#milkymist IRC log for Monday, 2011-11-28

qi-botThe Firmware build was successfull, see images here: http://fidelio.qi-hardware.com/~xiangfu/build-milkymist/milkymist-firmware-11272011-2319/00:10
stekernhmm, the bte signals of lm32 aren't connected, how does bursting work?00:53
kristianpaul*g*01:52
kristianpauli think we asume linear burst but not sure01:54
kristianpaulalso you dont see many cores implementing it as well01:54
wpwrakhmm. annoying. i connect the usb keyboard via an extension cable and even the rx timeouts go away.02:03
wpwrakextra cables are supposed to make things worse. not better !02:04
qi-botThe Firmware build was successfull, see images here: http://fidelio.qi-hardware.com/~xiangfu/build-milkymist/milkymist-firmware-11282011-0110/02:04
kristianpaulbetter ! cute :)02:07
kristianpaulbalance to the force ;)02:07
wpwrakwell, i'll debug the LV3 then. at least that one is properly misbehaving :)02:08
kristianpaulis the factory that cable came with LV3?02:09
kristianpaulah ok..02:09
stekernkristianpaul: (linear burst) yes, looks like you're right02:10
stekernit also seems it is assumed that if the first access in the burst is in fml cache, the rest of the burst will be so too02:13
stekerndoes lm32 do cache line refills with linear bursts?02:14
kristianpauli read somewhere cache need to be clean, let me confirm02:16
kristianpaulafaik i dont understand this topci very well (fml/L1/l2 caches...)02:16
stekernwpwrak: the errors I get with my kb/mouse are very inconsistent too02:17
stekernsometimes the mouse even works in fn (for a while)02:18
wpwrakstekern: yeah, my plan is to first try to kill all the bugs in can find locally. with a little luck, not too many of the hard to debug remote ones will be left then :)02:18
kristianpaulhaha, i just realized so  the mouse i use (genius traveler) always have a usb extension with it :)02:20
wolfspraulusb extension?02:20
stekernwpwrak: sounds like a good plan02:20
kristianpaulusb cable***02:21
kristianpaulextension cable**02:21
wolfspraulyou mean there is a connector on the mouse's side?02:21
kristianpaulM1 <--- usb extension cable --->|<--- mouse02:23
wolfspraulah ok02:25
wpwrakit's a USB A male to USB A female cable. very non-standard, i think :)02:45
kristianpauloh dhcp is background now, that should improve boot time by seconds02:57
wpwrakor more, if there's no network :)02:58
GitHub115[scripts] xiangfu pushed 1 new commit to master: http://git.io/GdMtQg03:55
GitHub115[scripts/master] fix typo - Xiangfu Liu03:55
GitHub31[scripts] xiangfu pushed 1 new commit to master: http://git.io/ifz-zQ04:34
GitHub31[scripts/master] ignore cvs update - Xiangfu Liu04:34
wpwrakhmm. that didn't sound nice. gunshots in the street ?04:50
xiangfuwpwrak, Hi. I just ordered four DMX devices from taobao:04:57
xiangfuhttp://detail.tmall.com/item.htm?id=624968990104:57
xiangfuhttp://detail.tmall.com/item.htm?id=625218518304:57
xiangfuhttp://detail.tmall.com/item.htm?id=624855562104:57
xiangfuhttp://detail.tmall.com/item.htm?id=631662117304:57
wpwrakwow. big collection :)04:58
wpwraklet's see what they look like ...04:58
xiangfuwpwrak, the   release_2011-11-13 don't include your 2 usb itfs code . :(04:58
xiangfuwpwrak, yes.04:59
xiangfuwpwrak, from my understand the dmx* variables in patches is read only.  and DMX port is like bypass for now. right?05:01
xiangfuwhat is the different between idmx* and dmx*?05:02
wpwraki think idmx is input (from a controller), dmx is output (to a light)05:02
rohdmx has different version05:03
rohs05:03
rohthe old one is one-way .. only from controller to dimmer/lamps etc05:04
wpwrakxiangfu: all this only USD 300 ?!? wow05:05
rohthere is a new variant which also does talkback.. and can do autosetting of addresses etc05:06
rohin the end its all rs485 based with some funny details05:06
xiangfuwpwrak, 366.504 USD :)05:07
xiangfu1 USD = 6.37920 CNY05:07
rohdmx512 is the simple old one i think05:07
wpwrakamazing05:07
xiangfuthere is one patch for DMX out: 'MilkDrop/Rozzor & Aderrasi - Canon (DMX out).fnp"05:16
xiangfubut there are only dmx1/2/3/4. needs learn more about dmx.05:17
wolfspraulxiangfu: please link to that equipment on the m1 accessories page too05:35
qi-botThe Firmware build was successfull, see images here: http://fidelio.qi-hardware.com/~xiangfu/build-milkymist/milkymist-firmware-11282011-0543/06:23
xiangfuthis new build boot time now it's ~16 seconds. (without network cables connect)07:37
xiangfuwpwrak, (about dhcp boot time) ^07:37
xiangfuthe new build. start rendering first. then start dhcp.07:38
xiangfunow it try to dhcp forever. : http://pastebin.com/gq2bSEe307:40
wpwrakso that's good, right ?07:51
xiangfuwpwrak, yes. at least when plug in the network cable. it will get a ip address.07:53
xiangfuwpwrak, before this. we have to manually click a button in GUIU07:54
wpwrakkewl07:54
wpwrakhmm, funny. i increased the number of SETUP retries and made it cycle back to USB reset on error, and now. i'm getting a 24 MHz USB bit clock. i wonder what i did to deserve that ...08:00
GitHub117[flickernoise] xiangfu pushed 1 new commit to master: http://git.io/0YY-Mw09:11
GitHub117[flickernoise/master] use it own language dispaly the Localization setting - Xiangfu Liu09:11
lekernelstekern: yes, the lm32 cache refills with linear burst09:50
lekernelstekern: and it is also - correctly - assumed that if the first access in the burst is in fml cache, the rest of the burst will also be there09:51
lekernelthis kind of assumptions really make you gain time, and results in smaller, faster and less prone to pesky bugs (because of improperly handled corner cases) hardware. of course, the wishbone spec ignores those four aspects.09:54
lekernelxiangfu: to disable translation of a string, prefix it with \e instead of messing the translation tables ...09:57
GitHub71[flickernoise] sbourdeauducq pushed 1 new commit to master: http://git.io/WR1ljQ09:58
GitHub71[flickernoise/master] Revert "use it own language dispaly the Localization setting" - Sebastien Bourdeauducq09:58
xiangfulekernel, just finish a new patch.10:02
xiangfulekernel, http://pastebin.com/9HA6WB0k. if this ok. I will commit it. :)10:08
lekernelDisable translation of a string, prefix it with \e -> To disable ...10:12
lekerneland leave the keyboard layouts translated imo10:12
xiangfuok10:16
xiangfulekernel, should be ok now: http://pastebin.com/mTJKuq9f10:23
xiangfu:)10:23
lekernelok, fine10:49
lekernelI think you can remove "English" from the translation table though10:50
stekernlekernel: yeah, sure, I didn't mean that the assumptions were wrong under the circumstances10:57
stekernbut wrapping bursts would seem more efficient for cacheline refills, but that is a choice made of lattice I guess10:59
qi-botThe Firmware build was successfull, see images here: http://fidelio.qi-hardware.com/~xiangfu/build-milkymist/milkymist-firmware-11282011-1019/11:00
GitHub83[flickernoise] xiangfu pushed 1 new commit to master: http://git.io/2Oljww11:03
GitHub83[flickernoise/master] use it own language dispaly the Localization setting - Xiangfu Liu11:03
lekerneldispaly ?11:05
xiangfuwiat. I commit again.11:06
GitHub1[flickernoise] xiangfu force-pushed master from 0b272f0 to f040ca7: http://git.io/kOuqRQ11:08
GitHub1[flickernoise/master] Use it's own language in the Localization setting - Xiangfu Liu11:08
xiangfuhow about add a handle/grip in the milkymist on box.11:35
lekernel?11:39
lekernelwhat do you mean? a handle on the case?11:39
wpwrakmaybe two little holes so you could run a lanyard through them ? :)11:41
wpwrakof course, a nice big handle, like on a frying pan, would also speak of extravagant style :)11:41
wpwrak"we think this is a really hot product. and we're not afraid to show it."11:42
xiangfulekernel, I mean the bit box. this one: http://en.qi-hardware.com/wiki/Milkymist_One_accessories#box11:42
wpwrakaah, a little plastic handle for the box ?11:43
xiangfuyes.11:44
xiangfus/bit/big/ (typo too much today)11:44
wpwrakhmm. would be nice for carrying the box home from a shop11:45
wpwrakbut may get people to try to use it as a carrying case for m1. is it robust enough for that ?11:46
xiangfuit is not robust enough.11:47
xiangfumaybe we have a expensive box for m1 in future. :)11:47
xiangfulike the expensive one for freerunner. :)11:47
wpwrakhave you seen jon's m1 travel case ? with projector, etc. ?11:48
wpwrak(freerunner case) oh dear. you saw that one too ? ;-)11:48
xiangfujon's m1 case. yes. I saw once. but didn't open it. so only know what is like.11:50
xiangfu(freerunner case) only in picture. :(11:50
wpwrakthere was something even crazier - and external gps antenna. sean presented it to harald and me when i first visited taipei. that thing came with something like 5 m of cable !11:50
wpwraks/and/an/11:50
wpwrakjon's travel case is nice, though. very compact.11:53
wpwrakbut shipping such a box would be $$$11:53
wpwrakin general, it should be possible to find such cases at places where they sell musical instruments or also at places where they sell cameras. there are several types. some have movable rubberized wall elements. these are nice and modular but have limited flexibility.11:55
wpwrakothers just have two big blocks of foam and you have to cut it according to your needs. these are very flexible. but you have to know precisely what you want. well, maybe you can buy the foam separately. should be relatively inexpensive.11:56
GitHub40[milkymist] sbourdeauducq pushed 1 new commit to master: http://git.io/KXR0Wg12:41
GitHub40[milkymist/master] standby: work around intermittent reconfiguration problem - Sebastien Bourdeauducq12:41
qi-botThe Firmware build was successfull, see images here: http://fidelio.qi-hardware.com/~xiangfu/build-milkymist/milkymist-firmware-11282011-1210/12:53
qi-botThe Firmware build was successfull, see images here: http://fidelio.qi-hardware.com/~xiangfu/build-milkymist/milkymist-firmware-11282011-1403/14:44
--- Tue Nov 29 201100:00

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