#qi-hardware IRC log for Tuesday, 2011-02-22

rohhe also is on freenode 'wpwrak'00:04
rohwpwrak: people playing with scpi. i pointed them to ahrt/tmc00:05
wpwrakgood ! free advertisement ;-)00:05
rohhehe.. i am happy to help people find each other in the foss community00:06
rohespecially if it serves my own (mostly not even evil) interrests ;) (in this case more foss sw for scopes)00:07
wpwrakheh :) yeah, it's surprising how little seems to be around for this00:09
wpwraknow .. let's see what happens when i try to restart X after that ubuntu upgrade ...00:27
wpwrakhmm. after some fighting, the system boots at least. lost one of my screens, though :-(01:28
wpwrakand xpdf still crashes. usrp2_fft works. phew. that's the one i currently need the most.01:30
rohwpwrak: still using xpdf?01:37
wpwrakroh: got anything leaner and quicker ? :)01:38
rohi use evince01:40
wpwrakevice uses a lot of screen real estate01:41
rohpress f1101:41
wpwrakhaha :)01:42
wpwraki wonder what happens if i check out an old git revision, git mv the files to a different place, then try to go back up to the present. will git let me do this ? will the moved files have their ancestry unbroken ? will this bring the world to an end even before LHC gets a chance to ? soon, we'll know ...02:15
kristianpaul(will git let me do this ?) i think i will02:20
larscthe history will be semi broken02:21
wpwrakkristianpaul: yeah, it usually assumes you know where your foot is in relation to your projectiles :)02:22
larscbut you can always go back02:22
wpwraklarsc: what i'm trying to do is to produce the same effect as if i had made the git mv in the past. in particular, then tracing the ancestors of the moved files, one should get back to the original ones. (i.e., it would be nice if this didn't detail schhist)02:23
larscgit log on a moved file will usualy not have the history before the move02:25
kristianpauland after?02:26
wpwrakyeah, it's a bit more complicated. i stitch it together from .... (looking) ...02:27
kristianpaullarsc: so the history just dissaper? in my current "git time"02:28
wpwrakfrom  git diff-tree02:28
kristianpaulbut if i got back some commits before i guess i can see it?02:28
larsckristianpaul: you can still see it, when you display the whole history02:29
larscjust not in the history of that one file02:29
wpwrakhmm. git seems to apply the changes to HEAD even if they've been made to an older revision02:37
wpwrakah well. i'll just copy then.02:38
wpwrakah .. maybe not ... confused HEAD and master02:41
qi-bot[commit] Werner Almesberger: Updates CAM processes of various sub-projects http://qi-hw.com/p/ben-wpan/125055b02:44
wpwraknaw, easier to just copy. the other thing doesn't feel right.02:48
wpwrakgrmbl. ubuntu introduced yet another little nuisance change: directory<TAB> completes to directory<spc> instead of directory/  luckily, there's complete -r to bin all that junk ...03:03
qi-bot[commit] Xiangfu Liu: add zgv, MPlayer, brainless http://qi-hw.com/p/openwrt-xburst/f521ceb03:23
qi-bot[commit] Werner Almesberger: atusb-sil/: moved C8051F326-related material out of atusb/ http://qi-hw.com/p/ben-wpan/2642b9e03:27
qi-bot[commit] Werner Almesberger: moved atusb/fw3/ to atusb/fw/ http://qi-hw.com/p/ben-wpan/78cfc8f03:27
qi-bot[commit] Werner Almesberger: atusb/fw/Makefile (CFLAGS): simplify -I../fw/include to -Iinclude http://qi-hw.com/p/ben-wpan/313fb7c03:27
qi-bot[commit] Xiangfu Liu: create pango.modules when system first boot http://qi-hw.com/p/openwrt-packages/a7d074603:28
qi-bot[commit] Xiangfu Liu: update nanonote-files version http://qi-hw.com/p/openwrt-packages/78ccf8603:31
qi-bot[commit] Xiangfu Liu: touch utmp for byobu, add -b ~/ ~/ to mc params http://qi-hw.com/p/gmenu2x/3ab2a4f04:17
qi-bot[commit] Xiangfu Liu: remove custum banner, focus on 100% upstream, http://qi-hw.com/p/openwrt-xburst/003127404:42
kyakmirko is so not here -\ i have these commits, burning my hands!06:52
larscuse some cold water06:53
kyakxiangfu: why is it only necessary to set g_ether.host_addr via fw_setenv bootargs? what about g_ether.dev_addr? What's the difference between the two?07:17
xiangfukyak: the usb ether gadget driver will create two device. NANONTOE:/dev/usb0   PC:/dev/usb008:17
xiangfuthe g_ether.host_addr == PC:/dev/usb0   g_ether.dev_addr == NANONOTE:/dev/usb008:18
xiangfusince we only want PC:/dev/usb0 have static mac address. so only needs setup g_ether.host_addr08:18
kyakoh, ok! it's intersting that developers decided to have two mac addresses08:19
kyakwhat could it possibly be used for?08:19
xiangfudid you saw the email from Ruben? if someone using mac system. once mac address changed. mac system think there is another new device. then re-configure that device. all the setting are lost. :)08:22
kyakyeah i know08:22
kyaksame is in linux without appropriate settings, it would name devices usb0/1/2 etc08:23
kyaki mean, what's the need of two mac addresses?08:23
xiangfuoh. sorry for mis-understand08:23
xiangfuit is not two mac addresses. there are two NICs.08:24
kyakor is it because there is actually networks emualted between g_ether.dev_addr and g_ether.host_addr?08:24
kyaktherefore they can't have the same mac08:24
kyakok, but if the second NIC is on PC, why it's mac address is set up via Ben?08:25
kyakit's pretty confusing. The ether gadget driver on PC has such settings, too08:26
kyakwhat would happen if they contradict?08:26
xiangfuhmm... think about as Ben is a NIC. that is why mac address can setup by Ben.08:28
xiangfuif you plus a NIC to your PC, there is a mac address from NIC. if you plus a Ben to your PC, there is a mac address from usb0 (which is set by Linux inside ben)08:30
kyakok, it's more clear now08:31
kyakbasically, there are two NICs in Ben :)08:31
xiangfukyak: yes.08:32
xiangfuI have test setup same mac address on both NICs.  ssh works fine not 'ping'08:42
xiangfuping will give all packet loss.08:43
xiangfukyak ^08:43
kyakhm. shouldn't work at all, theoretically :)08:43
kyakif it's two same macs on the network08:43
kyakxiangfu: try to clear arp cache on both ben and PC08:44
kyakok, this is all just a speculation. I ahve no idea how ether gadget driver works08:45
xiangfuhmm.. the mac address only works in local network. not through gateway, right?08:48
xiangfuif PC:/dev/usb0 and NN:/dev/usb0 have same mac address. I can understand why 'ping' not working. don't understand why 'ssh' works fine.08:50
kyakme too. Have you restarted the ssh connection? Are you sure you are not ssh'ing to localhost? ;)08:51
kyakxiangfu: is the internet connection on Ben working fine at the same time?08:52
xiangfuworks fine.08:52
kyakare you using bridge or NAT?08:53
kyakmaybe that's the thing? Ben appears to be behind the NAT for you08:54
kyakcould debug it with tcpdump on Ben's side08:56
kyakto see what mac address Ben sees when you try to ping it or when ssh'ing to it08:57
xiangfukyak: I think there some differents in Ben and PC. in the arp table. it's only record target IP -- MAC. not itself.09:23
xiangfuso the mac address for itself(system) is meaningless.09:24
kyakok, so at least the arp table is correct09:27
xiangfuin bnn: -- e6:ff:a4:c0:4d:0f09:27
xiangfuin  pc: -- e6:ff:a4:c0:4d:0f09:28
xiangfu101 is nanonote ip address.  100 is pc ip address.09:28
xiangfuI think the mac problem happens where there are a 'switcher' . like three nanonotes and one pc connect to switcher, all nanonotes have same mac address. then there is problem.09:30
kyakyup, i just started to think about the same09:31
kyakduplicate macs is onyl a problem, when more than one _other_ hosts have same mac. Your own mac, which could be the same, doesn't count :)09:32
kyakbut then, why is ping not working?09:33
kyakcould be some other problems09:34
xiangfuping is working. I mis type the ip address.09:36
xiangfutype wrong ip address.09:36
kyakall right :)09:36
xiangfukyak: I build new images today. : http://fidelio.qi-hardware.com/~xiangfu/compile-log/openwrt-xburst.full_system-02222011-1005/packages/09:45
kyakis it to be released?09:47
xiangfuno. still not fix 'gnuplot' and 'plplot' compile problem.09:48
xiangfukyak: but maybe you can try 'opkg upgrade' for update 'gmenu2x' and 'nanonote-files' :)09:48
xiangfuneeds setup /etc/opkg.conf to http://fidelio.qi-hardware.com/~xiangfu/compile-log/openwrt-xburst.full_system-02222011-1005/packages/09:49
xiangfukyak:  back online later. see you09:49
kyakhm. it seems that it's time we start increasing the PKG_RELEASE number, so that opkg upgrade will notice it09:59
jow_laptopyes indeed10:23
mirkokyak: pong11:15
kyakmirko: gnop11:29
mirkokyak: "07:49 < kyak> mirko is so not here -\ i have these commits, burning my hands!"11:31
kyakmirko: yeah, let me find it in scrollback11:33
kyakmirko: i already adapted all qt4 ports to work with TARGET_INCDIRS+=$(STAGING_DIR)/usr/include in qmake.mk :) So if you could review this change and commit it, this would be great!11:34
kyakearlier, i said11:35
kyakit seems that QMAKE_INCDIR_QT is not designed to hold multiple directories11:35
kyakthis is the patch, basically11:37
mirkook, let me take a look...11:37
kyakbut actually, having the first three directories in TARGET_INCDIRS ($(STAGING_DIR)/include $(TOOLCHAIN_DIR)/include $(TOOLCHAIN_DIR)/usr/include) is not necessary11:38
mirkokyak: was that the reason, why QtCore, QtGui was appeneded to the toolchain-paths, but not to the stagingdir-path?11:38
kyakit gets appended only to the last directory of TARGET_INCDIRS11:39
mirkothank you! :)11:39
mirkokyak: same with LIBDIR - e.g. "-L"-flags?11:39
kyaknp :")11:39
kyakyes, should be the same there. But i didn't notice any problem with it. I thought it is because all libs are installed in one palce?11:40
kyaki guess it just appends -lQtNetwork etc and it works11:41
mirko 11 TARGET_INCDIRS+=$(TOOLCHAIN_DIR)/include $(STAGING_DIR)/include $(TOOLCHAIN_DIR)/usr/include $(STAGING_DIR)/usr/include11:42
mirko 12 TARGET_LIBDIRS+= $(TOOLCHAIN_DIR)/lib $(STAGING_DIR)/lib $(TOOLCHAIN_DIR)/usr/lib $(STAGING_DIR)/usr/lib11:42
mirkothat way it should work, right?11:42
kyakyes, it will work like this11:42
qi-bot[commit] kyak: freedroid: add icon http://qi-hw.com/p/gmenu2x/b21899117:38
qi-bot[commit] kyak: update apps to profit from the latest changes in qt4 http://qi-hw.com/p/openwrt-packages/598a81d17:38
qi-bot[commit] kyak: kinyin: it builds now http://qi-hw.com/p/openwrt-packages/6ae061017:38
qi-bot[commit] kyak: w3m: add PKG_RELEASE http://qi-hw.com/p/openwrt-packages/c19e78617:38
kristianpaulpower failure yday, and isp modem malfuction today :-/17:55
qbjectHey all.18:23
cmwdoes anyone know where i could purchase a Spectec SDW-823 for my Ben NanoNote?20:33
tuxbrainmaybe on pulster.eu still have some.20:36
urandom__they seem to have some http://www.pulster.de/engl/index.html20:37
urandom__it isnt very cheap though20:37
cmwyeah but i can understand why... they seem to be ultra rare20:38
cmwthey only except wiretransfer, im guessing people have used them before and they20:40
tuxbrainI have use it, it works but it's still buggy and don't hold your breath for further development20:42
wpwrakroh: hmm, does mail.openmoko.org have a problem with relaying ? i'm getting timeouts trying to send things through it20:46
wpwrakroh: and it may have been unusually slow already more than half a day ago20:46
wpwrakrjeffries: did you already send the missing files to paula ? i tried to reply but it seems my mail relay is malfunctioning21:44
viricHey qi people22:12
viricDo you know of any mobile phone that can act as walkie-talkie?22:12
viricis that forbidden by law, maybe?22:12
wpwrakrjeffries: well, let me try another relay ...22:12
Jay7viric: any nokia with Push-to-talk in network which support it22:12
wpwrakviric: i've seen push-to-talk services22:13
viricbut they all depend on a network deployed?22:13
Jay7network should have support22:13
Jay7I see no prob with software implementation in 3G network22:14
viricis that broadly deployed?22:14
Jay7but haven't seen any..22:14
Jay7viric: iirc, european networks should have it22:14
virichm I read it works over GPRS22:14
Jay7voip over gprs22:15
viricit's more for speaking to a big number of listeners22:15
Jay7but you know that vendor lock-ins :)22:15
viricwhat I meant was some kind of p2p radio link22:15
viricJay7: yes yes.22:15
viricgood. I'm quite sure 'push-to-talk' is not what I want then :)22:16
viricI meant of some readio link independant of the network.22:16
wpwrakviric: probably easier to get a walkie-talkie and carry two devices :)22:17
viricI see :)22:17
Jay7mobile phone with walkie-talkie glued together :)22:17
viricI'd like a single device like so.22:18
viricI'm still asking google for a device like that. Do you come up with some words that would keep away PTT?22:18
wpwrakroh: hmm, or is it the hoster ? lists.en.qi-hardware.com also responds very reluctantly. i can reach other sites without trouble, so i don't think the problem is at my end22:20
--- Wed Feb 23 201100:00

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