspraulsorry about that, and thanks for bringing it up here23:14
wolfspraulif I see junk, I tend to quite radically delete whatever I can in the minimum amount of time, sometimes with some collateral damage...23:14
arossDOTmethanks, i think :)23:27
arossDOTmeShall I resend them? Or do you send me dinner to the list?23:31
arossDOTmeI think i can see how my last post looked like spam.23:33
wolfspraulresend please, I think that's the easiest23:35
arossDOTmeok spam time! lol23:35
arossDOTmeok i will quite the spam bad jokes23:35
wpwraknow i'm curious about the content :)23:49
wpwrakwas it the "Anti Thief" ?23:49
wpwraki mean, was it the "Anti Thief" that triggered wolfgang's spam finger ?23:52
arossDOTmeIt was my happy,keen "my nano note ordered" post.23:54
arossDOTme"Yea My Nano Note Is Coming!"23:54
wpwrakaah ! yeah, happy people are suspicious.23:55
arossDOTmeall the ads have given happy a unhappy name.23:56
arossDOTmethe =those23:56
--- Thu Oct 11 201200:00

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