#milkymist IRC log for Wednesday, 2011-04-20

lekernelguyzmo: where do you get that?07:42
wolfspraullekernel: I've thought about the 'sharing forum' idea a bit more, and I think initially we don't need a web interface for it08:00
wolfspraulwe just need to add a feature to upload and download patches from a shared public pool (url)08:01
wolfspraulthe best way to access this pool is on the m1 itself, then you can also try the patch immediately after downloading it08:01
wolfsprauland it should be possible to upload a patch to that pool as well08:01
wolfspraulhow does this sound to you?08:01
lekernelok08:02
guyzmo'lo08:38
guyzmolekernel - in compile-lm32-rtems in xiangfu's compile script08:38
guyzmobut I'm running it on OSX, and it may be a platform issue08:38
lekernelif you're on mac it's in macports iirc08:39
guyzmowhat's on macport ?08:40
lekernelthe lm32-rtems toolchain08:40
guyzmonot the lm3208:41
lekernel?08:41
guyzmoI got arm, i386, i960, m68k, mips, powerpc, sh and sparc08:41
lekernelhttps://trac.macports.org/ticket/2742008:41
lekernelhmm seems there isn't gcc... wtf08:42
guyzmoI'm doing selfupdate in case08:42
lekernelask Fallenou08:42
lekernelFallenou: do you have a macports file for gcc?08:54
Fallenouhumm I think so but I didn't send it yet for commit08:54
FallenouDon't have it here sorry08:54
Fallenouatm I just got commited the binutils-lm3208:55
Fallenouthat's something I should put higher on my todo list :) (to submit the remaining parts)08:56
lekernel_ok, just make sure you add the multilib patch09:02
lekernel_http://www.milkymist.org/wiki/index.php?title=Flickernoise_build_instructions#RTEMS_and_toolchain09:02
guyzmoFallenou - have you encountered the missing link I had earlier ?09:05
Action: Fallenou reading the conversation09:06
Fallenouguyzmo: I never used the script to generate the toolchain yet, (on mac or linux or whatever)09:06
FallenouI always did it by hand09:06
guyzmoah ok09:06
Fallenousorry09:06
Fallenouthe script is pretty new ^^09:07
Fallenoumy toolchain is already generated :p09:07
wolfspraullekernel_: for the DMX TX connector whose push button is hard to press down, can you take a look at this url? http://www.neutrik.com/de/de/audio/204_45791/XLR_Einbaubuchsen_productlist.aspx09:42
wolfspraulwe are currently using NC3FAH109:42
wolfspraulthere are two ways to improve the button, one is NC3FAH1-0, which replaces the push button with a 'retention spring'09:42
wolfsprauland the other one is NC3FAHL1, which turns the connector 90 degree to the left, leaving the push button09:43
lekernel_if you want to improve that, turn 90 degrees09:43
wolfspraulwe have enough space, so we can turn the connector. or if you think the 'retention spring' is enough, we can leave it upright.09:43
lekernel_and turn the other connector as well (there's a button on the cable)09:43
lekernel_but then it bumps into the MIDI cables etc. so tbh I'd prefer to leave everything as it is09:44
wolfspraulok, so you think that 'push'/lock mechanism is better than 'retention spring'?09:44
lekernel_i've never had problems unplugging the dmx cable on several occasions09:44
wolfspraulwhy does it bump into the MIDI cables?09:45
lekernel_if you turn the DMX RX 90 degrees too09:45
wolfspraulyou mean inside the case?09:45
lekernel_the button ***on the cable*** bumps into either the other DMX plug or MIDI09:45
wolfspraulyes, on top is best for RX09:46
wolfspraulalso far easier to press than the TX push button, because it's on the cable, i.e. further out (away from the box)09:47
lekernel_it would look weird to have only one of those connectors turned 90 degrees09:47
wolfspraulmy question was about push/lock vs. retention spring, and you answered it - retention spring is worse. correct?09:47
lekernel_probably more trouble than the current position of the release button09:47
lekernel_if you really want to change things, the retention spring is more acceptable than turning only one connector 90 degrees09:48
wolfsprauland if you had to choose between retention spring or push button, and you knew the difficulty of reaching the push button with our case, you would still choose the push button?09:50
lekernel_btw there's also the option of making a notch in the top cover plate of the case09:50
wolfspraulyes, definitely. That's another option.09:50
lekernel_well I know the difficulty. I have used this thing several times already :)09:50
wolfspraulso you would still choose this instead of the retention spring?09:50
lekernel_yes09:50
wolfspraulmaybe at the time we chose this connector we weren't aware of the retention spring solution?09:50
wolfspraulor we were aware and still prefer the push button, the way it is now?09:51
lekernel_ok well09:53
lekernel_use that retention spring connector :)09:53
wolfspraulhe :-) I'm just trying to understand, not kick anybody around.09:53
wolfspraulI also need to defend this product to sell it.09:54
lekernel_hopefully it's even footprint compatible09:54
wolfspraulwe can also leave it as is, push button on top09:54
wolfspraulI guess we have a total of 4 options now09:54
wolfspraul1) push button on top, acrylic as is09:54
wolfspraul2) push button on top, cut out acrylic a little09:54
wolfspraul3) retention spring on top09:54
wolfspraul4) push button on left, without rotating the RX one09:55
wolfspraulif we rotate RX, we cannot reach the unlock button on that cable09:55
wolfspraulmaybe we just go for option #1 until we hear more customer feedback09:57
wolfspraulI'm fine with that. It's not outright bad, it's just a little squeezed in there...09:57
lekernel_or just use the retention sprint version... if it's drop-in compatible just go ahead09:59
wolfspraulok, will check. if it's compatible, we switch to that then.10:00
carlobarhi, i was trying synthesize the HPDCM controler, but i got an error related to the ODDR2. In spartan-3e fpga the DDR_ALIGNMENT="C0" is not supported: http://www.xilinx.com/support/answers/24478.htm10:10
carlobaronly DDR_ALIGNMENT = "NONE" is supported, can i use that option, or is necessary implement  DDR_ALIGNMENT="C0"?  thanks10:12
lekernel_iirc all DDR_ALIGNMENT does is enable an extra register built into the ODDR block... so you might check that and try implementing that register in the fabric10:26
carlobarok, ill do that, thanks10:28
kristianpaullarsc: had you worked with nuttx ?15:46
kristianpaulor knew about it..15:46
kristianpauli just read it was like luclinux but can fit on low memory footprint devices, like microcontrollers15:49
guyzmo\o/15:54
guyzmofinally compiled gcc15:55
kristianpaul:-)15:55
guyzmoactually I had to add --with-libiconv-prefix to ports prefix15:56
guyzmo(as well as mpc, gmp and mpfr)15:56
Fallenoucongratz guyzmo15:56
guyzmo;)15:57
Action: kristianpaul wonders if a netbook can support ISE sistensis..16:02
kristianpaulfor milkymist :D16:02
Action: kristianpaul need a laptop with more speed, 799Mhz are getting short for firefox !!! :(16:03
guyzmokristianpaul - lm32-rtems4.11-gcc: bsp_specs: No such file or directory16:09
guyzmowhat is bsp_specs and where can I find it ?16:10
guyzmo:)16:10
kristianpauldid you compiled the rtems toolchain also?16:12
kristianpaulcompiled and installed16:13
kristianpaulnot just the gcc and binutls16:13
guyzmoah16:13
guyzmojust gcc+binutils for now16:13
guyzmook, let's go for more compilation :)16:13
kristianpaulhttp://milkymist.org/wiki/index.php?title=RTEMS_build_instructions16:13
kristianpaulyup :-)16:13
guyzmobut hm...16:17
guyzmowith the compile-scripts by xiangfu16:17
guyzmoI shall do all that, shalln't I ?16:18
kristianpaulwell, you took already the non-automated steps.. :-)16:25
kristianpaulalready the way*16:25
guyzmowell, not really, I used the script16:30
guyzmobut I hacked the Makefile with the right pathes :)16:31
kristianpaulnon-mmu flat address space is freedom ! :-)17:16
Fallenoudo micro controller =17:33
Fallenou=)17:33
kristianpaul1 2 3 4 - 5618:54
kristianpaulhmm18:54
kristianpaulah wait there is no 518:57
kristianpaullekernel: in conbus5x6.v what CTI means?19:02
lekernelcheck wishbone docs19:14
lekernelstreaming tonight hopefully on http://www.ustream.tv/channel/diemauermussweg19:17
kristianpaulcant you resume CTI from you current ind state?19:19
kristianpaul:-)19:19
kristianpauls/ind/mind19:21
carlobarhi, im trying to simulate the milkymist SoC, but i got this error: lm32_instruction_unit.v:154: sorry: constant user functions are not currently supported: clogb2(). I had asked it before, but i lost the link with the solution, can someone help me again?19:28
kristianpaulah, yes i remenber this chart19:30
kristianpaulchat*19:30
kristianpaulcarlobar: you need replace clogb219:30
kristianpaulsolution is not yet implemented19:30
carlobarhi kristianpaul: do i have to use another function?19:32
kristianpaulactually replace what it does19:32
kristianpaulthat was said by lars i remenber19:32
kristianpaullet me check19:32
kristianpaul10:07 < larsc> carlobar: you could replace the clogb2 call with it's result19:33
carlobarthanks, im going to do that19:42
mwallehttps://bugzilla.mozilla.org/show_bug.cgi?id=647959 << lol :)22:27
kristianpaulhahah22:30
kristianpauli wonder where is the bug for china's gov root CA too :-)22:30
rohkristianpaul: bug? its not A bug.23:20
rohits many.23:21
rohimho the whole idea of a prefilled chain is stupid23:21
rohchina isnt less trustable than t-sytems or verisign. surely not more.23:21
--- Thu Apr 21 201100:00

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