Sunday, 2016-12-04

*** tpb has joined #timvideos00:00
tumbleweedmithro: can I put a backport of my hdmi2usb-mode-switch package in ppa:timvideos/fpga-support ? It's a little different to your one00:24
CarlFKtumbleweed: what's the difference?00:27
tumbleweedit only uses udev, not the unbind helper00:28
tumbleweedthe dependencies are a lot saner00:28
tumbleweedand a few other subtle things00:28
tumbleweedmost were upstreamed, but the PPA package is old00:29
CarlFKcurrently the PPA doesn't have a hdmi2usb-mode-switch package ?00:35
tumbleweedit does, but it won't do the right things00:37
tumbleweedI think00:38
tumbleweed(for our ansible stuff)00:38
CarlFKI'd prefer to keep things 'best' for LCA until after LCA00:39
tumbleweedwhat does that mean?00:39
CarlFKgood Q... I'm struggling to figure out where it fits in the food chain00:40
tumbleweedI don't think there's anything better about what's currently in that PPA package. But mithro might, because he did it that way00:40
CarlFKhdmi2usb-mode-switch is used for flashing firmware.  which should be done days/weeks before the show.    right?00:41
tumbleweedits udev package is also used to set up alias devices, and permissions00:41
xfxfi'm fine using an improved package, just poke me so i can test it01:07
CarlFKxfxf: is the laptop hdmi capture projector opsis going to be near the lectern?01:17
xfxfyes, same as usual01:17
CarlFKxfxf: what sort of system do you plan on hooking the opsis to?01:18
xfxfThe venue's system? Not sure I understand the Q01:18
CarlFKwhat will the opsis usb cable plug into01:19
mithrotumbleweed: Sure I guess?01:19
xfxfA laptop01:19
xfxfSame as last LCA01:19
CarlFKI really like having a SBC in the same box.  but I also have bigger boxes that have room and a place to mount it.01:20
xfxfyes I bought a couple of rPI 3's and power adapter and started looking at 3D printing a mounting kit for it to fit inside of the opsis cases01:20
CarlFKand also I have no idea how to properly install stuff on it.01:20
xfxfBut not for LCA, prefer to concentrate on other things01:20
CarlFKaccording to the people in the room here, not gig e, and even if it is it is on a crappy usb2 buss so you won't get more than a few 10001:24
*** SamSagaZ has quit IRC01:25
tumbleweedCarlFK: which PPA can we put a voctomix-outcasts in?01:27
tumbleweedand voctomix for that matter01:27
tumbleweedwe had one for dvswitch, we should probably have one for voctomix01:30
CarlFKthat sounds ok01:31
* tumbleweed isn't an admin, though :)01:31
tumbleweedso someone else must create it01:31
CarlFKNot allowed here               Sorry, you don't have permission to access this page or     the information in this page is not shared with you.              You are logged in as Carl Karsten.01:39
tumbleweedmithro is the only admin
tpbTitle: Members : “TimVideos.Us - Live event streaming” team (at
CarlFKI don't mind putting -outcasts in my PPA, for my use and testing and maybe LCA don'01:39
CarlFKt care :p01:39
CarlFKbut I have to go to dinner now.  so never mind that01:40
CarlFKxfxf: get odroid C2s $40 each + $5 for the micro sd01:45
CarlFKxfxf: or whatever tumbleweed DC vid found that is an x86 based thing.01:46
CarlFKanway, Imm off for dinner - see ya01:46
*** CarlFK has quit IRC01:58
*** hyadez has joined #timvideos02:02
mithrotumbleweed: You should now be an admin02:40
mithrotumbleweed: You should be admin too02:44
xfxfdeeprave: i'm keen to start integrating the hdmi2usb daemon you wrote into our toolchain for LCA (mostly so we can log what the device is doing in an automated way), is it ready to be used for this use-case?03:32
xfxfimmediate usecase will probably be wrapping it up via systemd & using stdout/stderr from it as a simple log mechanism03:33
tumbleweedyou want to log what people did with the opsis?03:34
xfxfit's so we have more than empirical evidence when things don't work; i.e. actual useful information from production usage that will help with solving issues03:34
xfxfi.e. we have a bunch of boards at the moment that don't work properly in various ways, all we have are scribbled notes on top of each device with vaguely describe what we *think* happened.  obviously during an event you don't want to sit there and hold up a talk and debug exactly what's going on, you just swap it out and on the show goes03:35
xfxfand of course, these problems seem difficult to replicate in non-production03:35
xfxfbut having the debug info off the opsis will help somewhat03:35
xfxfhaving this in place is a pretty hard requirement set by mithro :P03:36
xfxfso that's probably my biggest aim03:36
xfxfwe started working on a supervisor like setup last LCA for monitoring etc but systemd very likely solves it better03:36
xfxfi'm not sure how much i trust systemd to do this, but if you guys have used it successfully then it increases my confidence03:37
tumbleweedwhat's not to trust about it?03:37
tumbleweedinit systems have to be pretty robust, or everything goes pearshaped03:38
xfxfthe usual rantings of those who are staunch init fans :P  i use it myself fine, and i'm already far further along drinking the koolaid given i'm a heavy docker user03:39
xfxfyeah, indeed03:39
xfxfi just tend to take a super conservative view with AV systems, given the scope of what can go wrong03:39
xfxfso i'm slightly hesitant03:39
xfxfbut logically it's fine03:39
tumbleweedfor the job of restarting something if it dies, I'd trust it more than anything I whipped up myself03:39
xfxfindeed, agree03:39
tumbleweedit's the corner cases around that were things break03:39
tumbleweedi.e. hardware that doesn't come back on its own03:40
xfxfyes, i suspect our use of it for what we're doing probably falls into corner case territory03:40
cr1901_modernMy main complaint about systemd, which has fortunately not quite come to pass yet, is that applications start hard depending on it and reducing portability b/c Linux is the One True Unix. And my complaint about Linux is that it's too complicated and I can't confidently grok the source anymore.03:40
tumbleweedcr1901_modern: yeah, I think we all worried about that, but haven't seen much evidence, yet03:41
xfxfdidn't somebody port systemd to one of the *BSD's?03:42
xfxfnot that this is an ideal solution, but it probably is one ;-P03:42
tumbleweedupstream didn't want the patches03:42
cr1901_modernehhh... it's a shim03:42
cr1901_modernI'm fairly okay reading, say NetBSD's source code, by contrast (although yes, there is a learning curve to the BSD source tree layout, I don't think it's any worse than Linux)03:42
*** faulteh[m] has quit IRC03:52
*** faulteh[m] has joined #timvideos04:25
*** CarlFK has joined #timvideos04:44
*** ChanServ sets mode: +v CarlFK04:44
CarlFKtumbleweed: thank you! \o/05:00
*** CarlFK has quit IRC05:04
*** CarlFK has joined #timvideos05:05
*** ChanServ sets mode: +v CarlFK05:05
deeprave@xfxf: re hdmi2usbd - had some fixes to go in, just not sure whether i've pushed them to github yet. Remind me again after Thurs this week, won't have much time until then (and next weekend looks pretty fubar atm too :-/).05:17
CarlFKWhat i the plan for streaming?  I am not really sure what was used last year07:22
jeai'll look into streaming. not too sure what it will be yet08:31
xfxfflumotion - mithro and micolous will know more08:46
mithroI think last year we just used some gstreamer pipelines directly to YouTube08:47
jeathat does sound vaguely familiar08:50
*** rohitksingh has joined #timvideos15:04
*** rohitksingh has quit IRC17:16
*** key2 has joined #timvideos18:31
*** key2 has quit IRC18:43
*** puck`` has quit IRC19:15
*** puck`` has joined #timvideos19:19
*** paradisaeidae_ has joined #timvideos23:05

Generated by 2.13.1 by Marius Gedminas - find it at!