#milkymist IRC log for Monday, 2013-03-11

n0carri3rwhats up guys01:24
n0carri3rbeen a while01:24
n0carri3rwould love to hear whats new with milkymist01:27
kristianpaulhey01:31
kristianpauli also would love to hear whas new with you and your milkymist on the wild :)01:32
n0carri3rhey!01:33
n0carri3rwell, i haven't used it at a gig since last may01:33
n0carri3rbut i'm heading to rio in a few weeks - and i plan to use it01:33
n0carri3rfinally got USB-MIDI working with the nanoKontrol201:34
n0carri3rits pretty awesome01:34
n0carri3rbut i did notice when i went looking, it was *hard* to find documentation on USB-MIDI01:37
n0carri3ri figured it out from the example FNP, which wasn't commented01:37
n0carri3rbut later found werners paper01:37
n0carri3rwhich did lead to one question: how do i use the MIDI out?!01:41
n0carri3r(not USB-MIDI, regular MIDI out)01:41
kristianpauli'm not sure but i tought i was meant to tap a midi cable, as the dmx in out ones i guess01:52
kristianpauli think thats for lekernel to answer..01:53
n0carri3ryeah01:53
n0carri3rwhat do you mean by "tap a midi cable"?01:53
n0carri3rhss3jb: https://bleeplabs.com/store/hss3jb/02:53
larsclekernel: The SimActor source is deasserting it's stb signal for one cycle after it got an ack. This means it will only produce a token every second clock cycle at max09:56
larscIs there any specific reason for this?09:56
lekernelno, I just didn't try to optimize09:58
Fallenoukristianpaul: actually the dmx is not just for sniffing, you can actually send dmx messages with the M110:56
Fallenouyou can control spots etc with dmx10:56
Fallenoulekernel: did you see that n0carri3r came by ?10:57
Fallenouasking about how to play with midi out (real midi, not usb midi)10:57
larsclekernel: btw. the bit width converter works pretty good (at least for the simple cases). But I do have to manually specify input and output width. Is it somehow possible to make a endpoint of variable width?12:06
lekernelyou can do it with AbstractActor, but the rules are hardcoded atm12:22
lekernelit only works with simple buffers etc.12:22
larscok, I'll have a look12:25
GitHub42[migen] sbourdeauducq pushed 1 new commit to master: http://git.io/j24EMQ16:13
GitHub42migen/master 80970b2 Sebastien Bourdeauducq: bus/asmibus: use implicit finalization16:13
lekernelFallenou, yes, saw that. with the current software, midi-out is just a pass through.16:14
Fallenouok16:15
kristianpaulFallenou: ah yes, but i meant it  /sniff/ just for MIDI18:38
lekernelhttp://www.aes.tu-berlin.de/menue/aktuelles/#50935018:41
larsclekernel: I've modified the SimActor like this, which seems to work http://pastebin.com/yzg6nztP. Or are you concerned that the constant writes are going to put to much additional load on the system?19:16
lekernelwell it's not like the simulator is efficient to start with... :)19:17
lekerneland this sort of issue, if it appears, should be resolved at the lower level with e.g. some sort of caching mechanism, so the constant write is fine19:18
lekerneland eventually, we want to get rid of iverilog and simulate fhdl directly from python, so that write would be just an assign19:18
lekernelhave you tested it when the generator returns None?19:20
larschm, on the other hand this probably will cause issues if the generator doesn't yield anything anymore19:20
larscor yields None19:20
larscThis seems to work better: http://pastebin.com/2qetbgeQ19:55
GitHub79[migen] sbourdeauducq pushed 2 new commits to master: http://git.io/mWbtiA21:27
GitHub79migen/master c4d9734 Sebastien Bourdeauducq: README: update21:27
GitHub79migen/master 3c75121 Sebastien Bourdeauducq: fhdl/tracer: remove leading underscores from names21:27
--- Tue Mar 12 201300:00

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