#qi-hardware IRC log for Sunday, 2010-10-03

qi-bot[commit] Werner Almesberger: Firmware-free C2 programmer, with USB interface. http://qi-hw.com/p/f32xbase/88f3dcc02:10
qi-bot[commit] Werner Almesberger: ben-pcb-domes is done at 1 mm and 500 um. http://qi-hw.com/p/ben-scans/77fdf5a02:39
qmasterrrhi, i have problems to compile gnuplot with the latest git13:20
qmasterrrseems that gnuplot needs libpng 1.4 but we are at libpng 1.213:21
larscsend a patch that updates libpng13:22
qmasterrrnow directFB wont compile against libpng 1.4.413:25
larschm :/13:25
larsclooks as if you have to update directfb as well.13:28
qmasterrridirectfbimageprovider_png.c:207: warning: 'jmpbuf' is deprecated (declared at /home/qmasterrr/xburst/openwrt-xburst/staging_dir/target-mipsel_uClibc-0.9.32/usr/include/png.h:109813:29
larscthe lastest version is said to compile against libpng 1.4.x13:29
qmasterrrthat could be an option13:29
qmasterrri will try that13:29
qmasterrrthanks :)13:33
qmasterrri would love to see gnuplot working on the NN13:35
rafaqmasterrr: well, if you have problems for a long time without success then you could try jlime or debian.. both have already gnuplot for nn13:54
kristianpaulrafa: how do you manager to boot SAKC is usbbot?14:21
aisaGood day qi-hackers.15:01
aisaI've recently updated my NanoNote to the latest source, and sdl is behaving oddly.15:01
aisaI'm not sure whether I am doing something wrong,15:01
aisaor whether I did not compile sdl properly.15:01
aisathings like dgclock load and hang.15:02
aisawithout doing anything to the screen.15:02
aisaI've even grabbed the sdl benchmark attached to this ticket:15:02
aisaand I get the same behavior.15:02
aisaIs there some special thing I must to do load SDL applications?15:02
aisagmenu2x isn't loading for me either.15:02
aisahello kristianpaul15:16
kristianpaulhad you defined no mouse?15:16
aisaSome programs do give me an error about not finding a mouse.15:16
aisaso I must have.15:16
aisalet me review my .config15:16
kristianpaulexport SDL_NOMOUSE=115:17
kristianpaulexport SDL_VIDEODRIVER=fbcon15:17
kristianpaulIn last sources some files are not copied by the dafult in the image15:17
aisaah, is that the message that came across the mailing list recently?15:17
kristianpaulyou need make a symbolic link before type make15:17
kristianpauli guess15:17
aisaI was going to look it up...15:17
kristianpaulif was mine :P15:17
aisait must be the same thing.15:17
kristianpaulyou can now15:18
kristianpaulin data/qi_lb60/files15:18
kristianpaulcopy all that to the / of your ben15:18
kristianpaulmenu2x wil not start automatically (not checked hwy yet)15:18
kristianpaulbut SDL will work15:18
kristianpaulas usual15:18
aisaIt is you.15:18
kristianpaulyou catch me ;)15:19
aisaI didn't understand why that suddenly stopped working,15:19
kristianpaulbecause of that15:19
aisaunless it was just never documented and the official image build (xiangfu?) did it without thinking about it.15:20
aisaif I make this symlink, do I need to delete anything before rebuilding?15:20
aisaI'd rather not run make clean.15:20
kristianpaultry make distclean15:20
kristianpaulthen copy the .config again15:20
aisaisn't that worse?15:21
kristianpauland create the symlink15:21
kristianpaulis it clean :)15:21
aisamy cross-compiler/&c are fine though.  this will cause those to rebuild, no?15:21
aisasurely this dosen't have to happen every time a file in data/qi_lb160/files changes...15:21
kristianpaulditclean i think just erase temp configs and some binaries15:22
aisaah, ok.15:22
kristianpauli think15:23
kristianpaulaisa: http://github.com/xiangfu/scripts15:23
aisaawesome, this I did not have.15:24
aisathe files in this dir get created with owner 0:0 and permissions 664?15:25
aisawith directories being 775?15:25
aisaor is it more subtle than that?15:26
kristianpaulwich dire15:26
aisathe ones copied to the image by the build.15:26
kristianpaulin my build i own those files by default15:26
aisaI don't mean in the repository,15:26
aisaI mean inside the rootfs.15:26
kristianpaulneither me15:27
aisagood to know :-)15:27
aisawell, gmenu2x works anyway!15:32
aisaok, loads anyway.  :-p15:33
kristianpauldid scp?15:33
aisascp as in secure copy?  Yes, that works.15:33
kristianpaulbbl i ned take a shower15:33
aisathank you for your help.15:34
kristianpaulnp thats why we are all here :)15:34
qmasterrrrafa: i couldn't find gnuplot in jlime :/15:42
aisaIf I had somehow gotten myself in the settings=>skin portion of gmenu2x, how would I exit?15:46
qmasterrrI updated libpng to 1.4.4 and now i get:configure: WARNING: Could not find libpng in the pkg-config search path16:06
aisaWhere is your libpng?16:09
aisaI'll check my repo to see if it is in the same place.16:10
aisacat that file?16:14
larscwhats in config.log?16:14
aisathat part looks ok to me.16:17
aisaso larsc has the next idea.16:17
qmasterrrwhat do you need exactly? its a big file :)16:22
qmasterrrthere is the same error as above16:24
aisaThis isn't an answer to your question directly, but I've found that libpng's .so location tends to move around, so my first guess would be that the version bump means the file is hiding somewhere it didn't used to be at.16:25
larscanything interesting in the lines above or beneath the error?16:25
qmasterrrconfigure says configure:27999: error: recommended PNG functions feature could not be enabled16:26
aisathat is probably the result of an earlier error.16:27
aisaduring compilation of libpng.16:27
tuxbrain_awaywpwrak: you there?16:27
qmasterrrhum, libpng compiled fine16:28
aisaIndeed, but with which features?  When the configure (or configure-like) step was running,16:28
aisadid it report not finding a header file or something similar?16:28
tuxbrain_awayonce you tellme that to achive high serial speeds with the sd I should need a microcontroler.... well any clue where to start investiganting?16:28
qmasterrraisa: let me check that16:29
tuxbrain_awaywith out any knowledge I can deduce it sould be 3,3 V, but at the end I want to comunicate with a 5V UART so I should need also need rise Tx and rise Rx at some place16:30
qmasterrrconfigure:13997: WARNING: unrecognized options: --disable-nls, --disable-ipv616:31
qmasterrrbut i don't think they a critical16:31
aisaindeed, no.16:32
aisaDid you echo the exact message you are seeing?  I don't recall.16:32
tuxbrain_awayalso I have a quiestion(s)? how the sdio I/O should be connected to that microcontroler? better said, what I don't understand is why with a microcontroler sdio will accept higher amounts/speed of data than directly doing a soft serial?16:33
qmasterrraisa: yes configure: WARNING: Could not find libpng in the pkg-config search path16:36
aisawhere is this message happening at?  during the build of the libpng package, or from another package trying to link to it?16:36
qmasterrrduring configure of cairo16:37
qmasterrrlibpng build fine16:37
tuxbrain_awaythis development will have two phases, phase one one breakout board (the custom one I made will fit) and a breadboard to do the circuitry without need to solder integrate, and one achive to do decent sdio to 5V UART comunication at 57600 bps at least then design a uSD adapter as you have done with wpan16:37
aisawhat is the path to cairo?  I can't find it in my tree.16:37
qmasterrrits in xorg -> lib -> cairo16:38
aisathank you for walking me though this...16:38
tuxbrain_awaybut the wold of microcontrolers is to new and my only experience is with ATMEL 328 at arduino level16:38
tuxbrain_awayit would be cool if I can use one of those micros due I plenty of source of that, but I can guess than being 5V will have problems with our 3,3V NN16:40
tuxbrain_awayI think I better ask on the list.16:40
qmasterrraisa: i needed a package from xorg for gnuplot :/16:40
aisaI would need to learn more about how pkg-config works, I don't see anything obvious to fix...16:44
aisaI'm very new to OpenWRT/NanoNote development.  ~1 month.  I'm still finding my legs.16:45
qmasterrrhehe, i'm new to the whole crosscompilation stuff16:45
qmasterrri'm now trying to build just gnuplot16:46
rafaqmasterrr: what do you mean with "i couldn't find gnuplot in jlime" ?16:51
rafaqmasterrr: http://pastebin.com/MafWfvXx16:51
qmasterrrsomething must be wrong with my jlime, after opkg update; i just got an "killed" for opkg list16:53
qmasterrrand opkg seemed to stop working16:53
rafaqmasterrr: ah, I see. You need at least 32MB of swap. That is nothing wrong with your system surely, just that it does not have swap. We will not need thatfix that in beta416:54
kristianpaul mrxvt saves my slow day !16:57
qmasterrrrafa: ah thankyou :)16:59
wpwraktuxbrain_away: now i am :)17:05
tuxbrainwpwrak: I'm writting a mail to the list but, have you read my intentions and cry for help?17:11
wpwrakyes, just read it17:12
wpwrak(3V vs. 5V) you may not need a level shifter. often, 3.3 V outputs are compatible with 5 V inputs. you need to check the respective data sheets for this.17:13
wpwrakfor 5 V -> 3.3. V, you need a voltage divider. some 3.3 V chips also have 5 V compatible inputs (no clamping diodes).17:14
wpwrak(controller type) i was thinking of a design that combines UART and SPI. you'd talk to the controller with SPI and the controller would take care of UART. that way, you have relaxed timing on the host side and the MCU's hardware takes care of all the hairy bit timing.17:16
tuxbrainso afaik, the ATmega328 is able to do that....17:17
tuxbrainspi interface + uart interface17:17
wpwraktuxbrain: it's probably massive overkill ;-)17:17
tuxbrainto much power comsumtion maybe?17:18
wpwrakway too much flash :)17:18
wpwrakand, hence, too much $$$ :)17:19
tuxbrainok ok, but let face two factors, I'm plenty of Atmega328 righ now17:20
wpwrakokay :)17:20
tuxbrainalso in the form factor to plug on a breadboard17:20
tuxbrainIf I achive to comunicate with this chip over SPI with the sdio, will be in fact a great success by it self17:21
wpwrakit has plenty or RAM, too. so you can buffer a LOT of data ...17:21
tuxbrainso we can narrow specifiation later, on the desing of the UART->sd interface17:22
wpwrakdoes the communication have to be 8 bit transparent ?17:22
tuxbrainehm? mmmmm? eeeeeh?... maybe?17:23
tuxbrainyou know what is mi objective, comunicate a arduino board with NN though uSD17:24
wpwrakthe question is how you indicate presence of data. e.g., when the host does one SPI transaction, can it just send/receive one byte and put a 0 if there's none to send/receive, or do you need to put any sort of "real" protocol on top of the SPI link ?17:24
tuxbrainfar from that I know barelly nothing right now17:24
wpwrakwell, of course, since you control the software on both sides, you can always figure this out later :)17:25
wpwrakbonus points for making the design such that you can flash the firmware of the board from the ben17:25
wpwrakyou should also let the ben provide the clock. so you don't need a crystal.17:26
tuxbrainyes :) that the final goal of all it17:26
tuxbrainyes since you tellme I allwais have clear this part :)17:27
tuxbrainand that's a clear advantage, so the question is , there is any SPI comm though SDIO righ now even in development?17:28
wpwrakone problem you'll have is that you don't have enough I/O lines on uSD17:28
tuxbraindamn it17:29
wpwrak(right now) if you count yourself, yes ;-) if not, probably, somewhere17:29
wpwrak(not enough IO) to solve this, you can generate the reset signal, e.g., when SS and SCLK both are high17:30
tuxbrainok so it only means than only one i/o is missing?, so I have to generate it though a combination of other two, ok17:32
wpwrakyou have a full SPI (atusd is the proof)17:32
tuxbrainbut those two signals will be ever been high both on "current" use?17:32
wpwrakand you can find an example of generating a reset signal, although using nSS and a different kind of pin, here: http://projects.qi-hardware.com/schhist/atusd/dd999ebba3354d0b9fdd3c098fb7c99b10ae580c/pdf/atusd.pdf17:33
wpwrakSPI defines that the clock is L when inactive and that nSS is H when inactive. so, unless you want to have multiple SPI slaves on the us, you can avoid both of them being high at the same time17:34
wpwrak(clock L) that is, when you lower a slave's nSS, then SCLK must already be L. before, it can be anything, e.g., when talking to another slave. but since you have only one slave, you avoid this condition17:35
tuxbrainhehe enough to undertand one to one comunication not intention of making  a bus :)17:35
wpwrakin case you really need a bus, you could always send a target selection at the beginning of an SPI transfer :)17:37
tuxbrain(I said ok as I understand a fuck of what you have said but, I think this words will be important at some point in the development)17:37
wpwrakwell, for now you can be content with just a single slave. that lone ATmega can already do a lot :)17:38
wpwrakthe device at the 5 V end (in the arduino) will also be an ATmega328 ?17:38
wpwrakor are there other gates in the way ?17:39
tuxbrainno is an atmega328 with some belts and whistles17:39
wpwrakah, the 328 isa ctually fairly cheap, too. thought it was more expensive.17:39
tuxbrainI sell it for 4¬ and I have some margin so yes is quite cheap17:40
wpwrakyup. even a single  piece is only USD 3.28 at digi-key. much less if you buy 25 or more.17:41
tuxbrainI the ones I have have already the arduino bootloader inside17:43
wpwraki hope you got the 28 MLF variant. that way, you can use the atusd form factor.17:44
wpwrakthat boot loader is probably not very useful. you can just program it directly, though. the memory programming interface uses the SPI signals. so, besides, nRESET, you already have everything17:45
tuxbrainI don't think so , but this will be phase two, integration, on phase one I will use a breadboard for tests and such17:46
wpwrakfor nRESET, use a 74...1Gsomething or just a pair of NPN transistors17:46
wpwrakyou can make this beast very very small17:47
tuxbrainyou mean I can program directly with NN trough spi....17:47
tuxbrainho ho ho, this will open some eyes in Arduino communitie man17:48
wpwrakyou need a programmer then, though. not sure if it's easier to just write your own or to tweak an existing one.17:48
tuxbrainprogrammer is soft?17:48
tuxbrainho ho ho17:48
wpwraksomething like avrdude17:49
tuxbrainbut though spi17:49
wpwrakthe protocol is always a kind of SPI. just that you use nRESET instead of nSS17:50
wpwrakit's all described in "serial downloading", section 27.8, of the atmega48...328 manual17:52
wpwrakmompls, gotta bring out the trash ...17:52
kristianpauljuan64bits: hey17:55
kristianpauljuan64bits: do you know how get boot mode in SIE?17:55
juan64bitsyes, with the jumper 15, and then press the reset button17:58
wpwrakback. let's check the voltage levels for compatibility17:58
wpwrakVoh(min@3V)  = 2.3 V. Vih(max@5V) = 0.6 * Vcc = 3 V. hmm, nominally, this is bad. however, Iih is so low that you'll have more then 3 V. so in practice, you shouldn't need a level shifter.18:01
kristianpaulthanks juan64bits18:03
kristianpaulah yes18:03
kristianpauljust looking the LCM righ side, the PCB labels are inverted :p18:03
tuxbrainon my serial test I can connect an arduino board directly to NN  , this means something?18:05
wpwraknow, for the voltage divider ... you need to take into account the parasitic capacitance at the input pin. so the resistors can't be too large. let's say we use a cutoff frequency of 200 kHz and a capacitance of 20 pF. then you get ... fc = 1/(2*pi*R*C)18:05
wpwrakabout 36 kOhm. not too bad. maybe just make it 10 kOhm.18:06
wpwrak(connect directly) yes, that should be okay. again, with a voltage divider in the 5 V -> 3.3 V direction18:06
wpwrakfor atmega to atmega, a 1:2 divider should do18:07
wpwrakso your bom is ... an atmega, 2 x 10 kOhm, a 100 nF cap, and a single NAND gate18:08
wpwrakplus whatever connector you use for the cable that goes to the aduino. maybe a 3 pin 100 mil header will do. you could even place it where i have the antenna in atusd, so it won't stick out18:10
qi-bot[commit] Alan Post: Remove BROKEN tag: wordgrinder is working with latest image. http://qi-hw.com/p/openwrt-packages/59d9e8818:10
wpwrakso you'd basically have the same form factor as atusd. same width, a bit shorter.18:13
tuxbraingreat :)18:13
tuxbrainDavid cuartielles in with me in #arduino_ES, he has just pasme an SPI slave code example to put on arduino to test SPI comunications18:14
tuxbrainnow I only need to know how to send trough sdio spi comunicatons on NN18:15
wpwrakwell, if you have an SPI interface on the arduino, you could of course talk SPI with it directly18:15
wpwrakyou could also in-circuit program it much in the same manner as the SPI-UART MCU we've been talking about18:16
tuxbrainyes, I know, but for first test turning on a led should be far than enough18:16
wpwrakwhat i mean is that, if you can use SPI instead of UART to talk to the arduino, you don't need an extra MCU for protocol conversion.18:17
wpwrakthen all you need is a bunch of passive components18:18
tuxbrainok understand, but there are also a lot of amount of serial gadgets you can connect to that dongle a part of arduino18:18
wpwrakwell, the NAND if you also want to be able to in-circuit program the arduino18:18
wpwrakyes, that's true. it would be quite versatile18:19
tuxbrainnand is Negative AND gate nor a NAND memory isn't it?18:19
wpwrakit's a gate, no memory chip :)18:19
wpwrakyou need it to generate nRESET from nSS and SCLK18:20
tuxbrainok (uff)18:20
wpwrakthen, there's of course the Mobile DDR3 ... ;-)18:20
tuxbrainyes and and a parabolic antenna to program it though satelitte18:21
tuxbrainin the middel of the jungle18:21
wpwrakyup. 12 GHz band.18:22
wpwrakit's always those small things that make the stuff expensive :)18:22
tuxbrainheheh well let's focus on let NN talt to an arduino board though SPI first, make a video and then parabolic anthena18:23
wpwrakdesign should be a piece of cake. you'll need two layers, though. (or bridges)18:24
tuxbrainby the way any clue in the soft neeeded in the NN part, this is my last piece in the puzzle18:24
wpwrakto start, just bit-bang the gpios. later, write a proper driver so that you can use an interrupt.18:25
wpwrak(interrupt) to signal arrival of data in the mcu's fifo18:25
tuxbrainthere is no spi kernel driver?18:26
wpwraksince you have tons of RAM in the atmega328, you can of course take things very very slowly ...18:26
wpwrakyou can build things around the existing driver, yes. you'd still need to handle the interrupt and the "magic" nRESET logic18:26
tuxbrainok and use the SDIOs as ends instead of whatever the driver has right now right?18:28
wpwrakwell, you tell the driver which gpios to use18:28
wpwrakbut you can do everything in user space first. no need to delve into the kernel right now.18:29
tuxbrainand ok, replace the direct gpio nRESET with the combination18:29
tuxbrainright right18:29
wpwrakthe nRESET needs some "magic" on your side. e.g., create a character device with an ioctl that sets the reset. or make a sysfs node.18:30
tuxbrainany piece of code of spi bitbanging in you grimoire ?18:30
wpwraksure ...18:31
wpwrakmuch of this is fancy at86rf230 reset logic and the various read/write operations, so you can ignore this part18:33
tuxbrainand due nReset is only needed in case I need to program it , It's save to think that with you code (the spi functions) seems "easy" to think it can work barelly out of the box ....18:35
wpwrakkristianpaul: grumble. seems that openscad can't extrude a 3D surface18:35
kristianpaulwpwrak: :)18:35
wpwrakwell, you'll have to program it sooner or later :) so better plan for it18:35
tuxbrainok :)18:36
wpwrakkristianpaul: is has 2D DXF extrusion and some geometric extrusions for 2D shapes18:36
wpwrakkristianpaul: but nothing for things in 3D space. it has a projection of 3D on 2D, so it's very close18:37
kristianpauli know ;)18:37
tuxbrainwell guys enough neuron struggling for my mind for now, a lot of piece of info to study and mess :) thanks a loooooot18:37
wpwraki read some of the discussions on the mailing list. amazing how upset some people can get about some quirks in variable handling :)18:37
wpwraktuxbrain: you'll have interesting dreams tonight :)18:38
tuxbrainnightmares , man, are called nighmares18:38
kristianpaulnoo :)18:39
kristianpaulwpwrak: had u tried heeks python?18:49
kristianpaulat least it can do things in 3D space18:50
wpwrakyes, but i don't think heekscad can extrude 3D things either, or can it ?18:52
kristianpaulyou mean extrude froma 2D shape?18:52
wpwrakno, extrude a 3D surface18:55
wpwrak(i.e., the mesh)18:55
kristianpaulheeks dont uses meshes18:56
wpwrakwhat i need is either a function that lets me extrude each face and then do the intersection of them, or a function that lets me build a solid out of two parallel meshes (or similar)18:56
wpwrakwell, it can read them18:57
wpwrakso whatever they become :)18:57
kristianpaulheekspython allow create you parametric primtives18:57
kristianpaulall you can do with the GUI18:57
kristianpauleven boolean operations18:58
wpwrakyes, heekspython can do many things the gui can do. but can you do what i just described with the gui ?18:58
qi-bot[commit] Juan José Díaz Vecchio: Libraries location fixed, added 2 screw holes http://qi-hw.com/p/sie-ceimtun/1b9f20619:35
DocScrutinizerwow, http://projects.qi-hardware.com/schhist/ is really some decent stuff20:09
aisaDocScrutinizer: I'm glad you understand all that ;-)20:11
wpwrakDocScrutinizer: the only way to survive in the schematics jungle :)20:42
wpwrakhmm, there's an interesting problem in the AUDIO sheet of AVT2. the transistors keep on changing. i first thought that was just an oddball change, but this doesn't look right.20:44
wpwrakwolfspraul: btw, as how problematic do you consider dsv ? it currently lives a somewhat hidden life in ben-wpan/scripts/, even though it's not at all specific to that project. eda-tools would be better place. of course, when moving it there, i should also mention this. so this increases visibility. do you think this is a problem ?20:48
wolfsprauldsv? what is that?20:49
wpwrakthe data sheet viewer20:49
wpwrakthe one that encourages "deep links"20:50
wolfspraulwhy should it be 'problematic'?20:50
wolfspraulahh, OK20:50
wolfspraulnow I got it20:50
wolfspraulwell it's not the fault of the dsv that some websites have stupid policies, is it?20:50
wolfspraulso I don't consider it problematic at all20:50
wolfspraulpolicies against deep links are somewhere off the chart on my stupidity scale that I cannot even think about them, really20:51
wolfspraulcan you?20:51
wolfspraulit's just too much20:51
wpwrakyeah, for me it's also bordering on insanity. i can understand certain types of concerns with them, but the sort of use we make should be very far from that20:52
wpwrak(valid concerns) e.g., if someone pretends its their work or such20:53
wolfspraulif these companies had any self respect, they would just turn off their HTTP servers20:54
wolfspraulyes, please, move the dsv to eda-tools20:54
wpwrakwell, they could always check the referer ...20:54
qi-bot[commit] Werner Almesberger: schhist/Makefile: added sie-v2 http://qi-hw.com/p/eda-tools/2cfc6ef20:59
qi-bot[commit] Werner Almesberger: Moved dsv over from ben-wpan (now in eda-tools/dsv/dsv) http://qi-hw.com/p/eda-tools/c247df920:59
qi-bot[commit] Werner Almesberger: Moved dsv to eda-tools. http://qi-hw.com/p/ben-wpan/b1f1b4821:00
qi-bot[commit] Werner Almesberger: Merge branch 'master' of projects.qi-hardware.com:ben-wpan http://qi-hw.com/p/ben-wpan/bcbd6d821:00
wpwrakhmm, now that we have multiple committers for ben-wpan and thus sometimes branches, it will be interesting to see how schhist holds up :)21:02
wolfspraulit will only run on the default/master branch I would think, no?21:03
wpwrakhmm yes, but what is the "true" master branch in this case ? :)21:06
wpwrakin any case, git log and git rev-list sees all the changes, so things should be fine21:09
wpwrakwolfspraul: btw, perhaps it would be nice to have a link from somewhere on the project page to the schematics diffs. that way, people don't need to remember the location.21:11
wolfspraulyes agreed, would need to find how to do this easily though21:19
wolfspraulI'm currently cleaning up the community news/blogging stuff21:20
wolfspraulafter that want to finally make this wiki->OpenWrt link, so one can write scripts (or other apps) in the wiki, and have them packaged up as an OpenWrt opkg automatically21:20
wpwrakah, another area of mild to strong obsolescence :)21:20
wolfspraulobsolescence where?21:21
wpwrak(scripts in wiki) wow, sounds scary :)21:21
wolfspraulscripts in wiki, package them via openwrt on the backend, yes :-)21:21
wpwrak(obsolescence) the blog seems a bit abandoned, no ? well, haven't checked recently ... lemme see ...21:21
wolfspraulyes and no21:21
wpwrakalso the community news seem to be designed for a weekly update yet happen more like monthly21:22
wolfspraulwe have a few people in the community now that I believe can (and want to) really step up on the wiki/news side21:22
wolfspraulyes I think monthly is better21:22
wolfspraulthat's what I'm cleaning up right now21:22
wolfspraulthe planet is not bad http://en.qi-hardware.com/planet21:22
wolfspraulI am now feeding the community news from the wiki into the planet21:23
wpwraki think a "reporter" would be good to have. i'm not so sure just an editor who waits for input from the developers works so well.21:23
wolfspraulyes agreed21:23
wolfspraulI was hoping maybe aisa can grow into that role? :-)21:23
wolfspraulI'm just putting some prerequisites in place now21:23
wolfspraulso that the works makes sense21:24
aisaa role?  for me?21:24
Action: aisa raises hand21:24
wolfsprauldid you see the Milkymist 0.9 news? nothing better for me than see a long list of little fixes & improvements21:24
wolfspraulaisa: yes! our 'reporter'21:24
wolfspraulthe more investigative the better21:24
aisaah!  tell me what this involves.21:25
aisaoh, I understand.21:25
wolfsprauljust don't steal trash cans like Oracle or SAP or whoever did it... :-)21:25
aisaIt is that people are too quiet about what they are working on.21:25
aisaand so they need someone to write/e-mail and say "I really want to know what you are doing>'21:25
wolfspraulwell from my perspective there are 2 things21:25
aisaplease tell.21:25
wolfspraulfirst is the wiki homepage, in fact the Qi homepage21:25
wolfspraulit's horrible right now21:25
wolfspraulstatic jungle of words21:25
aisaI was looking at it just now.21:25
wolfspraulthat page needs to become really really good21:26
wolfspraulwhich is hard :-)21:26
wolfspraulit needs to give entry points to the various projects that are going on, and for people who are coming from different backgrounds/motivations21:26
wolfspraulit's squarely on my todo list, but I do need a partner to discuss things21:26
wolfspraulotherwise I'm just in my own soup, and things can't be good21:26
wolfspraulthen, as Werner just said, 'reporter' is a role that has a big big OPEN POSITION sign on it in our community21:27
wolfspraulwe have a lot of great, very serious, technical/engineering people21:27
wolfspraulbut for a variety of reasons, they are not so good on the communication side21:27
wolfspraulI don't even think that should change, there should be others that look at their work, think about it, translate what it means to others, etc.21:27
aisaah, ok.21:28
wolfspraulso for me it's a mix between wiki editor/news writer/investigative reporter21:28
wpwrak"planet" looks pretty good21:29
aisaI'm willing to try this, by which I mean I'm willing to participate in this sort of activity.21:29
aisaI won't be able to operate effectively at a high level, like getting articles published in magazines, &c.21:30
aisabut I can certainly manage the wiki and pull information out of people into it.21:30
aisaout of people or their projects.21:30
wolfspraulwpwrak: yes planet is nice, I think21:31
aisaMy wiki permissions don't permit me to do mid-level work, like moving pages.21:31
aisaI ran into this problem this weekend.21:31
aisabut please, tell me the second thing.21:31
wolfspraulit's a good mix between copyleft hardware oriented technical articles, and the people behind them and what else interests them in their lives21:31
wpwrakaisa: i think one problem is that, while the information is often communicated in some form, it may not be very accessible. e.g., because things relevant to many are mixed with less relevant or overly specialized ones, or because something interesting is spread over a lot of little bits21:31
aisawpwrak: does an example of this come to mind for you?21:32
wolfspraulthe 6502 javascript mask-level simulator is cool!21:32
wolfspraulaisa: permissions!21:32
Action: kristianpaul wonders why there is not a SIE branch in openwrt-xburst21:32
wpwrakaisa: then there are also the things that are cool but that you don't expect to be cool so you don't look at them. a brief presentation/summary can help there.21:32
wolfspraulaisa: sorry I am sometimes mixing up people, your wiki username is Alan Post?21:32
wolfspraulI give you _all_ permissions possible, at any time. just ask.21:33
wolfspraulroot account everywhere21:33
wolfspraulI'm pretty radical on these things21:33
aisaMy wiki username is User:Alanpost, yes.21:33
aisaIt has been quite wonderful, the permission you have given me so soon.  It has been very inspiring to continue working.21:33
wolfspraulthe only thing I don't want to do is to give you permissions for technologies that you will never use, or don't understand21:33
aisaOn User:Alanpost, you can see I mention my freenode nick to be 'aisa'.21:34
wolfspraulthat would be a security risk (because others can hijack your access)21:34
aisaindeed, I don't want them.21:34
wolfspraulso please let me know what you are comfortable with, and know how to operate, and the permissions are yours in 5 minutes21:34
aisayou seemed to know immediately what projects to give me access to.21:34
aisaI want to be able to delete and move pages on the wiki.21:34
aisaI needed to do that this weekend, and could not.21:34
aisaplus, I think our documentation is very poorly organized.21:34
wolfspraulsorry about that, one sec21:34
aisaAnd I wish to fix this, while maintaining links that still work.21:34
wpwrakaisa: (example) hmm, actually pretty much everything :) as a developer, you have a fairly steady stream of things to address, so even the "milestones" don't pop out all that much in the daily routine. and when reaching one, you may not want to take a break to brag about it.21:35
wpwrakaisa: so the information may just be the commit message posted to this channel.21:35
aisasince we're holding two conversations, may i add a third?  I'm trying to make my first commit to openwrt-xburst. Previously, I've been adding only things in packages.21:35
aisaI get the following error:21:35
aisaTo git@projects.qi-hardware.com:openwrt-xburst.git ! [rejected]        tracking_backfire -> tracking_backfire (non-fast forward)21:35
aisaerror: failed to push some refs to 'git@projects.qi-hardware.com:openwrt-xburst.git'21:35
aisawpwrak: perfect, I understand that.  This makes sense, the idea being to take raw work done by others and tell a story around it.21:36
wpwrakaisa: up, exactly this21:36
aisathat story being something that people outside the project can understand and relate to, with one goal being to draw them in.21:36
wpwrakaisa: it's often hard to make a weekly story of what you're doing. particularly also since you may not feel you've reached just the right point. so some distance to the project helps.21:37
aisaindeed, I have experienced this.  My first software prject I worked on for many years and still it isn't published.  Because I just didn't know when to stop.21:38
wpwrakyears is bad :)21:38
aisaI learned a very valuable lesson ;-)21:38
wolfspraulrejected? maybe permissions problems21:38
wolfspraulfirst step: I made your account a Wiki admin21:38
qi-bot[commit] Alan Post: Replace OpenWrt banner with a custom NanoNote banner. http://qi-hw.com/p/openwrt-xburst/af01e0421:39
aisayay!  I fixed my problem.21:39
wolfspraulaha :-)21:39
aisaI hope you let me get away with that commit ;-)21:39
wolfspraulyou changed the splash screen?21:39
aisaI had my local branch and remote branch mismatched.21:39
wolfspraulah, I see. just the drink recpie.21:40
aisaNo, I changed the whole splash screen.21:40
aisaremoving the drink recipe, which I felt the worst about.21:41
wolfspraulwhy that? (curious)21:41
aisaThat I could certainly put back in, and I can put back in a reference to OpenWrt.21:41
aisaI want my machine to tell me I'm running a NanoNote when I log in.21:41
aisasort of "branding" our distribution a bit more.21:41
wpwrakregarding wiki, i'm still dreaming of a shell integration. one thing that prevents me from doing things on the wiki (such a maintaining a project overview page or such) is that it's very intrusive in the workflow. instead of just editing a file on the side, i need a whole screen with a browser, have to worry about edits that won't be saved until I "commit" them, and so on. a wiki that lives off git or such wouldn't have this problem.21:42
wolfspraulthere is (was) a wikifs wiki filesystem project21:42
wolfspraulbut it's abandoned/dead21:42
aisaI did not know that project was dead.21:42
wolfspraulit would show you the wiki in a filesystem format, and handle the revision control on the backend21:42
wolfspraulwell, free software never dies21:43
aisaIt is very fortunate that we are using mediawiki, as if there was going to be an easy way to work on it from the shell,21:43
wolfspraulso let's say it's not dead21:43
aisait will happen with mediawiki first.21:43
wolfspraul'unmaintained' :-)21:43
wolfspraulyes agreed21:43
aisaI too wish that mediawiki was a DCSS, like git.21:43
wolfspraulaisa: first for your commit & branding, some feedback21:43
aisaand that I could read/work on it from the command-line.21:43
aisaplease please.21:43
wolfspraulI only see you changing the drink recipe, not the graphical OpenWrt splash screen (the city skyline at night)21:43
aisaah, indeed.21:44
wolfspraulmost important: you should feel at home. if you feel the drink recipe is not good, please go ahead and change it. it's good.21:44
aisaI haven't approached the boot image yet.21:44
wpwrakwhat one would need is basically a mediawiki -> HTML renderer. probably with some slight extensions of mediawiki syntax, to make things more natural for a directory hierarchy21:44
wolfspraulaisa: I think we should leave the openwrt splash logo (city skyline) as is21:44
wpwrakmaybe put a better drink ? ;-)21:45
wolfspraulso about branding, my thoughts are that we should not make ourselves look bigger than we are21:45
aisawpwrak: heh, or wait for the next release ;-)21:45
wolfspraulwe (me) have no capacity to maintain any software, really, for the long run21:45
wolfspraulit's a whole OS we are talking about!21:45
aisawolfspraul: I did wonder about this.21:45
wolfspraulwe need help help help21:45
aisabecause lots of people work in this space,21:45
wolfspraulso the only way to do that is to see our resources realistically, and focus on packaging, leveraging, and praising the free software work and projects done by others21:46
aisaand in our community, there is no agreement on what distribution should be on the NanoNote.  There is OpenWrt, Debian, Jlime.21:46
wolfspraulI'm agnostic to that. If one day the production setup permits, I wouldn't have any problem in letting people upload images to our servers, and then in the online shop customers could just click on which image they wanted on their NanoNote.21:46
aisaGood, perfect.  I'm glad to hear you say this, because we already struggle to maintain a working OpenWrt on the NanoNote.21:46
wolfspraulyes, less less less21:47
wolfspraulI only work on doing less21:47
wolfspraulthat's enough work :-)21:47
wpwrakwolfspraul: hmm, interesting. when i generate avt2, i don't get those phantom changes of the transistors in AUDIO21:47
aisaWell said.21:47
wolfspraulso for me personally, I would not have changed the drink recipe21:47
wolfspraulno time21:47
wolfspraulbut like I said, please leave it in if you feel better the way it is now21:47
aisaI will think on this overnight, and in the morning either revert or leave it.21:47
aisaI will likely leave it in my own repo, but this perspective is very helpful to hear.21:48
wolfspraulanother thing, when I picture our user, I always try to think of the 'most normal' person I can come up with21:48
wolfspraulof course that's a cliche, but it helps me focus21:48
aisaI struggle with that, because I bought this thing to hack on.21:48
aisaand create my own end-user experience.21:48
aisaand that is the "person" I know using the device.21:48
aisaAllin and I both (Allin Karhl lives near me) use our NanoNotes for this.21:48
wolfspraulso I picture a mom/dad watching over a kid, or struggling with the kid over not doing homework21:48
wolfspraulor someone taking care of their elderly parents21:49
wolfspraulor a guy under crazy pressure from a crazy boss21:49
wolfsprauletc. etc.21:49
wolfspraullike I said, a cliche21:49
wolfspraulbut then I think 'what could be meaningful to them'?21:49
wolfspraulhow can we explain what we are doing to 'normal' people21:49
wolfspraulis the drink recipe meaningful?21:49
wolfspraulit's cool21:49
aisaLet's go back to the earlier topic,21:50
wolfspraulwhat about the new message you have now? it's very dry, does it really _explain_ what this is?21:50
aisaignore what I just said.21:50
wpwrakwolfspraul: ah, this is the pit of obsolescence i had in mind: http://sharism.cc/blog/21:50
aisano, I borrowed phrases from the wiki and sharism.cc21:50
aisaIt hurts me that that blog is not up to date.21:50
wolfspraulyes! will get to that too21:51
wpwrakwolfspraul: similar problem to the forum: there's a place that looks as if should have up to date content but doesn't. so people can draw the conclusion that the project is dead.21:51
wolfspraulone sec, first about the software thing21:51
aisathat is like the first thing people look at, when they search for the NanoNote.21:51
aisaand they see "Oh, this project is dead."21:51
xiangfuaisa: another feedback about commit.21:52
wolfspraulso my priorities are: do as little as possible myself, reuse free software, reuse free software branding. if anything make the free software more accessible/more understandable to regular people.21:52
wolfspraulaisa: does this make sense?21:52
aisaxiangfu: I'm am all and always ears.21:52
wolfspraulthat's just my approach...21:53
xiangfuwe better add Ben stuff to "data/qi_lb60/files/"21:53
aisawolfspraul: It makes sense on the surface.  I will again sleep on it to come to a deeper appreciation, and I may have a comment after that.21:53
wolfspraulbut please, the best way is if you think from your own perspective. that's fine! if you are the user of the NanoNote you have by far the best intuition of what to change to make it work better for you.21:53
aisaxiangfu: that I can understand, and can move this change there.21:53
wolfspraulsince we all do this collectively we might create a really cool device over time21:53
aisawolfspraul: I wonder about this:21:54
aisaOpenWrt has slightly different goals than Qi Hardware.21:54
aisaspecifically, we have a need to craft a more interactive end-user experience,21:54
aisawith things like web browsers, document readers, &c.21:54
aisaI worry that OpenWrt won't support us in doing this.21:54
wolfspraulthe OpenWrt people I know are all very happy to be able to grow out of the router only/headless corner21:55
aisawonderful!  This was not obvious to me, but is good news.21:55
aisabut then, we "compete" with other distributions that are already in this space, no?  Like opie?21:56
wolfspraulI agree that OpenWrt is weak on the interactive/gui side (in terms of packages), but right now I wouldn't know what more support we would like to see from OpenWrt21:56
wpwraki wonder if it wouldn't be better to push in the direction of jlime. solve the issues with encumbered software. and then make it the standard choice.21:58
wpwrakpro jlime: designed just for the kind of use the ben has, rich environment, mature.21:58
wpwrak(use) both in terms of platform capabilities and user experience21:59
wolfspraulfor me personally OpenWrt works fine, and I will continue with it21:59
wolfspraulI would have no problem flashing and selling Nanos with Jlime preinstalled21:59
wolfsprauleven dual booting22:00
wpwrakcontra jlime: small group of developers/maintainers, not what the "core" of qi distro developers use22:00
wolfspraulbut the problem right now is that our #1 distributor, Tuxbrain, wouldn't want me to ship him devices with Jlime on it, because of patent encumbered software22:00
wpwrakof course, if there is to be a change, making it later only makes it harder. cf. the story of make and the infamous tab in makefiles ...22:00
wolfspraulno change, if anything another option22:01
wpwrakwell, the change would be in what is the recommended distro22:01
wolfsprauland I understand copyleft hardware as good for _any_ free software22:01
wolfspraulOpenWrt is not the 'recommended' distro22:01
wpwraksure. but people often need to know what will give them the most. they may not be able to appreciate all the finer details until much later (if ever)22:02
wolfspraulI never liked the wiki page named 'official software image'22:02
wolfspraulshould be just 'OpenWrt software image'22:02
wpwrakhmm, openwrt is what comes with the box, it's what prominently features in all the documentation, and sharism even pays people to maintain it. doesn't get much more "official" ;-)22:03
wolfspraulaisa: we are diverting a bit, you still there? :-)22:03
aisaI'm watching and waiting ;-)22:03
wolfspraulwpwrak: I will change what comes out of the box any day someone wants to order such units22:03
wpwrakyes, but that's chicken and egg22:03
wolfspraulmy goal is a shop interface that lets you choose which image you want22:03
wolfspraulthat even lets people upload their own images, and when a NanoNote with that image is ordered, they get a royalty :-)22:04
wolfspraulthat would be cool22:04
wolfspraulyes sure it's chicken and egg. so I made the first step with OpenWrt.22:04
wolfsprauland so far I (me personally) am very happy with OpenWrt22:04
wolfspraulit solved some big problems, such as the PATENT flag they introduced, very good work on keeping things upstream, excellent kernel development environment22:05
wolfspraulfor example with Milkymist One or Xue, I will definitely continue with OpenWrt22:05
wolfspraulaisa: for the wiki & reporter stuff, one thing that may not be obvious to you, but where you could still be very helpful, is that you are an English native speaker22:06
wolfspraulQi hardware is a very global community, but many of us struggle in finding the right words to express what we mean in nicely readable English22:06
wpwrakas i understand things, jlime is currently a bit without hardware platform. the original devices are all very obsolete. and they don't seem to feel entirely welcome on the ben.22:07
wolfspraullike the email your friend Allin just sent about nostalgia! :-)22:07
wolfspraulok we need to make them feel more welcome22:07
wolfspraulone idea I has was dual booting22:07
wolfspraulbecause I am not able to offer 2 types of NanoNotes right now, too much work on the logistics side22:08
wolfspraulthe funny thing is, everytime I propose that, when it comes down to really doing it it turns out the patented software is still in22:08
wpwrakthe welcome issue seems to be that jlime can't be openly promoted (at the level of openwrt) due to the encumbered software22:08
wpwrakhowever, they don't seem to be quite sure what it would take to eliminate that concern22:08
aisathis really ought to be an easy thing to fix.  :-922:08
wolfspraulwe had the same problem with OpenWrt, but after long discussions OpenWrt understood the situation and introduced the PATENTED flag upstream, which is very helpful22:09
wolfspraulmp3, mpeg4, h.26422:09
aisaI was playing with config.full_system and noticed that flag.22:09
wolfspraulaisa: are you interested in working with me on the wiki/qi homepage?22:10
wolfspraulnot today, but over the next weeks/months...22:10
aisaI am.22:10
wolfspraulah, great! thanks a lot!22:11
wpwrakhmm, too bad rafa doesn't seem to be around at the moment22:11
wolfspraulif I come across too opinionated, shoot me down!22:11
wolfspraulwe need to try to make it a place as many people as possible feel invited to22:11
aisaI play by the core protocols, which means I prefer doing to not doing, and deciding to not deciding.22:11
aisaIt also means I believe the person doing the work makes the final decision.22:12
wolfspraulthat means you accept to make mistakes, and sometimes to need to pull back and reverse decisions22:12
aisaso you can't be too opnionated for me, but I won't like it if you need me to do things your way.22:12
aisabut I would very much like to do things we agree on, regardless of whose idea it is.22:12
wolfspraulperfect, good plan22:12
aisait also means what you said, above yes.22:13
aisaWhat day and time is it for you?22:13
wolfspraulMonday morning22:13
aisaI am October 3rd, 20:11, which is 8PM.22:14
aisathat is what I thought.22:14
wolfspraultime don't know, let me see... :-) bright sunshine, can tell you that already!22:14
aisaso when I'm in the office tomorrow, it will be Monday evening for you.22:14
aisathis means our best time is your evening and my morning.22:14
aisaas I am not usually online this late.22:14
aisabut am always online early.22:14
wolfspraul10:15 AM22:14
aisawhich is the opposite for you.22:14
aisaour best time meaning our best time to collaborate in real time.22:14
wpwrakthere's definitely something wrong with the schhist process. lemme upload my version of AVT2 ...22:14
aisayou can always send me an e-mail and i'll be back in 12-24 hours.22:14
wolfspraulno rush22:15
wolfspraulaisa: the second goal after the wiki homepage (which is an ongoing task), are the monthly community news22:15
aisaI actually care about this a lot.22:15
wolfspraulbasically as we go through the wiki, and try to distill what is important and how to represent it on the homepage, we will write up what we find, and publish once a month (on the first of each month)22:16
wolfspraulyes I'm working on it right now, doing some plumbing so that the community news can go out via an RSS/Atom feed, into the planet, etc.22:16
aisaI have a question about his.  There is a *lot* of spanish language material on the wiki.22:16
aisamy spanish is very poor.22:16
wolfspraulCarlos students22:16
kristianpaulyeah mostly22:17
aisakristianpaul: should my spanish get better, or will you work on this with me?22:17
kristianpaulinteresting projects around SIE board22:17
wolfspraulthe long term value of this stuff remains to be seen, students also produce a lot of crap which they don't care about after the end of the semester22:17
aisaif I have to learn spanish, it will take me some time.22:17
wolfspraulbut, like kristianpaul says, there may well be valuble stuff in there22:17
wolfspraulI suggest for this content, we always first wait about a year or so22:17
wolfspraulthe students that are just leaving a mess behind, and don't care about their project after they got their grade, well we can't help them, or cleanup their projects22:18
wolfsprauland for those who want to create something more lasting, or want to communicate it to others, we are there to help22:18
kristianpaulwich is the most afaik :( hope be wrong22:18
wolfspraulbut they need to do the first step, I think22:19
wolfspraulotherwise we end up chasing hundreds of students, 9x% of which only care to finish their studies with good grades, the projects are throw-away projects for them22:19
aisaOk, I feel better about my current level of understanding with these projects.22:19
wolfspraulif you see content with questionable licensing, we should flag and delete it right away22:19
wolfspraulthat's the exception to the 'let's first wait a year' rule22:20
wolfspraulso if there are images that are clearly not theirs, and no clear source/author/license info22:20
wolfspraulsame with text that is copy/pasted from non-free sources22:20
wolfspraulor long texts that are copy/pasted from wikipedia (better link to wikipedia then)22:20
wolfspraulon the other hand I am not very actively looking for such content, just glancing over, especially the file uploads22:21
wolfspraulit's relatively easy to delete junk later :-)22:21
wolfspraulaisa: ok I'll do a bit more plumbing work on the community news and feeds today, then we can chat again tomorrow22:22
wolfspraulwpwrak: schhist not working. alright! let's see..22:22
aisawolfspraul: deal.  Thank you for the conversation about my commit, I will make a decision tomorrow (your tonight) and fix or not this commit.22:22
aisaand we can talk about the nesw aggregation.22:22
wpwrakstill uploading ...22:23
wolfspraulnothing wrong with your commit22:23
wolfspraulI appreciate that you make the first step.22:23
wolfspraulit gets us talking, which is great22:24
wpwraki also see one strange change in my version. (there are a dozen or so in the one on projects.) maybe that provides a clue.22:24
wolfspraulaisa: I do think you should look at the NanoNote from the perspective of yourself as the user. That's the only thing that will keep you happy and in a position to quickly improve things.22:25
wolfspraulI'll do the same.22:25
wolfspraulif we have enough people with that approach, who still talk to each other, I believe we create a great device.22:25
wolfspraulwhat do you think?22:25
aisaThis very much works for me.22:27
aisaI will also stay in touch with Allin, the other NanoNote user I currently know.22:27
aisaand as I do news aggregration, I will find other users.22:28
aisaand from this be able to learn where I am clearly a NanoNote user and where I am not.22:28
aisaBut you are correct in that seeing myself as a user will make me more committed to keeping things going.22:28
wolfspraulfor sure22:29
wolfspraulI know exactly what (more) I want from my NanoNote!22:29
wolfspraulwe already work in a way that others can (partially) benefit from what we do for ourselves22:29
wolfspraulthat's why have have all these tools, wiki, irc channel, revision control, upstreaming efforts, etc. etc.22:30
aisaI will speak to everyone later tonight, I am off to get dinner and then sleep.22:32
aisaI had a wonderful and very useful conversation, thank you.22:32
wpwrakfor communication channels, i think less would be more at the moment. i'm sceptical about being able to populate the currently semi-abandoned ones with meaningful content. what makes this worse is that they're the ones that look inviting. so they serve as perfect traps.22:33
aisaagreed.  Less is most certainly more.  I'd like to focus on places we can really contribute value to.22:34
wpwrak(schhist: still uploading ...)22:34
wolfspraulwpwrak: which ones look inviting?22:37
wpwrakthe "official" blog, the forum, ...22:37
wpwrakupload done. your version: http://projects.qi-hardware.com/schhist/board-qi-avt2/22:39
wpwrakmy version: http://downloads.qi-hardware.com/people/werner/schhist/avt2/22:39
wpwraki have a mystery diff in my tree, too. (AUDIO, commit 3cce...) maybe that will shed some light on what's going on.22:40
wpwrakif we're lucky, it's just a bug in ppmdiff :)22:40
qi-bot[commit] Werner Almesberger: Corrected the renumbering warning in sanitize-profile. http://qi-hw.com/p/eda-tools/4c8171722:43
qi-bot[commit] Werner Almesberger: Added AVT2 to the Makefile and the demo index. http://qi-hw.com/p/eda-tools/781c92d22:43
wpwrakhmm. my "ghosts" are real pixel differences. probably rounding errors somewhere, but ppmdiff did the right thing.22:50
wolfspraulwow that looks totally different22:53
wolfsprauleven the commits are different22:53
wpwrakthe jury's still out on the one on my side. i see differences in the postscript as well. maybe be a real change after all. we'll see.22:56
wolfspraulhmm, I guess the commits are different because ghost diffs are picked up22:56
wpwrakyes, you have a lot of false positives there22:56
wpwrakwhat's odd is that the same change repeats22:57
wolfsprauloh, the >>> links are also not correct in my version22:58
wolfspraulI'll start with that22:58
wpwrak$! ;-)22:58
wpwrak$1 even22:58
wolfspraulno it's because of ', which is because of {}23:06
wolfspraulcan you run the scripts over mmone-jtag-serial-cable?23:06
wolfspraulI'm getting an error in sanitize-profile line 5823:06
wpwrakadding ...23:08
wpwrak_@%$#^! brownout23:12
wpwrak_and it killed the scan, too. 26 hours :-(23:14
wpwrak_the mm1 problem is that .pro wasn't committed. lemme catch that ...23:14
wolfspraulargh. that's bad [scan]23:19
wpwrak_yup. it'll delay completion of the main pcb by about a week. i need to make some pcbs. so i'll switch the mill back to being a mill for probably the rest of the week.23:21
wolfspraulat least you don't suffer from over-capacity :-)23:22
wolfspraulyour factory is running 100%23:22
wpwrak_yeah, the scans keep the little machine quite busy :)23:23
wpwrak_already had to oil it a few times. next will be some more thorough cleaning. (i'm beginning to worry about cuttings from milling jobs)23:24
wpwrak_then i should also check the tension of the cable. not quite sure how to accomplish that. (the cable needs retensioning from time to time)23:24
qi-bot[commit] Werner Almesberger: Make schhist more forgiving if profile or top-level sheet is missing. http://qi-hw.com/p/eda-tools/02bb69623:32
wpwrak_mm1-jtag should be good now. schhist only generates a single commit, but i guess that's the best we can do in this case.23:33
wolfspraulmore are coming, I'll try to get this hooked into kicad-libs as well23:35
wolfspraulit's a simple little board anyway, but such tiny projects are also nice to introduce newbies into the system, servers, processes, etc.23:35
wpwrak_you mean to trigger a rebuild of everything if kicad-libs changes ?23:36
wpwrak_yup. "hello world"23:36
wolfspraulfor example I want to document the steps from kicad files to pcb & smt factories better, generation of gerbers, ai files, and so on23:36
wolfspraulI'll do that with such small projects first23:36
wolfspraulgood point, I hadn't thought about it23:37
wolfspraulyes I guess if kicad-libs is updated, everything should be updated23:37
wolfspraulin my script, so far I only make sure that kicad-libs is always checked for updates, and before any other projects23:37
wolfspraulbut if it is updated, everything else should be too23:38
wpwrak_actually, kicad-libs should follow the date of the project being diffed23:38
wolfspraulyeah that would be ideal, but maybe for later...23:39
wolfspraulwe have too many things already, really hard to prioritize23:39
wpwrak_rebuilding the latest version can be tricky. if it's a change that comes from kicad-libs, it should ideally have no effect. if it does, that may be known and the project just hasn't reacted yet23:40
wolfspraulok then I leave it as is right now23:40
wpwrak_if the change comes from the project, the kicad-libs update may just be a misordered push23:40
wolfspraulI first fix that >>> url problem23:40
wolfspraulthen look into the ghosts23:40
wolfspraulURL problem is fixed, switched from single quotes to double quotes23:44
wolfspraulsanitize-profile also good now23:44
--- Mon Oct 4 201000:00

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