Wednesday, 2017-07-26

*** tpb has joined #timvideos00:00
mithroxfxf: if you stick with LCA firmware, any bug reports or similar are useless00:03
tumbleweedxfxf: it has hdmi2usbmon running if you configure it to do so00:04
mithroHrm, I can't see how listing modes would cause a crash00:04
mithrojea: if the problem really is just listing the modes, then it should only be the firmware.00:05
mithrojea: but I can't see how this is happening, so it is hard to know00:06
mithroAlso, jea you are probably the last person to touch that area of the code with the dual mode additions, so it is possible that it is you who broke it :-P00:07
*** deep-book-gk_ has joined #timvideos00:07
*** deep-book-gk_ has left #timvideos00:10
* CarlFK[m] uploaded an image: file1501028147072.jpg (445KB) <>00:16
CarlFKhalp... I can't keep track of what I am doing :p00:16
tumbleweedtoo many things? :)00:17
CarlFKhey look what I found:   Linux odroid-stretch64 3.14.79+ #1 SMP PREEMPT Thu May 11 15:12:03 CEST 2017 aarch64 GNU/Linux00:18
CarlFKany interest or are you good with c2 stuff?00:18
tumbleweedyeah, I'm good with it00:19
xfxfmithro: i'm happy running the latest firmware as long as you're around to solve issues :P00:27
xfxftumbleweed: actually i want to hear from you on top of CarlFK, should I be redeploying all of my laptops?  did you guys use the usb installer for debconf recently?00:28
CarlFKm l crashes v0.0.3-698-g7b7714d-dirty          built: Apr  2 2017 07:31:0900:28
xfxfi'm hesitant mostly because i'm barely going to get enough time to reinstall all of the machines, let alone debug issues that arise00:28
tumbleweedxfxf: debconf is still ahead of us00:28
xfxfah, right00:28
xfxfi took up a new job a bit over a month ago and it's challenging and interesting but it's kicking my ass time wise00:29
tumbleweedyou don't have to reinstall from scratch, but you'll probably want to push out some ansible changes00:29
xfxfand there's been a stupid amount of unforseen logistical complexities come up with pycon that have been eating my time00:29
tumbleweedif you make big config changes, you may want to reinstall too00:29
xfxfoh, that's interesting news, i should just be able to do an ansible update on the existing image you say?00:30
xfxfbecause i think i tried that on one of my laptops and i tanked the whole thing00:30
xfxfit replaced the repo's with debian repo's00:30
xfxfand totally fscked the dpkg db00:30
tumbleweedsounds like you did the wrong thing :)00:30
xfxfwhat should i be doing?00:30
tumbleweedI don't know what you did00:30
xfxfi am perhaps running the wrong command00:30
xfxfignore what i did00:30
xfxfwhat *should* i be doing00:30
xfxfalmost all laptops are still the same as they were at LCA00:31
tumbleweedthere's an ansible-up command00:31
tumbleweedbut I think you've moved repos around since LCA00:31
tumbleweedso that probably won't work00:31
xfxfalso, what kept me awake last night is realising some of the machines i'm intending on using weren't actually live machines at LCA - so they probably have older installs with some of the bugs we solved during lca00:31
xfxfso i'm gonna have to do something00:31
xfxfand yes00:31
xfxfit's totally broken now00:31
xfxfi ended up pointing it to the new repo's and then it changed the apt sources to debian and reinstalled stuff and broke everything00:32
CarlFKyou should get comforable bringing up one machine with just the on-site role00:32
xfxfCarlFK: you are assuming time I don't have00:32
xfxfthis isn't LCA where we have a one week hackfest leading up into the event00:32
xfxfit's just me until the day before the event00:32
tumbleweedxfxf: sounds like you used the plain playbook without any of the lca config00:32
xfxfwhat's the command i should be typing?00:33
tumbleweeddepends on what you want to do00:33
tumbleweedbut you'll probably want to point at the site.yml in both the playbook and your config repo00:34
tumbleweedand the inventory in that config repo00:34
xfxfat this stage i just want to make sure all of the already installed laptops have any of the latest fixes00:34
xfxfi'm not against reinstalling them, but if problems come up, i'm then in a worse situation due to not having sufficient time to resolve00:34
tumbleweedthat also means you'll get the latest bugs00:34
xfxfthat's the part that concerns me :P00:34
CarlFKxfxf: you are going to spend more time trying to go fast00:34
xfxfi am actually tempted to take both ansible repo's00:34
xfxfrevert them to a date during late LCA00:34
xfxfand deploy those00:35
xfxfCarlFK: no, i'm just trying to take the conservative route - this assumes i have a team before the event to delegate stuff to, i do not, this isn't LCA00:35
tumbleweedyou want to configure some stuff first00:35
tumbleweedlike room names and conference name00:35
CarlFK*you*  should get comforable bringing up one machine with just the on-site role00:36
xfxfCarlFK: you are assuming i have time that doesn't exist00:36
CarlFKreally.. do that.  now.  less chatting and angsting. more bringing up a box00:36
tumbleweedif you want to play with this stuff, it'll take some time getting used to it00:36
tumbleweedit isn't exactly packaged up for people without time00:36
xfxfCarlFK: can't bring up a box when i'm at work and the devices are in a box at home00:36
xfxfand what tumbleweed said00:36
xfxfto be clear, absolutely want to spend time on this and contribute to it00:37
xfxfi just ain't gonna have time to do that in the next week00:37
xfxfthe new job has kinda fscked the time I had assigned to do this stuff00:37
mithroxfxf: I think you have the problem that it is impossible to accurately reproduce the exact setup at LCA.00:37
CarlFKxfxf: this will test the stuff in qemu
tpbTitle: video-stack-deploy/ at master · CarlFK/video-stack-deploy · GitHub (at
tumbleweedthe ansible stuff will get you what was on the machines at LCA mostly00:38
tumbleweed(except secrets)00:38
tumbleweedbut if your environment is not exactly like LCA, that's probably not what you want00:38
xfxfit's close enough to LCA00:39
xfxfsame equipment00:39
xfxfideally i should be doing this properly though00:39
xfxftime to look at my calendar and find who i can piss off the least I guess00:39
CarlFKmithro: m l crashes on v0.0.3-698-g7b7714d-dirty          built: Apr  2 2017 07:31:09 - on Atlys00:40
CarlFKadding to the issue...00:40
mithroCarlFK: do you know how to do a bisect?00:40
CarlFKmithro: yes.00:41
xfxfCarlFK: do you have any tine to see if you can ingest pyconau's json into veyepar?00:42
xfxfif not, i'll attempt today00:42
CarlFKxfxf: url?00:42
mithroOkay, so you should be trying to find an old firmware which works? I think it worked on the LCA firmware, but I would want to confirm that...00:42
tumbleweeddidn't we have a litex firmware at LCA?00:43
mithroCarlFK: then take a firmware half way between the working and know working firmware and test that...00:43
xfxfi thought the litex firmware was what we all should be using moving forward00:43
xfxfand the misoc one is now depreciated00:43
mithroxfxf: correct00:44
CarlFKmithro: do you have a ver number for lca?00:44
CarlFKmithro: or really one to test00:44
mithroxfxf: we used LiteX firmware at LCA00:44
xfxfmithro: ya00:44
mithroCarlFK: it should be somewhere around where we tagged 0.0.300:45
mithroCarlFK: it might be harder to flash the very old firmware, as we didn't have a single flash.bin back at LCA00:46
mithroCarlFK: maybe find the earliest firmware with a flash.bin and try that?00:48
CarlFKmithro: k00:48
mithroCarlFK: Problem is that I don't know the status of most firmware versions, so I can't advise what the correct version to look at are00:50
mithroCarlFK: You might want to put together a wiki page which tracks which versions you have tested and what problems exist in them?00:53
CarlFKwoo, does not crash   git describe: v0.0.0-639-g956a8b3-dirty          built: Apr 11 2016 15:53:0900:55
mithroCarlFK: oh.. that is _ancient_00:56
mithroLook at the build date00:57
CarlFKoh look at the year00:57
CarlFKI did. April :D00:57
CarlFKmithro: btw - that was what was already on one -00:57
mithroI would guess v0.0.3-664-gfab22b9 is somewhere around LCA...00:57
mithrov0.0.3-664-gfab22b9 is before flash.bin existed...00:58
mithrov0.0.3-696-g2f815c1 looks like the first version with a flash.bin ...00:59
mithroCarlFK: Would be good to test that version...01:00
mithroI'm not sure it will work at all however01:01
*** tsglove2 has joined #timvideos01:04
CarlFKmithro: there is no hdmi2usb dir.. just video - should I try anyway?01:07
*** tsglove has quit IRC01:07
mithro ?01:10
tpbTitle: HDMI2USB-firmware-prebuilt/archive/master/v0.0.3-696-g2f815c1/opsis/hdmi2usb/lm32 at master · timvideos/HDMI2USB-firmware-prebuilt · GitHub (at
CarlFKah, Im flashing Atlys01:12
mithroCarlFK: I would always start with Opsis at the moment...01:12
CarlFKmithro: I am doubling up testing them - I thought I had a bad one, but I lost track of it01:14
CarlFKfor sure I have a bad host machine, which may have been the problem01:14
mithroCarlFK: ?01:22
CarlFKmithro: I am preping for PyOhio, also this weekend - trying to work out what boards I am going to use01:23
CarlFKmithro: have 7 Atlys.  one with broken ports, I thought one of the other's was bad somehow, so I am also testing each of them as part of this01:24
xfxfCarlFK: hooray!  do you have time to ingest that schedule or should I?  (i was about to do it so i can knock another thing off the TODO...)01:25
CarlFKxfxf: I'll take a shot at it01:25
xfxfthx, any idea when?01:26
xfxfi wanted to do it asap so i can bug the organisers if data is bad01:26
mithropaddatrapper: I've enabled travis for the HDMI2USB-fx2-firmware repo01:26
CarlFKmaybe in a few hour - I got about 6 hours left in me today, I'll let you know if I am faling asleep01:27
mithropaddatrapper: First build on travis is here ->
mithropaddatrapper: Maybe we should have a separate travis build for each firmware?01:29
CarlFKmithro: Opsis - after I flash, do I need to do something, or wait, before connecting to the console?01:34
xfxfCarlFK: dance01:34
mithroCarlFK: you need to power cycle after a flash01:35
CarlFKmithro: thought I did.. k, v0.0.3-696-g2f815c1-dirty on opsis now01:36
CarlFKmithro: m l crashed it.  built: Mar 31 2017 04:57:2601:37
mithroCarlFK: So it looks like it has always crashed with flash.bin stuff...01:37
mithroCarlFK: before flash.bin you need to flash both the gateware (top.bin) and the firmware (firmware.fbi) files -- but I'm unsure if that works correctly any more....01:44
CarlFKmithro: i would rather not spend more time on that this week01:45
mithroCarlFK: Finding a version which works would be really useful in figuring out what is causing the issue01:46
CarlFKmithro: I know.. but lets ignore that issue for a week01:47
mithroBut I suppose it isn't urgent if it doesn't prevent you using it....01:47
mithroIt does worry me however....01:48
CarlFKim not thrilled, but it is low on my list of things to care about01:48
*** vys has quit IRC02:20
*** vys_ has joined #timvideos02:20
*** vys_ is now known as vys02:21
CarlFKxfxf: do you have a log in that gets you access to the emails?02:21
xfxfi have admin, so probably.  i'm not clear on how that works - is it http auth, or a special token used?02:24
xfxf(how symposium implements the private details in the json feed, i mean)02:24
CarlFKxfxf: put this in
tpbTitle: Ubuntu Pastebin (at
CarlFKadd a closeing }02:31
CarlFKfill in your user/pw02:31
xfxfah, right, sweet.  so i'll do that after you've verified all is OK, and i'll update it with the extra info?02:31
xfxfi had a look through the JSON last week and i couldnt' spot any of the slots nonsense we dealt with at LCA - outside of actual genuine breaks / things we're not recording02:32
CarlFKxfxf: I had to # out reviewer ;)02:44
CarlFKotherwise, looking good
tpbTitle: veyepar - Show: PyCon Australia Pycon_au_2017 (at
xfxfoh ugh, the opening and closing talks are showing as slot02:45
xfxflook at the bottom02:45
xfxfis that something you can fix, or should I bug them?02:45
CarlFKhang on.. room 105 106 0- guessing they get joined?02:46
xfxfthey do yes02:46
CarlFKwhich rooms gear will record then?02:46
xfxfroom 105 is the box that'll be used for recording02:46
xfxfoh wow02:49
xfxfthe keynotes and closing sessions are completely missing02:49
tpbTitle: PyCon AU 2017 | Conference Schedule (at
xfxfnote first and last sessions are not there02:49
tpbTitle: Mini ATX power supply for Raspberry Pi OSH Park (at
CarlFKxfxf: yeah, tell them to fix their stuff ;)02:52
xfxfit's there in the conference.json02:52
xfxflooking at it now02:52
xfxfconf_key 25402:52
xfxfi'm blind02:53
xfxfit's there02:53
xfxfwtf how did i miss that02:53
CarlFKtime zonez my be whack02:54
CarlFKfor lca I had to set the web servers tz to Hobart02:54
xfxfyeah i noticed timezones make things fun02:55
xfxfwhen i added a conference the other day02:55
xfxfi ended up setting the laptop running the schedule ingest to the same timezone as your web server02:55
CarlFKah, that makes sense.. I think...02:55
xfxfthx for ingesting in any case02:55
xfxfCarlFK: the fix for that is to store the timezone as UTC (i assume it already is in django) and have the frontend call JS to render the time in the local users timezone02:57
xfxfbut that requires work02:57
xfxfand who has time02:57
xfxfeither that or letting a user specify their timezone in the user model, and rendering it appropriately02:57
xfxfmithro: i'm struggling to see the purpose of that given the power draw of a rPI is bugger all.  i assume it's for powered projects given the breakout board on the top02:58
CarlFK@#[email protected]#!!!  the 2 are ever so slighty different03:19
CarlFKlike pyonio gives me a list of twitters, one for each presenter - au just gives one item03:20
CarlFKso this happens   'twitter_id': '@p @i @p @c @l @e @a @v @e @s'}03:20
xfxflightning talks shows up as slot, but I can see it fine in the json file, not as a slot?03:23
CarlFKshows up as slot?03:25
*** fishman1 has joined #timvideos03:26
*** rohitksingh_work has joined #timvideos03:34
mithrorohitksingh_work: ping?03:36
rohitksingh_workmithro: pong03:38
fishman1boing and bam03:45
paddatrappermithro: I can't do today until much later, so may have to be Saturday. I should get my umpiring allocations for Saturday tomorrow so will know when I'm available then04:23
mithropaddatrapper: Okay - have you done a weekly status update yet?04:24
paddatrappermithro: thanks for reminding me, I'll do it today04:31
*** deep-book-gk_ has joined #timvideos04:58
*** deep-book-gk_ has left #timvideos05:00
tumbleweedmithro: I see 0.0.3 doesn't have an hdmi2usb.hex any more, does that mean the atlys is flashing the fx2? or should I get it from an older build?06:04
mithrotumbleweed: No, the atlys should still have a hdmi2usb.hex but I only remembered that needed to be fixed when CarlFK did testing recently06:06
tumbleweedbut the old one should work because we haven't changed the interface?06:07
mithrotumbleweed: Correct06:07
*** sb0 has quit IRC06:59
*** sb0 has joined #timvideos07:45
*** rohitksingh_wor1 has joined #timvideos08:34
*** rohitksingh_work has quit IRC08:35
*** fishman1 has quit IRC10:03
*** rohitksingh_wor1 has quit IRC10:15
*** sb0 has quit IRC12:20
*** sb0 has joined #timvideos13:05
*** rohitksingh has joined #timvideos13:17
*** CarlFK has quit IRC14:13
*** CarlFK has joined #timvideos14:14
*** ChanServ sets mode: +v CarlFK14:14
* CarlFK[m] uploaded an image: file1501089429940.jpg (337KB) <>17:18
paddatrappermithro: weekly progress report is up18:51
*** rohitksingh has quit IRC18:57
CarlFKmithro: it would be super helpful to link to someting we can buy18:57
CarlFKsame with power supplies  and cases18:57
CarlFKso we can stop asking each other18:57
tpbTitle: HDMI2USB-numato-opsis-docs/ at master · timvideos/HDMI2USB-numato-opsis-docs · GitHub (at
mithropaddatrapper: so where are we with data transfers?19:01
mithropaddatrapper: do you think you'll be able to get an audio tone soon?19:01
paddatrappermithro: I'm getting some data, though currently it's just what ever is in the FIFO for endpoint 2. I am yet to actually submit my ears to it though...19:10
paddatrappermithro: how do I fill the FIFO internally and not from an external master?19:13
mithropaddatrapper: There are docs for that19:14
mithropaddatrapper: look at the serial example19:14
mithropaddatrapper: I would eventually like you to setup your two FX2 boards back-to-back, one as audio output and one as audio input19:15
paddatrappermithro: that would be pretty cool19:19
paddatrappermithro: the serial example I wrote or the one in the fx2lib?19:19
mithropaddatrapper: btw since I'm up super early would you have a chance to sync up in like 15-20 minutes?19:24
paddatrappermithro: sure19:25
mithroI'm just heading into the office now19:27
mithropaddatrapper: how much longer do at have in GSoC?19:29
mithroAnd when is DebConf?19:30
mithroAnd when do your uni classes start again?19:31
mithroI would like to put together a quick roadmap of what we want to have finished before you go back19:32
paddatrappermithro: Debcamp starts Monday, Debconf finishes 13 August. (though currently I'm still waiting for a visa... So I may not be going) I go back to varsity 14 August. Think final evaluation is end of August19:34
paddatrapper14 August19:35
mithroWhat does that word mean?19:36
paddatrapperAh University, sorry19:36
mithroRegarding DebConf - that really sucks :-(19:37
paddatrapperIt does... I'm hoping it goes through before I'm supposed to fly on Monday, but we'll have to see19:37
mithroI was hoping that you could borrow an Opsis after DebConf19:38
paddatrapperSo was I... E219:38
paddatrapperUgg dropped my phone there...19:39
mithroI would send you an Opsis, but I'm afraid it wouldn't get through the SA post19:39
paddatrapperI may still be able to organise, but it may mean only getting it much later in the year19:39
mithroWith the current supply issue, I don't want to loose any19:40
paddatrapperSo won't help with gsoc, but should help later19:40
paddatrappertumbleweed is coming to SA for pyconza, so can try organise with him. No via-via19:41
paddatrapperThe trick is getting it back to Debconf next year, as I won't be going19:42
mithropaddatrapper: I'm just having some breakfast19:44
paddatrappermithro: no problem. Just shout when you're ready19:45
mithropaddatrapper: So, I think we need to figure out what needs to be done to make your work usable19:45
mithroAugust 21 - 29 16:00 UTCFinal week: Students submit their final work product and their final mentor evaluation19:46
mithropaddatrapper: So it looks like we have exactly 4 weeks until the end of GSoC?19:47
paddatrappermithro: yes. For it to be fully usable it needs the HDMI capture side of it, but I have never even looked at the fgpa code...19:48
paddatrapperWow, that puts things in perspective!19:48
mithropaddatrapper: Yeah :-P19:50
paddatrappermithro: do we want to merge it into the current firmware or leave it separate until the HDMI side is done?19:50
mithropaddatrapper: That is what we are currently trying to figure out :-)19:51
paddatrappermithro: this is true :)19:52
mithropaddatrapper: I'm editing our GSoC weekly meeting notes19:52
* paddatrapper goes to look19:53
tumbleweedpaddatrapper: if debconf-video people are happy with it, we could get you a debconf opsis after debconf / at pyconza19:54
tumbleweedI was planning on sending them back with olasd, but we also need *something* at pyconza, and we should support a team member doing development on it19:55
olasdthere goes my pretense to come to pyconza19:56
olasd(no I'm not serious :p)19:57
paddatrapperolasd: You can always come to pyconza,we have three rooms to set up!19:57
mithropaddatrapper: So - I think by the end of GSoC we definitely want to have the "audio demo" firmware working in both modes19:57
mithropaddatrapper: My rationale is explained in doc19:58
paddatrappermithro: yeah. The two fx2's should be a fairly simple change. Though I need to work out the FIFO slave/master stuff19:58
tumbleweedolasd: if I'm going, you should go :P19:59
mithrotumbleweed: Does DebConf have a MilkyMist TOFE board?20:00
tumbleweedoh? when did we get one?20:02
mithroolasd: You think yes?20:02
tumbleweedah, right20:02
olasdmithro gave us one20:02
mithroDid I give you one at that?20:02
mithroahh cool!20:02
mithroYay past me!20:02
olasdit's in the stuff I intend to put in my suitcase tomorrow20:02
olasdit's not in the import papers so if paddatrapper makes it to DebConf he can go back to .za with it20:03
tumbleweedor send one back with highvoltage20:04
paddatrapperah yes, I had forgotten he was going20:05
mithrohighvoltage is a SA person?20:05
tumbleweedpreviously a montrealer, too20:05
tumbleweedit would make pyconza easier if we could borrow 2 debconf opsis :P20:06
tumbleweedalternatively we could borrow kit from CarlFK again...20:06
CarlFK[m]im good with that20:06
mithroCarlFK mainly has Atlys boards20:07
mithropaddatrapper: So I think we should aim to have good progress on one of the two "potential" options I just listed in the doc20:08
olasdI don't especially mind the debconf hardware being borrowed; it's not like we need them to be idle in a cabinet in my office 11 months out of 1220:08
tumbleweedit is a bit annoying to have them in a far corner of the world, when they're needed20:08
tumbleweedbut we do have lots of opsis now \o/20:09
paddatrappermithro: ok. which one depending on what hardware I can organise and what the limitations on endpoint numbers are?20:09
olasdtumbleweed: far corners of the world like Cambridge20:09
mithropaddatrapper: Not sure yet20:09
mithropaddatrapper: Another options is to finish off / improve the serial port stuff20:11
paddatrappermithro: CDC over USB or USART?20:12
mithropaddatrapper: See the stuff I'm entering....20:12
mithropaddatrapper: I'm almost ready for a VC20:17
mithropaddatrapper: Anything more we should add to the agenda?20:18
paddatrappermithro: Can't think of anything20:19
mithropaddatrapper: Okay, I'm going to find a room now20:22
tpbTitle: HDMI2USB-fx2-firmware/hdmi2usb at master · timvideos/HDMI2USB-fx2-firmware · GitHub (at
tpbTitle: HDMI2USB-fx2-firmware/hdmi2usb at master · timvideos/HDMI2USB-fx2-firmware · GitHub (at
mithropaddatrapper: Don't know if that is correct in any way :-P20:42
mithropaddatrapper: Hopefully you are asleep by now - we should also add a script to the third_party/ which is similar to the one in HDMI2USB-litex-firmware/third_party which updates the submodules21:58
mithroCarlFK: Well, I've been able to repo the video_mode list crash / lockup22:59

Generated by 2.13.1 by Marius Gedminas - find it at!