#qi-hardware IRC log for Tuesday, 2012-04-24

wpwrakcladamw: heya ! i played a bit with board-m1 and got it to pass with 0 ERC problems. what i did was that i deleted FLASH_A0...A3 and drew them again. but i'm not sure this actually made the difference.00:29
cladamwyeah...00:29
wpwrakhere's what changed in FPGA.sch: http://pastebin.ca/213957000:29
wpwraki comitted my change. would be interesting to know if this causes any troubles on your end or of the ERC problems are gone for everyone00:30
cladamwwhen i modified somethings in eechema then DRC and modified back then DRC again, the results will be different sometimes, 00:30
wpwrakhmm, scary.00:30
wolfspra1ljust ran into the "zhaga consortium" which is developing mostly mechanical standards for future led lighting systems00:30
wolfspra1lquite nice of only the standards were more open... http://www.zhagastandard.org/specifications/book-7.html00:31
pabs3wolfspra1l: I follow it because I'm aware that the hardware I rely on (Thinkpad) is completely secret/non-free (including the software before GRUB and various bits of "firmware"). for the same reasons as I use free software, I would like to be using free hardware. so I'm watching the free harware space00:31
wolfspra1lpabs3: yes ok, that is great. but I want to encourage you to step up a little, now that you are already here in Qi...00:31
cladamwand but the discoveries about FLASH_A0 ... A3, you are right too. I also found that from my DRC histories. :-)00:31
wolfspra1lwhat is 'hardware'?00:31
wolfspra1ldive in00:31
wolfspra1lthere are things like baseband firmwares (say wifi)00:32
wolfspra1lmechanical00:32
wolfspra1ldsp00:32
wolfspra1lmemory, flash00:32
wolfspra1land and and00:32
wpwrakwolfspra1l: sebastien should like "book 7". finally a standard with no trace of overengineering ;-))00:32
wolfspra1lwell, I do think those standards could become relevant even for us, since if they really catch on, it would make sense to engineer small logic boards for those standards to be embeddable in furniture etc.00:33
wolfspra1lbut unfortunately they think they will only open at the end, members pay 1000 EUR per year, etc. etc.00:33
wolfspra1lthe usual industry stuff00:33
wolfspra1lif it takes off, thousands of companies will just copy the simple mechanical specs anyway :-)00:34
wpwrak(relevant) maybe ... not sure if it even makes sense to standardize at the level of furniture. rather at the level of components. like lamp sockets.00:34
wpwrak(copy) yeah :)00:34
wolfspra1lthat's what this is I think00:34
wolfspra1lindoor, outdoor. just lots of mechanical standards to make interoperation between different companies easier.00:35
cladamwwpwrak, btw, about last Zero DRC results, i have an idea. :-) since I was pretty sure about J17's duplicated pins in my old KiCad version, at that time I removed pins and made sure they are gone, then added new then still got that same warnings(duplicated) But yesterday, after I used DRC function in newer Library Editor, it found there's duplicated there. so I much doubted it's my "OLD" KiCad versions in here.00:35
wolfspra1lgiven that LED lighting essentially frees up designers to shape things totally freely00:35
wolfspra1lbut mechanically it still has to fit...00:36
wolfspra1lso there's need for new standards00:36
wpwrakwolfspra1l: hmm. or you could just include it in the design. if it's just a bunch of LEDs (in standard packages) and a standard power supply ...00:37
wpwrakcladamw: hmm. not sure what you mean :) i can confirm the duplicate entires - when i tried to delete a pin, i was given two choices, and after deleting, the pin was still there (that is, one of the two)00:38
cladamwwpwrak, i wanted to copy all files locally here( which generated by OLD 2448 KiCad), then use new 3494 KiCad to save them all again( by slightly move some pins then move back then save it), then compare OLD and New files.00:38
wpwrakxiangfu: i made a new gadget: http://projects.qi-hardware.com/index.php/p/wernermisc/source/tree/master/bin/pidiff00:39
cladamwwpwrak, yes, i knew there's two choices you can select in 3494 KiCad, like i said current board-m1 files mostly generated by my OLD 2448 KiCad. 00:39
wpwrakit takes this sort of nightmare: https://github.com/milkymist/board-m1/commit/09871289563b0eaa79f8300a94235fbb4e86ecb100:39
wpwrakand turns it into this: http://pastebin.ca/213957000:40
xiangfucladamw, I have install 3494 kicad.00:40
wpwrakcladamw: (compare old and new) okay, then pidiff may be useful for you, too :-)00:40
cladamwwpwrak, I used 3494 to select 'clear' choices it wanted me to pick yesterday to solve J17's duplicated pins.00:40
cladamwwpwrak, i need now to sync yours commit. moments :)00:42
wpwrakcladamw: i think you may just have accidently done a "copy block", which would duplicate everything00:42
wpwrakcould of course also be a kicad bug that saved things twice. but i haven't run into anything like that so far.00:43
cladamwwpwrak, yes, but like i said I was thought they are duplicated and tried to move pins in J17, but actually it DIDn't, but alll this confirmation was done by my OLD 2448 KiCad, then yesterday used new 3494, I successfully discovered and removed those duplicated, so super strange, that's why i now want to compared all files. :-)00:44
wpwrakhmm, could also be that the old kicad just moved both pins00:45
cladamwwpwrak, "could of course also be a kicad bug that saved things twice." <<< about this, i can pretty sure I met them when I just clicked DRC button used 2448. :)00:46
wpwrakat least some time ago, it did such things. later, it started to ask.00:46
wpwrakokay, good :)00:46
wpwrakwell, let's see if the problem is gone now. then we can worry about it when and if it strikes again.00:46
cladamwbtw, since we all already synced now to use 3494, so I should to use it to save all files then compare to see if any difference, otherwise I have no confidence to move on further, especially in my laptop it installed at least two times KiCad, so i'll ask xiangfu to help me uninstall them to avoid collision among KiCads. i don't know. But at least to remove them or then reinstall 3494 again.00:49
wpwrakyeah, can't hurt to remove the old one00:50
wpwrakonce schhist is up, it'll also be easier to track changes00:51
xiangfucladamw, do you need me run 'Test Erc' in my computer?00:51
wpwrakxiangfu: at least i would be curious :)00:51
cladamwxiangfu, why not ! :-)00:51
xiangfuWarnings Count: 800:51
xiangfuErrors Count: 000:52
wpwrakyeah ! :)00:52
cladamwxiangfu, did you git pull latest ones ? :-O00:52
xiangfulet me check. 00:53
xiangfuthere is r4/FPGA.sch update.00:54
xiangfulet me do TestErc again.00:54
xiangfuafter update kicad-libs/board-m1   'Test Erc' give nothing.00:54
xiangfuall 0.00:54
wpwrakeven better :)00:54
xiangfu'Markers:' NONE.00:54
cladamwha00:54
cladamwwpwrak, xiangfu this is the one I got yesterday by 3494 : http://downloads.qi-hardware.com/people/adam/m1/tmp/kicad/m1.erc00:55
wpwrakxiangfu:  is there a *cache*lib ?00:55
wpwrakxiangfu: if yes, please kill it and try again00:55
xiangfuwpwrak, no. 00:55
xiangfugit clean -xdf 00:55
cladamwnow after update, still Zero. :-O00:55
wpwrakexcellent :)00:55
cladamwxiangfu, what does "git clean -xdf" mean ?00:56
xiangfuremove everything that git not track, ignore. 00:57
xiangfulike the .bak. m1-cache.lib etc.  00:57
xiangfucladamw, I got the same report: http://pastebin.com/T2JaAyvq00:57
xiangfuwpwrak, so pdiff is for file like .sch?00:58
xiangfucladamw, 'git clean -xdf' is like the a fresh git clone. if there are no change/commit locally.00:59
cladamwwpwrak, seems now results/trouble turn to your site. Amazing ... :-) btw, I still need to compare and do uninstallation. 01:00
xiangfuwpwrak, we still have the m1-cache in some .sch files: http://pastebin.com/1b4Bhprf01:00
wpwraki think the cache keeps coming back. not sure what exactly produces it. just delete the local copy from time to time.01:01
cladamwso maybe m1-cache.lib in each *.sch caused that different results when DRC ? don't know. 01:03
xiangfucladamw, no.01:04
wpwrakit shouldn't matter01:04
xiangfuwpwrak, if I remove m1-cache.lib and click save whole project. it give me: pastebin.com/VAxxEmXz01:05
cladamwxiangfu, later, when you are free, could we skype sometime to help me uninstall KiCad ?01:05
xiangfucladamw, sure.01:05
wpwrakcladamw: looks reasonable01:06
xiangfuwpwrak, I think I can commit this change.01:06
cladamwwpwrak, yeah, the results from just xiangfu gave us was what I met before. Sorry that I should have described more clear. :(01:07
cladamwbut I didn't notice they could caused troubles. :(01:08
wpwrakoh, that was for xiangfu :)01:08
wpwrakyes, commit is good01:08
wpwrakthat way, we have that out of the way01:08
wpwrakyou usually get a lot of changes without much meaning when saving with a significantly different version of kicad. so it's good to have them all in the same commit, and not get them bit by bit scattered over the next week or so01:10
xiangfunew version kicad have BIG icon. good. 01:12
wpwrakand opening FPGA.sch is super-slow :-(01:13
wpwrakabout 4-5 seconds01:13
xiangfuwpwrak, no.01:13
xiangfuwpwrak, something wrong in your side.01:14
xiangfuwpwrak, super faster. same with other .sch files.01:14
wpwrakhmm01:14
cladamwwill m1-cache appear in each *.sch after we modify sch ? or after xiangfu's commit, they won't be committed 'forever' even .sch has them ?01:14
wpwrakxiangfu: do you have 3D hardware acceleration ?01:15
xiangfuwpwrak, yes. I think so. I use the NVIDIA driver.01:15
wpwrakcladamw: i don't think you need to worry about the m1-cache in *.sch. as far as i remember, kicad only writes this information but never uses it when reading01:15
wpwrakxiangfu: ah nice. no kernel hacking for you then :)01:16
cladamwwpwrak, okay, nice !01:16
xiangfuwpwrak, download from nvidia website. then install. no kernel hacking. the ben kernel already take me too much time. :)01:17
wpwrakno atusb then :)01:18
xiangfuno. not yet. 01:18
xiangfuI use 2 atBen for test for now. 01:18
kristianpaulwpwrak: do i need recent version kicad to avoid this http://paste.debian.net/164769/, when opening m1-board sch ?02:32
kristianpaulor..02:32
kristianpaulah,  3494 kicad !02:33
cladamwkristianpaul, no, you need this http://projects.qi-hardware.com/index.php/p/kicad-libs/source/tree/master/components02:36
kristianpaulahh02:36
kristianpaulown libs interesting :)02:37
cladamwkristianpaul, we created specific KiCad compoents for our own from now on, and will not use KiCad's sytem libs., so please git clone them. :-)02:37
kristianpaulsure sure02:39
kristianpaulnow where i put it after type make..02:39
xiangfukristianpaul, ?02:40
kristianpaul/usr/share/kicad/library/ ?02:41
xiangfukristianpaul, no02:41
kristianpaulah easy, root dir for m1 projects :)02:42
xiangfuput those components under "board-m1/rc4/../../kicad-libs/components"02:42
xiangfukristianpaul, but this is better :  "board-m1/rc4/../../kicad-libs/components"  :)02:42
kristianpaulyes works now02:44
cladamwxiangfu, thanks for your helps on my uninstallations on previous versions03:18
cladamwwpwrak, now yes, i still got Zero Err from 3494 KiCad. nice !03:18
xiangfucladamw, :-)03:26
wpwrakcladamw: very good ! :)06:39
wpwrakkristianpaul: the libs are in the git project kicad-libs. the idea is that board-m1 and kicad-libs (as well as all the other qi-hw projects) are checked out under the same parent directory. e.g., i use /home/qi/06:40
wpwrakkristianpaul: the various kicad-using projects then access the libs with relative paths, ../../kicad-libs/... or similar06:40
JCGsphi everyone07:22
JCGspa pygame package from an older build will install and work con newest software image?07:26
xiangfuJCGsp, maybe. not 100% sure about that.07:45
JCGspthanks, after the reflash I'll try to install it and I'll tell if it works. 07:46
JCGspxiangfu, the reset command on "usbboot" needs any argument?07:47
xiangfuJCGsp, no argument is fine.07:49
JCGspstrange thing then. I try to run it with no argument, but stays with black screen and doesn't boot. I  have to unplug usb and remove battery as usual xD07:50
JCGspxiangfu, I've tested pygame and works, but i can't create the data partition (other things work flawlessly)08:23
xiangfuJCGsp, there is a mtd.nn command for format or mount data partition.08:24
JCGspanybody knows how to use mtd.nn to mount data partition?08:31
JCGspnevermind, I managed to mount it after fighting08:47
zhaiJCGsp: good luck to you09:28
JCGspto install a package on open wrt is "opkg install PACKAGE_NAME"?10:22
wolfspraulyes, sounds right12:22
--- Wed Apr 25 201200:00

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