#qi-hardware IRC log for Friday, 2012-05-11

wpwrakthe things people do for ~50 US-cents ... http://pics.nase-bohren.de/chinauser.jpg/133669021400:51
whitequarkwpwrak: mindsplosion01:16
whitequarkDocScrutinizer: fwiw, I'm already patching the driver01:17
whitequarkit was written by some poor Korean in a great hassle, so...01:17
whitequarkthere's a huge 'companion IC' which contains power management, usb phy, a LOT of switches (one usb port is used for: MHL, USB, USB OTG, audio, charging detector, car dock, regular dock and UART)01:19
whitequarkregulators and whatnot01:20
whitequarkit's an asic probably01:20
whitequarkand it's definitely a great mess of code01:20
whitequarkwhoops, overcurrent condition. not good.01:37
whitequarkhm. I enable charging and 0.2 seconds later the port reports OC01:39
wolfspraulwpwrak: I'm back at updating the schhist01:46
wolfspraulI am thinking about removing some outdated/unmaintained ones like Xue, what do you think?01:46
wolfspraulshould we leave it there, move to the bottom maybe? or mark as 'unmaintained'?01:46
wolfspraulsince we try to increase the quality and establish some sort of style guidelines, it wouldn't help if someone accidentally stumbled into unmaintained projects01:47
wpwrak(Xue) yeah, agreed. it's unfinished anyway.02:32
wpwrakthere may be some residual use for questions of the kind "how was this or that done over there", but even this may produce misleading results02:34
wolfspraulI'm running the schhist for board-m1 right now, it seems to take a while02:36
wolfspraullots of "cannot find crystal" "cannot find cx6slx45-..."02:37
wolfspraulis that normal? maybe it goes away later... (still running)02:37
wolfspraulI invoke it with just -S and --title "Milkymist One"02:37
wpwrakhmm yes, several component have changed name. since schhist walks the history of the board-m1 but keeps all other projects static, all non-monotonous changes may cause false errors02:38
wpwraknon-monotonous: e.g., if something is renamed or removed.02:38
wolfspraulyes, I vaguely remember02:38
wolfspraulso I just wait for now02:38
wpwrakyes. it should all converge in the end :)02:40
wpwraknot really sure if this is worth "fixing". sometimes we also have the opposite effect: something wasn't up to date at the time and then gets corrected later.02:41
wolfspraulyes, agreed02:41
wolfspraullet's wait for the outcome first02:41
wolfspraulso I remove xue from schhist02:41
wolfspraulalso SIE CEIMTUN, whatever that was/is/supposed to be/etc02:41
wpwrakindeed ;-)02:41
wolfspraulsame for SAKC/SIE - unmaintained, potential confusion02:42
wpwrakno idea what the status of this one it ...02:43
wolfspraulI also think we should remove avt2 from schhist, not sure how much of that circuitry is of any value in the future02:43
wolfspraulwe are only talking about the automatically generated schhist now, of course the sources stay and anybody can run schhist locally (with some effort)02:43
wpwrakbut it's the one that has been built ? then i'd keep it02:43
wolfspraulwe keep for now, push to bottom02:44
wpwrakotherwise you have two milkymist projects and an army of my projects. a very tiny universe :)02:44
wolfspraulbut that's what it is02:44
wpwrakwe don't have to make it *that* obvious :)02:45
wolfspraulI will add kristianpaul's gps board02:45
DocScrutinizerroh: time on *.om.org is 25min off02:49
wolfspraulwpwrak: finished. can you look over http://projects.qi-hardware.com/schhist/board-m1/ ?02:57
wolfspraulone thing I noticed already is that adam is not writing commit messages with line breaks02:58
wolfspraulshould we auto break-around in that column?02:58
wolfspraulcladamw: you there? when you write a longer/multi-line git commit message, I think it would be better if you manually break around into new lines, maybe after 60 characters or so02:59
wolfspraulor maybe something is wrong with his newline?02:59
wolfspraulin github it looks better...02:59
cladamwhi, okay, so I'll seperate commits.02:59
wolfspraulnono, wait02:59
wolfspraulwe are just analyzing03:00
wolfspraullook at http://projects.qi-hardware.com/schhist/board-m103:00
wolfspraulsome commit messages go very far to the right, without breaking around03:00
wolfspraulbut I just look at the corresponding view in github (follow the >>>), and there it looks good03:00
cladamwyou meant if the commit is too long , I should keep maybe 60 characters ?03:01
wolfspraulfacts are not clear yet :-)03:01
wolfspraulmaybe werner's scripts format it into 1 line03:01
wolfspraulthis looks good (multi-line commit msg) https://github.com/milkymist/board-m1/commit/cb81048d2a0057e347a0b6b564665818d1c8abf003:02
wolfspraullet's see what werner says, he may remember what the plan was...03:02
wpwrakyeah, it's my script that does this03:12
wpwrakthe issue is space: multiline commit messages would increase the vertical spacing of everything03:13
wpwrakyou can get the commit message in all its glory by following the link marked with >>>03:16
wolfspraulok but you like them going so far to the right?03:18
wolfspraulit's really disturbing to see the horizontal scrollbar extend so far to the right03:18
wpwrakat least nothing is lost :)03:19
wpwrakwe could truncate them if this really bothers you03:19
wolfspraulno, not really03:20
wolfspraulI just noticed it03:20
wolfspraulno need to waste time over this imho03:20
wolfspraulcladamw: you don't need to change anything, all perfect on your end with multi-line commit messages...03:20
wolfspraulwpwrak: ok, so the output of board-m1 looks good to you?03:21
wolfspraulI can't spot any immediate issues03:21
wolfspraulit's very big though03:21
wolfspraulmaybe some renaming/reordering of columns?03:21
wpwraki was just thinking about merging columns where one timeline ends and then another timeline starts03:22
wpwrake.g., when renaming/splitting/merging sheets03:22
wolfspraulok, I clean up the other projects first, remove the ones discussed above, add kpaul's gps board03:22
wpwrakthat would get rid of three columns of thumbnails. still quite a lot left, though03:23
wpwrakmaybe we need another level of overviews :)03:24
wpwrakactually ... lemme see what this looks like in other browsers than konqueror ...03:24
wolfspraulneed to run out quick, the exciting and disturbing markets are calling :-)03:26
wpwraklooks good in all of them. i notice that there are a lot of components missing, though03:26
wolfspraulmissing because? not yet in kicad-libs? something wrong with paths?03:27
wpwrakmaybe you need to git pull in kicad-libs ?03:27
wolfspraulthe script does this automatically, before any of the actual board projects03:27
wpwrakah, good. my own foresight surprises me ;-)03:28
wpwrakof course, then this means something's amiss03:28
wolfspraulyes the markets are disturbing, a lot03:28
wolfspraulbut I made my peace, I think, with the way people are doing business here03:28
wpwrakyou mean supermarket ?03:29
wolfspraulso I just walk around now with an eternal smile on my face, floating on the air, using the markets for what they are good for, and not complaining about anything else, even to myself03:29
wolfspraulno other way03:29
wolfspraulno no, electronic markets of course :-)03:29
wpwrakah :)03:29
wolfspraulhow do they like to say, "with special chinese characteristics"03:29
wolfspraulha ha03:29
wolfspraullike the democracy here03:29
wolfspraulalso with "special chinese characteristics"03:29
wolfspraulover time you get used to it, and they don't change, for sure not... :-)03:30
wolfspraulnothing can disturb my smile anymore03:30
wpwrakhave fun ! :)03:30
wolfspraulhey I looked for ultrasonic cleaners the other day03:32
wolfspraulha ha03:32
wolfspraul"cleaning" is not something of high value...03:32
wolfspraulit must be somewhere, but the tool shops I found all tried to sell me some strong airblower instead03:33
wpwrakthey should be easy to find. there's a ton of cheap chinese ones even here on the market :)03:33
wpwraknaw, liquid, not air :)03:33
wolfspraulsure sure, I know03:33
wolfspraulwill get to it, today have some other things first03:34
wpwrakmaybe tell them you want to clean injectors (for cars)03:34
wolfspraulthen I will start the PCB project in earnest soon, but only after schhist is up and running again :-)03:34
wpwrakinjector cleaning uses stronger devices. but once you find a shop that has anything like this, they can probably tell you where to find weaker units. or maybe then even have them too.03:34
wpwrakthe best choice would be a device that sweeps the ultrasound frequency. these are specifically for electronics and have the lowest risk of causing damage. but they may be more expensive. (they are here)03:38
wpwrakotherwise, just for for the one with the lowest power to volume ratio03:38
wpwrakthe one i have has something like 35 W for 0.5 l03:40
wpwrakthe weaker the device, the less likely the resonances are to damage the chips03:41
wpwrakif you get one with a frequency sweep, you avoid the resonances. then you can go for higher power levels.03:41
wpwrakthere are also fancy (yet cheap) ones with built-in heating. it can't hurt but probably doesn't accomplish all too much03:42
wpwrakfor the cleaning fluid, you just use de-mineralized or distilled water. available at any car supply shop03:43
wpwrakyou can also use other liquids of course. e.g., alcohol. but water is usually sufficient. and very cheap ;-)03:45
qi-bot[commit] Werner Almesberger: schhist/sanitize-profile: treat LibDir as a list (master) http://qi-hw.com/p/eda-tools/9365a2303:53
wpwrakthis should fix some of the missing symbols03:54
wpwrakwolfspraul: (foresight) ah wait. that's your foresight, not mine. that's why it surprised me :)04:03
wpwrakyeah. looks better now04:06
wpwrakwolfspraul: pulling the sanitize-profile change and then running schhist again will fix the missing symbols. it'll finish reasonably quickly since it can get much of the data from the cache.04:11
qi-bot[commit] Xiangfu: add mjpg-streamer config (master) http://qi-hw.com/p/openwrt-packages/1982f9a04:25
qi-bot[commit] Xiangfu: cgminer update to 2.4.0 (master) http://qi-hw.com/p/openwrt-packages/ce8896604:25
qi-bot[commit] Xiangfu: cgmier: update it automatic detect icarus (master) http://qi-hw.com/p/openwrt-packages/6867d6704:25
qi-bot[commit] Xiangfu: cgminer: cgminer now working with pool that don't provide midstate (master) http://qi-hw.com/p/openwrt-packages/5f0428904:25
qi-bot[commit] Xiangfu: cgmienr: update POOL1 to ozco.in (master) http://qi-hw.com/p/openwrt-packages/f3e735404:25
qi-bot[commit] Xiangfu: cgmienr: add usb hotplug for Icarus (master) http://qi-hw.com/p/openwrt-packages/81a454904:25
wpwrakhehe. now that's what i call overview: http://downloads.qi-hardware.com/people/werner/tmp/durchblick.png05:24
qi-bot[commit] Werner Almesberger: schhist/schhist2web: truncate commit entries after 300 characters (master) http://qi-hw.com/p/eda-tools/b7816b905:51
qi-bot[commit] Werner Almesberger: schhist2web: column sharing with SCHHIST_ORDER=...:first=next=...:... (WIP) (master) http://qi-hw.com/p/eda-tools/1468a3305:51
wpwrakwolfspraul: i added a cutoff for very long commit entries (> 300 characters) now. in case someone writes a book ...06:01
wpwrakwolfspraul: there's also a preliminary extension of SCHHIST_ORDER that recognizes not old the old  name:name:name  sort order but also column sharing with  name1=name2=...06:04
wpwrakso  foo=bar:blah=gaga=etc  would put foo and bar into the first column, blah, gaga, and etc into the second, and then all the remaining sheets in alphabetic order06:05
qi-bot[commit] Werner Almesberger: schhist/schhist2web (ordered_names): fix typo in pattern (master) http://qi-hw.com/p/eda-tools/cc5732b06:18
wpwraka transition from one sheet to the next would look like this http://downloads.qi-hardware.com/people/werner/tmp/column-sharing.png (3rd and 4th column, with deletion (all read) followed by creation (all green))06:26
wpwrakthere are still a few issues: the names in a ...=...=... list must be in alphabetic order, the title bar only shows the first name in that list (no matter whether this is also the last surviving sheet), and there is no indication in the middle of the timeline when a new sheet occupies the column06:29
LunaVoraxHello everyone!15:51
kristianpaulNice description about how USB works for those like still unknow about it..  http://www.lendlocus.com/?q=16fusb20:01
kristianpauland those who likes PIC's like wpwrak :)20:02
viricdon't they have a c compiler?20:06
kristianpaulsdcc yes20:07
viricthen they used asm for fun I imagine20:08
kristianpaulthat mcu is too small for even code in c something serious20:08
kristianpauli remenber even some older ones you have to write your own uart in software20:09
wpwrakviric: in this sort of code you usually need assembler for the timing21:18
wpwrakkristianpaul: i didn't say i like them :) i just find them useful at times21:19
wolfspraulkristianpaul: you there?23:09
wolfspraulI included m1-gps-expansion in schhist at http://projects.qi-hardware.com/schhist/m1-gps-expansion/23:09
wolfspraulone small things I noticed - can you add some licensing information to the files in your git repo?23:09
wolfspraulright now there is nothing23:09
wpwrakwolfspraul: did you see my comments about column sharing, in the backlog ? does it look useful to you ? http://downloads.qi-hardware.com/people/werner/tmp/column-sharing.png23:14
wpwrak(assuming i add intermediate titles. still thinking about how to best pull this off. one of the things that get a little tricky in the shell :)23:15
wolfspraulah wait, no. didn't see that yet, or rather think about it yet.23:16
wpwrakall those missing libraries were a problem in schhist. i fixed that now.23:17
wolfsprauloh great23:17
wolfsprauldo I need to update the server?23:18
wolfspraulhow about kicad there?23:18
wolfspraulall schhist should be back up and running again now, including m1 and m1-gps23:18
wpwrakand i added a 300 character limit for the commit entry. in case someone feels a bit too chatty :)23:18
wolfspraulwow that's a big graphic23:19
wolfsprauldon't know what to look at23:19
wpwrak(update) yes, git pull eda-tools. then run things again.23:20
wpwrak3rd and 4th column, all in ted and then above all in green23:20
wolfspraulah yes23:20
wpwrakthat's a one sheet that got renamed. with SCHHIST_ORDER, i told schhist to put the old and the new sheet in the same column23:21
wpwrakit's all manual. didn't dare to automate it.23:21
wpwrakwhat this still need is a new title bar (like the one on top) each time the sheet in a column changes23:22
wpwrakthat is, the sheet name23:22
wpwrak(kicad update) kicad doesn't seem to need updating so far. maybe it's already the latest version anyway :)23:23
wolfspraulok, not sure about column sharing then23:24
wolfspraulone thing we can do in the m1 sheet is better order and naming of columns?23:24
wpwrakyes, we can reorder them. not sure if we should rename them in schhist. they already have two different names, which i find quite confusing.23:27
wpwrakthe order should be something like this:23:28
wpwrakwith the dead sheets placed somewhere (if you don't include them in the list, they go to the end, in alphabetical order)23:30
wpwrakthe dead ones are: FPGA, FPGA_P1, and FPGA_P223:31
qi-botThe build was successful: http://fidelio.qi-hardware.com/~xiangfu/build-nanonote/openwrt-xburst.minimal-20120510-1615 23:32
wolfspraulok I updated eda-tools, added the m1 order, and now just rerunning everything (will take a few hours I think)23:50
wpwrakyour machine is too slow :)23:51
wolfspraulit's on the server, going through all projects23:52
wpwrakstill sluggish. well, having removed the zombies will help23:54
--- Sat May 12 201200:00

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