2011-02-28 00:01 hehe.. 2011-02-28 00:01 that's not just wires.. :) 2011-02-28 00:01 wires + solder paste + solder iron 2011-02-28 00:01 have PL2303 usb<-> rs232 only here 2011-02-28 00:02 I don't know about voltage 2011-02-28 00:02 It dint sound 3.3V compliant.. 2011-02-28 00:02 I see no way to get FTDI-based cable easily 2011-02-28 00:03 hmm 2011-02-28 00:03 not locally? 2011-02-28 00:03 May be wolfspraul can sell you a multi funcional jtag-serial board ;-) 2011-02-28 00:03 hehe 2011-02-28 00:04 I think you can re-use the other cable if you can force it to work at 5v.. 2011-02-28 00:05 hm.. seems kexec isn't working 2011-02-28 00:05 after kernel selection nothing happens 2011-02-28 00:05 (5v) you can then make an voltage shifter http://ur1.ca/3cpv8 2011-02-28 00:06 But i recomend that as the last option 2011-02-28 00:09 no way of make qemu-system-mips work:( ... I will try those recomended by mth tomorrow , .... man I have spend tooo much time trying to have this avr stuff on NN ... tomorrow really is the last chance .... 2011-02-28 00:10 tuxbrain: seems native compilation was faster way :\ 2011-02-28 00:10 tuxbrain: why not cross compile? or bad things happen when trying that..? 2011-02-28 00:10 tuxbrain: btw fedora mips may have som rpm for arduino stuff i think 2011-02-28 00:11 you can take a look, just in case.. 2011-02-28 00:11 kristianpaul on cross(cross) compilation (aka Canadian compilation) I can't rid of configure check of default name gcc output check and fails 2011-02-28 00:12 btw if some one see something wrong here let me know : 2011-02-28 00:12 qemu-system-mips -M magnum -m 512M -kernel vmlinux-2.6.37-r4-qemumimps.bin -serial stdio -append "console=tty1 console=ttyS0,115200n8 logo.nologo debug root=/dev/sda rootfstype=ext2" -hda Jlime-native-sdk-image-2010.1-qemumips.rootfs.ext2 -net nic -net user,hostfwd=tcp:127.0.0.1:2222-:22 2011-02-28 00:13 -serial stdio 2011-02-28 00:13 better to remove it now 2011-02-28 00:13 -serial stdio is a desperate last time addiion to see if I have see something more without it does the same 2011-02-28 00:13 you can look at serial by pressing Ctrl + Alt + 3 2011-02-28 00:14 qemumi_m_ps ? 2011-02-28 00:14 is it ok? 2011-02-28 00:15 and what problem was with 'malta'? 2011-02-28 00:16 tuxbrain: what are you trying to accomplish? maybe xiangfu can help you. 2011-02-28 00:16 that bios problem? 2011-02-28 00:16 I would have not suggested you the qemu route 2011-02-28 00:16 there are too few people using qemu with NanoNote, if it works at all it is a bumpy road 2011-02-28 00:16 and we never made that any coordinated priority either 2011-02-28 00:17 werner said avrdude is already running on jlime and openwrt - so what else is missing now? 2011-02-28 00:19 wolfspraul:Â Â a cross tools for compile avr binaries on NN 2011-02-28 00:19 gcc, binutils and avr-lib 2011-02-28 00:19 compiled host=mips target=avr 2011-02-28 00:20 you want a compiler on Ben that can produce avr binaries? 2011-02-28 00:20 yes 2011-02-28 00:20 and how is qemu supposed to help you now? 2011-02-28 00:21 you want to build the mips2avr cross compiler in a mips qemu session? 2011-02-28 00:21 I have tried in NN but is toooo slow 2011-02-28 00:22 I have achieved the binutilsÂ Â indeed :) 2011-02-28 00:22 on NN 2011-02-28 00:22 in theory you should be able to build on x86/x86_64, with host=mipsel and target=avr 2011-02-28 00:22 yes in theory 2011-02-28 00:22 Canadian compilation 2011-02-28 00:22 but that doesn't work? 2011-02-28 00:23 it fails on a check on configure 2011-02-28 00:23 ./configure 2011-02-28 00:23 on checking default gcc output files 2011-02-28 00:23 does config.log provide any useful info? 2011-02-28 00:24 i says gcc doesn't produce executable binariesÂ Â or something like that 2011-02-28 00:24 totally true 2011-02-28 00:24 but logical 2011-02-28 00:25 let me check if I can revoer the config.log 2011-02-28 00:26 well 2011-02-28 00:26 so kexecboot is booted on NN but there are some problems with kexec 2011-02-28 00:26 that's all for today 2011-02-28 00:26 I'll investigate tomorrow 2011-02-28 00:29 -> sleep 2011-02-28 00:30 I'm going to sleep as well 2011-02-28 00:30 tuxbrain: if you can find or recreate the config.log, please put it on pastebin or a similar place 2011-02-28 00:30 then I or someone else can have a look at it 2011-02-28 00:38 here it is the config.log http://pastebin.com/QMYcSfY7 goodn8 guys 2011-02-28 00:45 tuxbrain_away: mpfr is missing 2011-02-28 00:51 wpwrak,hi 2011-02-28 00:51 I was curious about usrp 2011-02-28 00:51 and I wondered what was the differences between them 2011-02-28 00:53 You mentioned you're interested in Wifi networks debuging, right? 2011-02-28 00:54 he, you just turned on all your gadgets it seems ;-) 2011-02-28 00:55 lol 2011-02-28 00:56 the dream oomed when lookin at an usrp-nuradio pdf 2011-02-28 00:58 GNUtoo|htcdream: i think wpwrak is in a "asado", so may take a while to reply 2011-02-28 00:58 what's that? 2011-02-28 01:00 and I lack some notification system on SHR... 2011-02-28 01:00 I think i should add one 2011-02-28 01:17 kristianpaul, what do you know about gps-sdr? 2011-02-28 01:18 GNUtoo-n900: wich part exactly? 2011-02-28 01:18 does it work yet? 2011-02-28 01:18 GNUtoo-n900: you mean gps-sdr.com ? 2011-02-28 01:18 or my atemtp og doing gps-sdr on the nanonote? 2011-02-28 01:20 yes 2011-02-28 01:20 sdr.com 2011-02-28 01:33 GNUtoo-n900: it basically read the i/q data from serveral devices as usrp gn3s dongle, and well,correlate the data and give you a fix point, and navigation 2011-02-28 01:34 It can read from a file, but i really dint get a fix point in that way, a a proper sample file from usrp is kind of big.. 2011-02-28 01:34 (1Gb or so i think) 2011-02-28 01:35 yes i saw that 2011-02-28 01:35 I written in C++ and required SIMD.. 2011-02-28 01:35 600M for 1 fix 2011-02-28 01:35 yup 2011-02-28 01:35 but 2011-02-28 01:36 you can cheap and just put a small file, and track just PRN 2011-02-28 01:36 cheat also using gpredict to catch the gps satellites on the sky atÂ Â the time you take the sample 2011-02-28 01:38 GNUtoo-n900: my focus about a fully sdr process,Â Â is not this gps-sdr.com software 2011-02-28 01:38 There is one project called osgps. 2011-02-28 01:39 ok 2011-02-28 01:39 it worked with an old flahsion acquisition module (PCI bus) 2011-02-28 01:39 But the code is portable, no SIMD required. 2011-02-28 01:39 the sparkfun gps stuff need windows anyway....so not usable 2011-02-28 01:39 gn3s dongle? 2011-02-28 01:39 no windows ! 2011-02-28 01:39 yes 2011-02-28 01:39 no no 2011-02-28 01:39 ah? 2011-02-28 01:40 that dongle basically re-use the usb comunication module from the USRP 2011-02-28 01:40 sparkfun comments talked about a windows aquisition program 2011-02-28 01:40 I mean is a cypress mcu 2011-02-28 01:40 ah ok 2011-02-28 01:40 so no windows needed 2011-02-28 01:40 I mean is a cypress mcu, you can code it with sdcc and sadly it required a pre generated blob froma cypress tool 2011-02-28 01:41 but the dongle is not cheap.....for a single purpose chip 2011-02-28 01:41 check osgps code on SF.net they have a driver for that dongle 2011-02-28 01:41 indeed, no cheap.. 2011-02-28 01:41 i would prefer a gnuradio 2011-02-28 01:41 of course 2011-02-28 01:41 ah well that dongle is supported by gps-sdr.com software 2011-02-28 01:44 you've got that dongle? 2011-02-28 01:44 nah 2011-02-28 01:44 ok 2011-02-28 01:45 I thought as the images sounded similar 2011-02-28 01:45 i have a EVB from a similar IF chip from sige 2011-02-28 01:45 no acquisition hardware, just the EVB 2011-02-28 01:46 what is evb? 2011-02-28 01:46 devboard? 2011-02-28 01:46 evaluation board...right 2011-02-28 01:47 yes 2011-02-28 01:47 ok 2011-02-28 01:47 of course sige is not the only ic manufacture for rf fronted, there are more 2011-02-28 01:48 ok 2011-02-28 01:49 btw is pinguino project dead? 2011-02-28 01:49 he, no i donk think so 2011-02-28 01:49 ok 2011-02-28 01:49 Just irc is not wide used 2011-02-28 01:49 ahh ok 2011-02-28 01:51 You should looks at some GNSS/GPS related books: http://en.qi-hardware.com/wiki/GPS_Free_Stack/Books 2011-02-28 01:52 ok 2011-02-28 01:52 I dinget get all the theory and and still diguesting some things, but you got the idea from time to time. 2011-02-28 01:52 s/dingent/ dint get 2011-02-28 01:54 ok 2011-02-28 01:56 of course, any doubt i can resolve you if know i'm here :-) 2011-02-28 01:57 hmmm 2011-02-28 01:58 maybe I'll ask later if i need 2011-02-28 01:59 Btw, the only way or return an array from a function is using a pointer? 2011-02-28 02:00 an array is a pointer no? 2011-02-28 02:01 isn't it a pointer to the firsrt element 2011-02-28 02:01 ah, i can use a struct too 2011-02-28 02:02 or you can use a pointer 2011-02-28 02:02 yeap ! 2011-02-28 02:02 and return error code 2011-02-28 02:02 instead 2011-02-28 02:02 as everybody do 2011-02-28 02:03 he yes, i'm just taking my time, is first time i write in C actually. well formally :-) 2011-02-28 02:04 in C for GNU/Linux userspace you mean 2011-02-28 02:04 else you know microcontrollers C 2011-02-28 02:05 yeap linux/rtems :-) 2011-02-28 02:05 ok 2011-02-28 02:05 ansi c :-) 2011-02-28 02:05 ok 2011-02-28 02:06 i'll try to sleep 2011-02-28 02:06 you should ;-) 2011-02-28 02:07 dont remenber using pointers when coding with sdcc 2011-02-28 02:08 And lot of problems just solved embedding some asm (dont kill me that ;-)) 2011-02-28 02:09 xiangfu: There is a way in git to automatically push a master and testing branch when hitting, git push? 2011-02-28 02:10 Morning btw :-) 2011-02-28 02:11 kristianpaul: you applied the high-speed fix to your jtag-serial board, right? 2011-02-28 02:11 did you notice any speed increase after that fix? (did you verify it is now connecting in high-speed with dmesg or so?) 2011-02-28 02:12 how long does a complete reflashing take now? I remember before you said something about '1 minute'? 2011-02-28 02:12 wolfspraul: (fix) yes 2011-02-28 02:12 sure i verified 2011-02-28 02:13 (how long) i cant tell, i mean i dont re-flash everything 2011-02-28 02:13 Just change some parts of memory from time to time. 2011-02-28 02:13 does it feel faster now? 2011-02-28 02:13 sure 2011-02-28 02:14 ah, that's good then! so whatever exactly you reflash right now, it basically just takes 'a few' seconds? 2011-02-28 02:14 20 or so? 2011-02-28 02:14 yeah is fast 10 to 15 i think 2011-02-28 02:14 nice 2011-02-28 02:15 for a few megabytes, my rtems binary is not more than 3Mb 2011-02-28 02:15 doesn't matter. the faster the better I'd say. 2011-02-28 02:16 sure :-) 2011-02-28 02:16 good to know that the fix is improving things. 2011-02-28 02:20 kristianpaul: git push --all 2011-02-28 02:21 hmm 2011-02-28 02:22 all? 2011-02-28 02:22 what if i have other dirty branches i dont want to mek public ;-) 2011-02-28 02:22 make* 2011-02-28 02:29 kristianpaul: let me test a little. never using the --all before. 2011-02-28 02:29 oh, np 2011-02-28 02:34 kristianpaul: oh. --all will push all local branch to remote. 2011-02-28 02:34 just tested. 2011-02-28 02:34 yeah, it looked kike that 2011-02-28 02:35 may be git allow me to created aliases.. but anway i just wanted to be a bit lazy and save the typying of two commands ;) 2011-02-28 02:35 kristianpaul: another method is "git push [repository] branch1 branch2" :) 2011-02-28 02:36 kristianpaul: like "git push origin master testing" 2011-02-28 02:36 ah, how blind, i never tought i just can put one branch next to other 2011-02-28 02:37 me, I failed to set a subject with git-send-email 2011-02-28 02:37 xMff: when you add "--annotate" the mail will pop up for you to edit 2011-02-28 02:38 however, xiangfu: I sent some patches to the list to fix qi packages with most recent backfire 2011-02-28 02:38 you should add them when you sync next time 2011-02-28 02:39 xMff: thanks. 2011-02-28 02:41 xMff: do you want direct commit to qi packages? maybe we can add you to qi packages member :) 2011-02-28 02:41 I think I already have, but my git skills are still weak so I prefer to go the indirect route for now 2011-02-28 02:42 yes, I am listed at http://projects.qi-hardware.com/index.php/p/openwrt-packages/ 2011-02-28 04:37 whoa. turn your back to the channel for half a day, spend 1+ hour catching up ! 2011-02-28 04:37 now, let's see what happened over there at #milkymist ... 2011-02-28 04:37 kristianpaul: btw, what wast the cable length thing wolfgang mentioned ? 2011-02-28 05:33 [commit] Werner Almesberger: atusb-pgm: changed LED color; cleanup of silk screen http://qi-hw.com/p/ben-wpan/22f7eee 2011-02-28 06:01 wolfspraul: oh, one more very important news item: now we have statistics in the IRC channel ;-) (!stat !top10 !seen and such) 2011-02-28 06:12 wpwrak: the length of the upward-pointing usb cable we include with the jtag-serial duaghterboard 2011-02-28 06:13 wolfspraul: what's the issue with the cable ? too long ? too short ? 2011-02-28 06:13 no issue, it just turns out kristian paul thought for a long time 'cables from sharism are too short', and ew didn't realize 2011-02-28 06:16 aah, i see 2011-02-28 06:38 i thought "the shorter the more reliable", Moreover, Ben is usually near the laptop, so shortness is good to make it all compact 2011-02-28 06:39 if i really needed, i could go a buy 5m extension cable 2011-02-28 06:39 and now, do you really think you can easily find that short USB cables in shops? 2011-02-28 06:40 you can probably find a few different lengths in most well-stocked 2011-02-28 06:40 shops 2011-02-28 06:40 for fancy choices, you can try places like digi-key ;-) 2011-02-28 06:41 always forgets to put the strongest arguments at the end of the statement 2011-02-28 06:41 :) 2011-02-28 06:44 those (active, i hope) 5 m extension cables may be potentially troublesome. shorter - passive - ones, too. so think it's good to have a usable cable that comes with the machine. particularly in a case where it's not just the standard mini USB cable. (e.g., micro-usb, angled, and so on) 2011-02-28 06:46 by the way, needing an angled cable sounds like a bit of a design flaw to me ;) 2011-02-28 06:47 what can i say? i definitely have more than one (or even more than 5) spare mini-USB cables, but i'm using the one shipped with Ben. It's short and compact. Less wires on my table 2011-02-28 07:07 i don't even know which cable i'm using :) i tend to favour slightly longer cables, though, because my pc's front usb socket is the bottom, and my devices are usually on top of a table adjacent to the desk. so that's at least ~70 cm up and ~50 cm to the side. everybody's setup is unique. 2011-02-28 07:08 have 10cm usb cable :) 2011-02-28 07:16 xiangfu: and some people have a very unique setup ;-) 2011-02-28 07:18 aw_: i'm thinking about a production test plan for the wpan boards. when you mass-produce a board, do you normally test the connectivity of capacitors ? (i.e., if they've been properly soldered) 2011-02-28 07:18 xiangfu, wow, good patches for u-boot merge 2011-02-28 07:20 panda|x201: oh. you are in u-boot mailing list :) 2011-02-28 07:21 xiangfu, yeah, I also maintain u-boot for our board 2011-02-28 07:22 wpwrak, good question; this is depends on if automated program that testing/verifying the functionality of that concerned capacitor! 2011-02-28 07:22 panda|x201: great. then you can help me review our code. and give us some advice. 2011-02-28 07:24 wpwrak, some test equipment I met/used before. Like ATE system or ICT machine can measure dedicated smd capacitor by probing test points of that concerned capacitor. 2011-02-28 07:24 aw_: atben and atusb have the usual bypass caps around the chips, the two load capacitors for the crystal, and 2-3 DC blocking caps for RF 2011-02-28 07:24 aw_: (test points) ah, so they put extra test points into the layout ? or can the ate "grab" the terminals of the capacitor directly ? 2011-02-28 07:28 aw_: in the case of atben/atusb, some cap malfunctions can be easily detected by a functional test. e.g., if the ones carrying the RF signal aren't connected, the resulting signal will be very weak (although you may still have reception over a short distance). also, if analog bypassing is very bad, then the signal amplitude will change during a packet transmission. this should also be detectable by software. 2011-02-28 07:29 wpwrak, be noticed (such test points) is all simulated in advance not like you did on atben and atusb. This close knowledge are surely i used before to probe/'grab' the terminals of capacitor directly without supplying power. 2011-02-28 07:29 aw_: if the crystal load caps are unconnected/missing, the frequency will be a little off. about 250 ppm if both caps are missing. haven't measured what happens if only one is connected. 2011-02-28 07:30 you could imagine that ICT machine is a active multimeter as capacitor measurement. 2011-02-28 07:31 aw_: what is simulated in advance ? 2011-02-28 07:31 wpwrak, agreed your idea on detecting by a gunctional test. that's the conceipt of ATE machine/industry. 2011-02-28 07:32 aw_: (250 ppm) this may be difficult to detect in a functional test. need to try. you should be able to see the error in the carrier frequency, though (with a spectrum analyzer) 2011-02-28 07:33 wpwrak, yes, normally 250ppm malfunction which should be recognized on functional test stage if watch out carefully. :-) 2011-02-28 07:34 aw_: there is also one test point for clock output. on atusb, it's very ATE-friendly. on atben, i didn't have enough room, so it's a bit smaller. you can still reach it with a probe, but it may be difficult with pogo pins (at least the kind i have. i've seen a catalog with better ones at openmoko.) 2011-02-28 07:36 i think i'll abuse some of the older board for fault simulation. take a known to be good board (even if a bit different from the current ones), they rip out components and see what happens :) 2011-02-28 07:36 s/they/then/ 2011-02-28 07:37 wpwrak, you don't need to let atben/stusb be as a good design of ATE-friendly or ICT. surely you can see many such knowledges introduced on internet. Unfortunately i didn't involve a responsibility in om. 2011-02-28 07:39 zrafa mth take a look on the line 339 from the config log http://pastebin.com/QMYcSfY7 seems ther is some hardcoded path /mnt/sda6/programas/stuff/tmp/work/i686-mipsel-sdk-linux/gcc-cross-sdk-4.4.2-r2.1/gcc-4.4.2/configure --build=i686-linu [...] this can be the cause of the problem? don't know how to solve it. 2011-02-28 07:39 aw_: yeah, my goal is to have a quick functional tests that finds as many issues as possible. not sure what failure rate to expect for smt there 2011-02-28 07:43 wpwrak, you could survey ADS (advanced design system) on web for studying simulation. Unfortunately it's not open source /free then. 2011-02-28 07:43 aw_: heh, i think i'll just improvise :) 2011-02-28 07:44 maybe some other same idea has been discovered out to do same idea on free s/w. don't know. ;-) 2011-02-28 07:44 wpwrak, i think using a functional test which is good enough... 2011-02-28 07:45 aw_: i mean, the basic principle should be easy enough. remove component, measure what happens, then try to construct an automated test for it. 2011-02-28 07:46 always will have a potential failure rate during smt! 2011-02-28 07:46 aw_: yeah, just sending a few packets should already tell us a lot. plus a look at the spectrum to see if it's nice and clean 2011-02-28 07:47 (smt) yeah. i just wonder how many things could go wrong there that wouldn't be caught by visual inspection. 2011-02-28 07:47 well, we'll certainly find out :) 2011-02-28 07:48 when i dealt with Motorola for mass production: there'sÂ Â a lot of validation system review like your questions on what failures will be? how to prevent?....tons of them. 2011-02-28 07:49 well...good that you always has have solid precautions, studies before productions. :-) 2011-02-28 07:50 aw_: yeah, it's so much easier to fix a problem before MP than on a few million units afterwards ;-)) 2011-02-28 07:50 such thinking as you is rare and commendable. ;-) 2011-02-28 07:53 thanks ;-) 2011-02-28 07:53 wpwrak, if u like, later u can survey HP Momentum which is the Simulation of Power Amplifier Bias, Match and Housing Effects. 2011-02-28 07:53 well..it's CLOSE! 2011-02-28 07:54 sigh. microwave design simulation. how i'd love to have something like that ... 2011-02-28 07:55 since i couldn't find any useful simulation tools, i had to do everything by trial and error :-( 2011-02-28 07:56 wpwrak, we know. ;-) 2011-02-28 08:01 xiangfu: yesterday, i tried adding "C" to gmenu2x icons. It looks more than ugly.. What do you think about changing the name instead? E.g. "[c]abook" or "[abook]". This also has an advantage that console apps will be grouped in gmenu2x 2011-02-28 08:01 since they are sorted by name 2011-02-28 08:02 hmm, interesting. i could actually have made the atben clock accessible to the ben, by putting a ~100k resistor between CLKM (clock output) and SCLK. well, there's not enough room on the pcb to do this easily (i would have to make it quite a bit bigger for this) 2011-02-28 08:04 (or use microvias and place the resistor on the backside, to make production even more fun ;-) 2011-02-28 08:07 kyak: hmm... how you adding "C" got gmenu2x icons? 2011-02-28 08:08 kyak: just found one command: convert -size 32x32 xc:transparent -channel RGBA -blur 0x6 -fill darkred -stroke magenta -draw "text 4,25 'C'" console_c.png 2011-02-28 08:08 can create on C png. then we can merge the console_c.png and icon. I don't know how to merger too pictures. 2011-02-28 08:09 xiangfu: i was playing manually with gimp so far 2011-02-28 08:09 tried various options 2011-02-28 08:10 kyak: I think 'ImageMagick' can do this. 2011-02-28 08:10 it sure can 2011-02-28 08:11 the first question is what to do :) 2011-02-28 08:12 all icons are different. It is hard to find a combination of right size/background color/ foreground color and transparency of "C" to match every icon 2011-02-28 08:13 if all else fails, you can use pnmcomp (from netpbm) 2011-02-28 08:16 convert should know how to merge. But again, the question is what to do, not how to do :) 2011-02-28 08:22 kyak: asking in #imagemagic irc. :)Â Â I think we first test "C" or "Console" first. 2011-02-28 08:23 hope I can get answer soon. 2011-02-28 08:23 xiangfu, is GPC27 in NN as an input now? 2011-02-28 08:24 xiangfu: ok, seems you are aimed at adding this letter :) you'll see its ugliness eventually, or, if it looks great somehow, i will be glad 2011-02-28 08:26 aw_: output. 2011-02-28 08:28 aw_: sorry. it's input 2011-02-28 08:28 kyak: you could also try draw a large terminal symbol, shrink the existing icons, and put them inside the terminal. 2011-02-28 08:28 E< 2011-02-28 08:28 xiangfu, wow...it's not useful to configure it as output. :-( 2011-02-28 08:28 maybe a [c] prefix is even easier? 2011-02-28 08:28 this could work :) 2011-02-28 08:29 need to see how it looks like. 2011-02-28 08:29 square brackets could also be a good indication of console app 2011-02-28 08:30 or maybe a different font then ? 2011-02-28 08:30 wpwrak: or, developing your idea, just to use terminal.png for all console apps :) 2011-02-28 08:31 hehe ;-) 2011-02-28 08:31 it is also fair because most console apps jsut don't have official icons 2011-02-28 08:31 different font - won't work without gmenu code midification 2011-02-28 08:32 kyak: all console apps using terminal.png. sound better. 2011-02-28 08:32 yeah, this sounds good to me, too 2011-02-28 08:33 xiangfu, later you could try to let it as input, then s/w side to monitor this pin: if detects Low, means NN is in a charge cycle; it's strong pull-down (~10mA) by charge IC itself. 2011-02-28 08:34 aw_: it's input now. 2011-02-28 08:35 aw_: the LED is controlled by hardware. 2011-02-28 08:35 xiangfu, this is thus led ON. yes...LED is controlled by charge ic itself not s/w. 2011-02-28 08:37 xiangfu, but as s/w side, surely NN can detect GPC27 to know if it acts in chargin cycle. ;-) 2011-02-28 08:37 kyak: then we will have a lot of terminal.png . like vim, emacs :( 2011-02-28 08:37 xiangfu: do you have now other icons for vim/emacs/other console apps? 2011-02-28 08:38 xiangfu: maybe we can make exception for some of them? But actually, by seeing vim icon people could get wrong impression that they are about to see gvim 2011-02-28 08:39 maybe we could find a nicer terminal.png, some scale of grey. It won't be so black and eye-catching 2011-02-28 08:39 use svg of gnome's one or konsole's one 2011-02-28 08:40 to create template 2011-02-28 08:40 then just use 24x24 icons to place inside :) 2011-02-28 08:40 iirc, you are using 32x32 now 2011-02-28 08:41 this is what wpwrak suggested 2011-02-28 08:41 you won't know how it will look like until you try 2011-02-28 08:42 Jay7: all the icons are here: http://projects.qi-hardware.com/index.php/p/gmenu2x/source/tree/master/nanonote/skins/Default/icons 2011-02-28 08:43 kyak: don't ever say me.. I've spent a lot of time trying to choose right icons for kexecboot actions :) 2011-02-28 08:43 Jay7: it's Faenza. from gnome.org. I am using the same in my pc. :) 2011-02-28 08:43 I've finished with oxygen 2011-02-28 08:44 aw_: if the changing finished. if the LCD will change? 2011-02-28 08:46 xiangfu, not follow your meaning. ;-) are you saying that before today I mentioned, NN's GPC27 is output? now changes to input..then s/w side will get that what's happening? 2011-02-28 08:47 aw_: no. I am if the battery is full. the LED will change?Â Â sorry. 2011-02-28 08:47 s/sm/mean 2011-02-28 08:51 xiangfu, yes, if charge ic acts that as a finished charge on battery. charge ic's CHRG pin will go high impedance so that it lets LED OFF. Meanwhile it's thus GJC27 is as a INPUT high. then s/w should can be realized NN is known as a fully charged. 2011-02-28 08:52 kyak: how about shrink the terminal.png to the left corner of icon. maybe is better then just a "C" 2011-02-28 08:53 xiangfu: you can try. But i now have a strong feeling it won't work. Due to different icons - it will cause different contrast levels, and will look ugly 2011-02-28 08:57 aw_: ok. understand. thanks 2011-02-28 09:06 xiangfu: soem screeshots, i made yesterday: 2011-02-28 09:07 http://downloads.qi-hardware.com/people/kyak/tmp/gmenu2x.png 2011-02-28 09:07 http://downloads.qi-hardware.com/people/kyak/tmp/gmenu2x1.png 2011-02-28 09:14 kyak: this one I just make: http://downloads.qi-hardware.com/people/xiangfu/tmp/gmenu2x_C.png 2011-02-28 09:14 kyak: your idea :) 2011-02-28 09:17 instead of adding [c] to name you may just use [name] :) 2011-02-28 09:17 I mean text label under icon 2011-02-28 09:19 [c] increasing the length by three symbols, and it may look strange for short names, like [c]sc 2011-02-28 09:19 [sc] 2011-02-28 09:20 maybe [sc] 2011-02-28 09:20 Jay7: kyak: http://downloads.qi-hardware.com/people/xiangfu/tmp/gmenu2x_C_1.png 2011-02-28 09:20 looks better imho 2011-02-28 09:20 i think [aewan] is pretty distinctive 2011-02-28 09:21 but you should notice users that apps in [] are console apps :) 2011-02-28 09:21 yeah, i wonder if they will figure that out 2011-02-28 09:21 should be obvious, no? 2011-02-28 09:21 after 2-3 runs should be :) 2011-02-28 09:22 zrafa mth I have tried to rid off this paths with --with-build-time-tools=/usr/local/jlime-2010.1/mipsel/bin --with-mpfr=/usr/local/jlime-2010.1/mipsel/usr when calling configureÂ Â but it didn't help :( the ugly path still there like I have not put nothing 2011-02-28 09:24 as long as we make different with GUI icons. 2011-02-28 09:25 ah, so you want to change both icons and names? 2011-02-28 09:25 kyak: no. no. I almost give up change icons. it's need too much time. 2011-02-28 09:27 well, i counted 27 console apps icons :) 2011-02-28 09:27 xiangfu kiak , the idea of [name] looks nice and easy, once we got a user manual we can also clarify in advance but as has been said after start two [names] apps people can figure out it selfs 2011-02-28 09:27 kyak: Jay7 let's start with "[terminal_apps]" the only problem is the GUI apps will always at the end. 2011-02-28 09:28 xiangfu: i agree. For the gui apps at the end, could we prefix them with a "space"? 2011-02-28 09:28 to move them up 2011-02-28 09:28 or patch gcc to ignore [] on shorting 2011-02-28 09:29 hm.. thats may be problem.. 2011-02-28 09:29 sorry gcc->gmenu I kind of obsesed this days 2011-02-28 09:29 it 's a good point. gui apps must be higher 2011-02-28 09:29 maybe we should put [] on gui apps? :) 2011-02-28 09:30 {} 2011-02-28 09:30 hehe 2011-02-28 09:30 Jay7: {[ look pretty much the same on Ben's LCD :) 2011-02-28 09:30 btw 2011-02-28 09:30 find ./applications/ games/ programming/ terminals/ utilities/ | grep -v .gitignore | wc -l 2011-02-28 09:30 69 2011-02-28 09:31 may be patching gmenu2x is good idea.. 2011-02-28 09:31 other good idea to have some mark in configs that app is console app 2011-02-28 09:31 then gmenu2x will add [] automagically 2011-02-28 09:32 sounds like an extra work :) 2011-02-28 09:32 but a good idea 2011-02-28 09:32 yeah.. 2011-02-28 09:32 Jay7: as long as the patch accept by upstream :) 2011-02-28 09:32 shouldn't be very hard 2011-02-28 09:32 iirc, yesterday we was announced as upstream ;) 2011-02-28 09:33 should look into gmenu2x code one time :) 2011-02-28 09:33 hm, i wonder if we can use one of utf-8's square brackets. Not the ascii one 2011-02-28 09:33 may be there is something good for kexecboot 2011-02-28 09:34 or vice versa 2011-02-28 09:38 Jay7: i heard that gmenu2x's code is am awefull piece of shit. Besides, it's using SDL 2011-02-28 09:38 as far as i know, are drawing directly to FB :) 2011-02-28 09:38 *you are 2011-02-28 09:38 yes, we are.. 2011-02-28 09:39 have some kind of own UI library already.. 2011-02-28 09:40 Jay7: but what you should look into sometime is the new openwrt image :) 2011-02-28 09:40 you'll like it! 2011-02-28 09:40 kyak: yeah, I should :) 2011-02-28 09:41 I should spend more time to playing with NN anyway 2011-02-28 09:41 at least to find why kexec is not happens 2011-02-28 09:42 ah.. I should make nice pics for news :) 2011-02-28 09:44 Jay7: what is the flow with kexecboot? Is it u-boot -> uImage+kexecboot+kexec-tools (the boot menu) -> uImage ? 2011-02-28 09:44 interface is working ok 2011-02-28 09:44 but when I choose image to boot (kexec) it freeze 2011-02-28 09:45 I suspect kexec-tools vs uImage 2011-02-28 09:45 is kexecboot kind of a small layer between u-boot and the uImage you want to boot? 2011-02-28 09:46