#milkymist IRC log for Wednesday, 2011-12-07

aw_( second pre-rc4 ) just tested well on reset and 7181C, next to rework usb power switch.02:01
wpwrakaw_: nice ! meanwhile, my parcel is about 2 flight hours from BUE. will probably arrive tonight and hopefully clear customs during the day tomorrow. then it'll sit at some depot until next week because thursday and friday are holidays :-(02:45
aw_wpwrak, great ! Be noticed that I glued with surroundings ( ugly ) on all pins of both usb connectors to strengthen its structure : http://en.qi-hardware.com/wiki/File:M1pre-rc4-u-A019.png02:57
wpwrakaw_: how do you like the general mechanical stability of those USB connectors ?02:59
wpwrak(picture) ah, don't see anything unusual there02:59
wolfspra1lwpwrak: seems rejon has trouble with the mouse on his m103:00
wolfspra1ltoo bad, I don't know much just that he's under stress for the meeting/demo and trying to run out to buy a new one :-)03:00
wpwrakargh. he too ? all MY mice work :)03:00
wpwrakwell, one is a bit bumpy. but it always was. and it's nominally functional, although you wouldn't want to use it03:01
wpwrakso that would be three mouse regressions: stekern, aw_ , and rejon03:01
wolfspra1lI don't know whether rejon has the latest sw on his m103:02
aw_wpwrak, the original Molex USB connector mechanical with four pins are strong while soldering well on pcb, but I 'lifted' up about maybe 2mm high, so it won't be stronger even when I soldered those 4 pins of shielding case.03:02
wolfspra1land I don't think he should update right now, too risky03:02
wpwrakmaybe i should go to the computer shops and get a bag of rodents ...03:02
wolfspra1ldon't know why it suddenly doesnt' work for rejon, too bad03:02
wpwrakaw_: ah ... so does the case still fit for the USB connectors ?03:03
aw_wpwrak, ha ~ no more. so you need to mill those two usb holes by yourself. :-)03:04
wpwrak(upgrades) yeah, we currently have a bit of a mess. lots of things changing at different places, with interfaces between them changing as well03:04
wpwrakaw_: good that it's one of the small sides of the case, so i can do it with my mill :)03:05
aw_wpwrak, yes. :-)03:05
wolfspra1lwell, I can just hope the update button in the GUI keeps working and our update path is functional03:06
wolfspra1lit will not update the standby bitstream, for one03:06
wolfspra1lthen about soc, bios, rtems/flickernoise - don't know03:06
wpwrakthe general update mechanism looks good. the versio numbers probably don't amke much sense, though. better to just use the build date03:06
wolfspra1lis it properly tested? I doubt it03:06
wolfspra1lyes, build date is good03:07
wolfspra1lI don't think we have many users right now that would actually use web update03:07
wolfspra1lthe few we have we have to manually uptick, ask them to come into this channel etc.03:07
wolfspra1lit'll get smoother over time...03:07
wpwrakthey could probably use the web update for anything but standby03:08
wolfspra1lok good, Jon bought a mouse that works03:28
wpwrakgreat ! :)03:29
rohhrhr04:13
rohwpwrak: http://shop.embedded-projects.net/index.php?module=artikel&action=artikel&id=129604:13
rohcheck that schematic... and see how power sequencing is done *g*04:14
wpwrakduring our epic struggles with the PCF506xx, i was actually thinking of a very similar approach to put an end to all that misery :)04:17
rohyeah. i would also try that next time. i mean.. its not that we have had issues with getting such stuff to work as long as it was programmable and works when applying vcc (mcus)04:18
wpwrakyeah. and you have some wiggle room when it comes to things like inrush current. remember Vsys ? :)04:20
GitHub103[scripts] xiangfu pushed 1 new commit to master: http://git.io/DhoAjg05:01
GitHub103[scripts/master] copy the images to patchpool - Xiangfu Liu05:01
aw_( 2nd pre-rc4 ) results after uses of different mouses: http://en.qi-hardware.com/wiki/Milkymist_One_run_3_schedule#USB_Power_Switch_Circuit10:06
aw_seems the mouses from Logitech and ASUS work well.10:07
wpwraki think i'll go mouse chasing next week. go to the local computer mall and get a bag of rodents. then see how they do.10:58
wpwrakARGH. shipment dropped out of the "simplified" process :-(((12:14
wolfspraulwhich one?12:16
wpwrakthe things adam sent12:18
wpwrak"Paperwork available for non-FedEx broker" this is bad news12:18
wolfspraulhmm12:19
wpwrakthis means that they'll only deliver some paperwork (and charge extra for it)12:19
wolfspraulyeah now it's too late12:19
wolfspraul35 USD most likely too low12:19
wpwrakthen i have to go to customs, most likely at the airport, do the 3 doors game, and i'll find out within 2-3 hours what they didn't like. then they'll decide if it's a minor issue (maybe a technical verification, which only adds 1-2 hours) and allow me to proceed. or whether i have to involve a customs broker, which means bribes.12:21
wpwrak(and a second trip to the airport)12:21
wpwrakmaybe there's a hint in the letter fedex will send12:22
wolfspraulaw_: did you put an invoice into the shipment to Werner?12:22
aw_wolfspraul, yes.filled up an fedex formal invoice12:24
wolfsprauldo you have a copy?12:24
aw_no12:24
wpwrakit may not be the invoice. could also be that customs are putting up some other random obstacles12:25
wpwrakbut of course, anything with a value below the shipping cost is automatically suspicious12:26
aw_'bribes' ?12:26
aw_I filled up in that Fedex formal invoice with 'repair' purpose.12:27
aw_not sure if a very low value will be suspicious.12:28
wpwrakaw_: (bribes) argentine customs are very corrupt12:50
wolfspraulI think such a low value of 35 USD for so much stuff is not good. should be 200 USD or so.12:52
wolfsprauland it should roughly list the various things on the invoice, and I would not try anything about 'sample', or 'gift', or repair/damaged/used/etc.12:53
wolfspraulbasically all those words are over-used by people who try to cheat customs, and will be scrutinized a lot12:53
wolfspraulin German customs you can read entire books full of explanations what makes a correct 'sample', and what not12:53
aw_wpwrak wolfspraul hmm... I think i was not clear about Argentina customs, so next time will definitely declare the real price about them.12:54
aw_wpwrak, but the pre-rc4 is a damage board surely with diy soldering, maybe you can tell them those else stuffs are going to 'repair' that Milkymist One ? or customs there don't care anymore just bribes ?12:56
wpwrakwe'll see :)12:57
wolfspraulthere are many types of values12:57
wolfspraulit may not matter whether something passes customs for repair/exchange, whether money is paid or not, etc.12:57
wolfspraulfrom my experience, see above12:57
wolfspraul1) do not price too low12:57
wolfspraul2) do not use sample/gift/exchange/repair/damaged/used12:57
wolfspraul3) include an invoice that more or less lists each bigger item12:58
wolfspraulthen cross fingers and pray :-)12:58
aw_next I'll follow them above though.13:00
wolfspraulyeah but no worries, let's see how it goes now for Werner. not much we can do I think.13:01
wpwraki guess it has to go wrong from time to time so that people believe me that it's tricky :)13:03
wpwrakwhat i don't know is if the trouble floor has changed. e.g., if also things that would have passed in the past now bounce. that can also happen.13:04
wpwrake.g., for a long time, it wasn't allowed to import things that connect to mains that didn't have a argentina certification, but they only started to enforce this something like five years ago13:05
wpwrakand there are probably more such "dormant" restrictions. and of course, they can always make something up if they feel it may be worth it13:06
xiangfuanyone know that it the DMZ of Scopeuk-AFK?14:01
xiangfuI play a little with the DMX controller.14:01
xiangfuI needs buy more DMX cables for connect all those stuff together.14:02
xiangfuhere is the controller  : http://detail.tmall.com/item.htm?id=624855562114:08
kristianpaulwpwrak: do they track you what you import?14:41
kristianpauls/import/receive from exterior14:41
wolfsprauldon't make Werner think14:41
kristianpaul*g*14:41
wpwrakkristianpaul: they should. not sure if they do in all cases, though14:45
wpwraklekernel: i think i'll finish the first round of parser changes today. so far, all inside libfpvm. will you prefer something like 20 individual patches or rather one big patch with everything combined ? (or both ?)16:02
lekernelyou can send one big patch16:04
lekernelbtw, is it against the latest fpvm.c you sent to the ML? (but which I have not merged yet)16:05
wpwrakit would either be based on that set or include it, whatever you prefer16:05
wpwrakthe changes are all over libfpvm, though16:06
lekernelwell, if you changed everything, you can also send a tarball =]16:07
wpwrakbut what could have been the final stroke finally broke patch compilation. so i have to figure out what i did wrong there :)16:07
wpwrak(tarball) hehe ;-)16:07
wpwrakyeah, thinking of it, it's almost that much :)16:07
lekernelbut if you have intermediate local commits already, it's better to keep them16:07
wpwraki currently have ~17 individual commits16:08
lekernelok... then, send 17+ patches :)16:09
wpwrakalright :)16:09
lekernelshould I already merge your 6?16:09
lekernelnote that, those 6 don't include the macro change that mwalle proposed16:10
wpwraki think they're safe to merge, yes. lemme post the macro change ...16:10
wpwraksent16:14
lekernelthanks!16:16
GitHub41[milkymist] sbourdeauducq pushed 7 new commits to master: https://github.com/milkymist/milkymist/compare/36b1e07...2b3952e16:25
GitHub41[milkymist/master] libfpvm: wrap some overlong lines in fpvm.c - Werner Almesberger16:25
GitHub41[milkymist/master] fpvm.c: replace code patterns involving add_isn with macros - Werner Almesberger16:25
GitHub41[milkymist/master] fpvm.c: replace code patterns of register allocation with macro - Werner Almesberger16:25
wpwraki guess now would be the time to figure out how to use rebase to plant the remaining patches on the new head :)16:31
wpwrakhah ! a NULL pointer :)17:01
wpwrakhmm. interesting. a "make clean" later, it seems to be gone17:06
GitHub189[milkymist] sbourdeauducq pushed 2 new commits to master: https://github.com/milkymist/milkymist/compare/2b3952e...81cbd1417:15
GitHub189[milkymist/master] delete all script file. Makefile soon - Xiangfu Liu17:15
GitHub189[milkymist/master] use Makefile manage the build progress cleanup the gcc output, make V=1 will output all detail message - Xiangfu Liu17:15
wpwrakphew. that'll be tricky. we now have overlapping changes :-(17:41
lekernelif it's too messy, just send one diff or a tarball17:44
lekerneldon't waste time with the rcs17:44
wpwrakactually, the changes don't overlap. hmm. just can't pull or rebase17:45
wpwrakif i pull, it complains about fpvm.c. if i try to rebase, it tries to go back to some ancient softusb change17:46
wpwrakand one last build dependencies issue ... now, let's see if it still works18:02
lekernelyeah, the dependency system is lousy, to say the least18:06
lekernelactually, I don't know of any build system that handles that nicely... cmake isn't the best for cross compilations18:07
wpwrakmy dream is a nice top-level makefile that then recurses into everything we need18:07
lekernelyeah, but make doesn't handle #include (and others) automatically18:08
lekerneland updating makefiles everytime you add an include is messy18:08
wpwrakby why do it that way ? :)18:08
lekernelwell, any good tools that do that automatically...?18:09
wpwraka Makefile with a wee bit of instrumentation: http://projects.qi-hardware.com/index.php/p/fped/source/tree/master/Makefile18:09
wpwraksee the %.o: %.c rule (and the link above it)18:10
kristianpaul3 screens wow18:19
wpwrakand i want bigger ones ! :) dell have some nice 23" screens with ~2500 pixels horizontally. and want three of these. plus graphics hardware that can actually drive them18:22
wpwrak11 patches coming :)18:23
lekernelmore lurkers will leave the list =]18:27
wpwrak;-))18:28
kristianpaulhe, i was thinking same time ago better i dint said it first but i agree18:28
kristianpauli mean, guessing people buy M1 suscribe to list18:29
wpwraksomeof thesedays, we really ought to set up a forum18:32
wpwrakthe average user must be scared by all the hardcore tech18:32
kristianpaulalso the patch flood in his imbox ;)18:33
kristianpaulask.milkymist.org :)18:33
wpwrak:)18:39
wpwrakwell, the patch flood will only get worse. now i have relatively simple things for a while. this means many patches per week :)18:40
kristianpaulnice :)18:40
wpwrakjust got a mail from fedex. "A confirmation of the value of goods is required." so it seems it was indeed the declared value18:42
kristianpaulconfirmation = you are LIER !18:48
kristianpaulor, what is probably the solution in this case?18:48
wpwrakjust called fedex. yup. customs didn't accept the declared value.18:54
wpwrakwell i should get the documentation today. one problem is that i have to leave soon, so maybe i'll miss the fedex guy. not sure if he'll leave it with the porter.18:55
wpwrakthen, with the documentation, i have to go to customs. thursday and friday are holidays. so that'll be monday. not sure which customs office, but probably at the airport. that's some 45 minutes drive.18:56
kristianpaulso there is polite way to solve this?18:56
wpwraki'll find out during monday :)18:57
lekernelwpwrak: how much was that declared value?19:06
wpwrakUSD 3519:07
lekernelha, yes... pretty low indeed19:08
wpwrakand also customs know how to google ...19:12
kristianpaullol19:39
kristianpaulgoogle really? ;)19:39
kristianpaulthast game over...19:39
lekernelwpwrak: does this "unique string repository" actually improve performance?20:38
lekernelalso it introduces a list of patch variables into libfpvm :(20:38
GitHub139[milkymist] sbourdeauducq pushed 11 new commits to master: https://github.com/milkymist/milkymist/compare/81cbd14...92eefac21:19
GitHub139[milkymist/master] libfpvm: make scanner aware of function names and handle arity in parser - Werner Almesberger21:19
GitHub139[milkymist/master] parser.y: move node allocation and initialization to shared function - Werner Almesberger21:19
GitHub139[milkymist/master] libfpvm: pass token in struct ast_node - Werner Almesberger21:19
kristianpaulmwalle: but tilling trought all that windows... well,i  prefer the mouse sometimes..21:21
GitHub11[migen] sbourdeauducq pushed 2 new commits to master: https://github.com/milkymist/migen/compare/0e8d894...fab02f821:25
GitHub11[migen/master] fhdl: cleanup value bv - Sebastien Bourdeauducq21:25
GitHub11[migen/master] fhdl: fix implicit slice index - Sebastien Bourdeauducq21:25
GitHub97[datasheets-m1] sbourdeauducq pushed 1 new commit to master: http://git.io/tKKWSg21:33
GitHub97[datasheets-m1/master] Add ADV7181C manual - Sebastien Bourdeauducq21:33
wpwraklekernel: (performance) not sure. i saw compile time go from 6 to 5 seconds at one point, but that wasn't the state before anything else21:57
wpwrak(list of variables) we can get rid of that. but ... both sides should grow together anyway, so it may be natural to share them.21:58
wpwrakbut let's find out ...22:06
wpwrakwhat would you call the variables _Xi, etc.? and ... do you remember any more internal variables that aren't listed in pfv_names, pfv_from_name, and pvv_names ?22:13
wpwrakah, here are more: x, y, rad, and a whole crowd in finalize_pvv22:15
--- Thu Dec 8 201100:00

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