Monday, 2014-07-14

*** tpb has joined #timvideos00:00
*** mparuszewski has quit IRC01:12
mithroshenki: the FPGA is too small01:14
mithroshenki: these boards are the same only better01:14
tpbTitle: FPGA boards, Silicon On Inspiration webstore (at
mithro <- that is the equivalent01:15
tpbTitle: Spartan 6 LX9 + 32MB DRAM, Silicon On Inspiration webstore (at
mithrofound via Joelw :)01:15
shenkimithro: they don't have any HDMI io01:18
mithroshenki: true, but they have a very nice expansion port01:19
shenkimithro: i can upgrade it for $30 to a LX2501:31
mithroLX25 is much better01:32
mithroI would highly recommend it, but $30 is a just a tiny bit more than I'd pay for it.01:32
shenkithe netv is a lx901:34
shenkimithro: i don't know the product line off by heart. does lX9 vs 25 vs 45 just give you more gates?01:36
shenki"logic elements"01:36
shenkiwhatever we're calling them today01:36
JoelwYou get more DSPs, BRAM and I/Os, too.01:47
JoelwOh, and more clock management and MCBs01:48
*** tija has joined #timvideos03:19
mithroshenki / tija: were are we at getting your debug code merged?03:25
tijamithro: Currently, the debug output is displayed using gtkterm. I am writing a program so that it outputs in human readable format.03:30
mithrotija: what you writing it in?03:31
tijamithro: in C03:31
mithroOkay. C or Python were the two acceptable answers :)03:32
shenkias far as the existing pull request goes, we needed to add copyright headers03:36
shenkiand then it can get merged03:36
shenkitija: do you know how to edit commits in git?03:37
tijashenki: you mean via browser itself?03:37
shenkitija: nah03:37
shenkitija: using interactive rebasing03:37
shenkigit rebase -i03:37
shenkitija: do you want to learn, or do you want me to fix up your commits?03:38
tijashenkI; I will fix it up, no worries03:38
tijashenki: vc?03:39
shenkioh, it's that time03:39
* shenki has been awake a lot the past 24 hours03:39
shenkiwatched tour de france, then slept for a few hours, and woke up to watch the world cup final03:40
shenkithen went straight to work03:40
tpbTitle: Google+ (at
shenkitija: so, here's what i suggest you do:04:00
shenkigit clone
tpbTitle: shenki/HDMI2USB · GitHub (at
shenkicd HDMI2USB04:00
shenkigit checkout ajit-debug04:00
shenki*hack* *hack* *hack*04:00
shenkigit commit -av -m"Awesome improvements"04:01
shenkigit remote add ajit-github git://
shenkigit push ajit-github ajit-debug:debug04:01
shenkiso the end result is my patches + your awesome improvements will end up in your github repo, in a branch named 'debug'04:02
tijashenki: thanks04:02
shenkiwhen you break it, just give me a yell04:03
shenkias long as you don't type the words '--force' or 'rebase', you shouldn't lose any data04:03
tijasorry but what does this mean "give me a yell"04:04
Joelw'just tell me'04:04
tijaokay I will break and yell :P04:05
shenkithanks Joelw04:05
*** tija_ has joined #timvideos04:55
*** aps-sids has joined #timvideos05:23
*** aps-sids has quit IRC05:35
*** aps-sids has joined #timvideos05:40
tija_Did anyone notice this?
tpbTitle: Installing Exar Driver · timvideos/HDMI2USB Wiki · GitHub (at
shenkitija: yep, looks good05:42
shenkitija: it inspired me to hack on the driver while watching the world cup05:42
tija_shenki: you mentioned you got pull request for this. :D05:43
tija_I mean the exar_driver repo05:43
shenkitija_: i did, yeah, to package it for debian05:53
shenkitija_: i don't think it's very useful though, as it doesn't cooexist with the existing cdc-acm driver05:53
*** aps-sids has quit IRC06:02
*** aps-sids_ has joined #timvideos06:02
mithroYes I did.06:04
shenkitija_: have you pushed your debug program?06:04
tija_shenki: give me an hour  making some changes06:05
*** aps-sids_ is now known as aps-sids06:06
shenkitija_: ok. don't forget to push it :)06:07
tija_shenki: yeah sure06:07
aps-sidsmithro: I'm not able to figure out where the code for some of the flumotion components is coming. For example, x264-encoder
tpbTitle: streaming-system/tools/flumotion-config/pusher/encoder.xml at master · timvideos/streaming-system · GitHub (at
aps-sidsmithro: ah, got it. flumotion-ugly06:13
mithroshenki: I have two Zybo development boards now06:14
shenkimithro: shiny!06:16
mithroThey are based off the series 7 technology06:17
mithroI'm interested to see if it is possible to do 1080p on their pins06:17
shenkii imagine the ARM could run a software h264 encoder too06:21
shenkiperhaps not for 1080p06:21
shenkimithro: do you have them in your hands?06:21
mithrothey are way smaller than I would have expected06:22
mithroI will be bringing them to Adelaide06:22
*** tija has quit IRC06:33
mithroshenki: what license header are you proposing tija use?06:33
shenkimithro: not sure. that's part of the reason why i haven't added it myself06:36
shenkimithro: what do you suggest?06:36
mithroThe firmware is currently under a MIT license06:36
shenkiis that what we want new contributions to be put under?06:38
mithroyes, I'd like the repo to stay under the same license as much as possible06:38
mithroI don't know what the correct theory is regarding per-source-file copyright headers06:40
shenkiYeah. I'm a fan of the one liner appraoch06:41
shenkiCopyright 2013 Fred Smith <[email protected]>06:42
shenkiLicened under the MIT Licence06:42
mithroshenki: I think we should probably just be consistent with what currently is done06:43
*** slomo has joined #timvideos06:44
*** slomo has joined #timvideos06:44
shenkiwith less ascii art06:47
mithroASCII art?06:47
shenkiand more 80 line columns06:47
shenkii think copying this format is good:
tpbTitle: The MIT License (MIT) | Open Source Initiative (at
shenkiwhich is almsot what we hav06:47
shenkimithro: the current files look like this:06:47
shenki-- //////////////////////////////////////////////////////////////////////////////06:47
shenki-- /// Copyright (c) 2012, Jahanzeb Ahmad06:48
shenki-- /// All rights reserved.06:48
shenki-- ///06:48
mithroshenki: I think consistency should be our biggest goal at the moment06:48
mithroshenki: we can do a programmatic transform at a later date if everything is consistent06:48
shenkiit's not consistant06:49
shenki //  Date - revision : Jan. 2008 - v 1.006:49
shenki //  Xilinx, Inc. 2008                 www.xilinx.com06:49
shenki //  Description :     TMDS encoder06:49
shenkibut okay06:50
shenkiwe will cut paste from the other files06:50
mithroshenki: SGTM07:13
mithroI need to do an audit to understand everything07:13
tpbTitle: guvcview_video-2.mkv - Google Drive (at
tija_the HDMI0/1 is toggling between connected/disconnected which it should not08:07
shenkitija_: nice08:26
shenkitija_: where's the code? :)08:26
*** tija_ has quit IRC08:29
*** tija has joined #timvideos09:02
*** slomo has quit IRC09:05
*** tija has quit IRC09:29
*** wanig has quit IRC09:31
*** wanig has joined #timvideos09:33
*** slomo has joined #timvideos09:43
aps-sidsmithro: I need to port components in flumotion-ugly. Should I add them to our flumotion repo or should I commit in flumotion-ugly itself?09:56
*** tija has joined #timvideos10:16
mithroaps-sids: you should have your own fork of flumotion-ugly10:19
mithroaps-sids: have you chatted with CARAM_ recently?10:19
aps-sidsmithro: no10:19
mithrookay, I'd like him to package your work in a docker container10:20
mithroaps-sids: so, if you could guide him in getting that working, it would be awesome10:22
mithroaps-sids: I'd like him to build off Ubuntu Trusty10:23
mithroaps-sids: is there still that twisted issue?10:23
aps-sidsmithro: yes, some functionality of twisted that is used in flumotion was removed after twisted 1110:24
mithroaps-sids: Can you see if you could find a way to fix that?10:24
mithroaps-sids: I'm sure there must be an alternative way to do it10:24
aps-sidsmithro: I can try. Isn't porting the remaining components being used a priority?10:25
mithroaps-sids: well, that is a hard question10:25
mithroaps-sids: I think spending half a day to investigate the amount of work to fix the twisted issue is worth half a day delay in the remaining components10:26
aps-sidsmithro: one more problem. I'm working on "ported flumotion" by using ./env bash and activating the env in code and not actually installing this flumotion version. Now, I'm not sure how to use ugly components with this version since they are installed for system installed flumotion. I'm not sure if I'm being clear here.10:26
mithroaps-sids: that is a good question10:27
mithroaps-sids: yeah, I understand what you mean10:27
aps-sidsso, any ideas?10:29
mithroaps-sids: what env does the bash set?10:32
aps-sids[email protected]:~/workspace/gsoc/flumotion-orig$ which flumotion10:32
aps-sidsenv is in flumotion-orig folder10:32
*** tija has quit IRC10:40
*** rohitksingh has joined #timvideos10:40
mithroaps-sids: what I was trying to determine was if the environment variables are paths or not10:43
mithroaps-sids: IE can you just whack /home/aps/workspace/gsoc/flumotion-ugly into them?10:43
mithroaps-sids: is there any documentation in flumotion ugly?10:45
mithroaps-sids: do you still need the old flumotion installed anymore?10:46
aps-sidsmithro: Here's the env file -->
tpbTitle: Ubuntu Pastebin (at
aps-sidsmithro: I use old flumotion just for using the flumotion-admin10:46
aps-sidsso I can make config changes on the go10:46
mithroaps-sids: what is preventing the porting of flumotion-admin?10:47
mithroaps-sids: do you know how much work is involved in that?10:47
aps-sidsmithro: that involves a lot of work itself. we need a replacement for python-kiwi (which is unmaintained) and also need to port whole gui to Gtk310:48
aps-sidsmithro: this thing works?
tpbTitle: timvideos/flumotion-extra · GitHub (at
aps-sidsFlumotion WebUI10:50
*** tija has joined #timvideos10:52
*** tija has quit IRC11:04
mithroaps-sids: poorly11:10
mithroaps-sids: if we could improve that, it would be pretty awesome11:46
aps-sidsmithro: One thing at a time :P11:47
mithroaps-sids: yeah11:47
*** rohitksingh1 has joined #timvideos11:47
*** rohitksingh has quit IRC11:47
mithrohey rohitksingh111:48
rohitksingh1mihtro: Hi!11:51
rohitksingh1*mithro: Hi!11:51
mithrorohitksingh1: how are things going?11:51
mithrorohitksingh1: I was pondering if we could connect clamp / coast to the two I2C data lines. As I2C is only going to be clocking data in when the clock changes and during those periods we are clocking data we probably don't care too much about coast / clamp?11:52
rohitksingh1mithro: I'm on Plan B now. Writing FIFO based implementation for VGA Capture. Coding almost complete. Yesterday ran simulation which looked great. Will  run on hardware tonight.11:53
rohitksingh1mithro: those CLAMP and COAST are input (to AD9984A)11:54
rohitksingh1mithro: so, if any input is detected, the IC will clamp the ADCs or COAST the PLL (whichever input is active)11:54
rohitksingh1mithro: Which will result in erratic output (atleast color-wise). So, we can't connect them to I2C lines11:56
rohitksingh1mithro: But, if we ensure that AD9984A is configured *only* at startup and their is *no* I2C activity after startup, them we may connect them. But, I'll not recommend going this way as we may need to dynamically changed settings or need I2C lines other devices (like EEPROMS)11:58
mithrorohitksingh1: got your schematic somewhere for me to look at?12:00
mithrorohitksingh1: so I'm going to be dedicating from late Friday (18th) till the 30th to work on HDMI2USB related stuff12:02
mithro(full time)12:02
rohitksingh1mithro: One more *major* problem with this implementation is that I2C lines are pulled-up at IDLE times. So, when there is no I2C activity, those CLAMP and COAST will also get pulled-up. So, this just puts off this plan.12:02
rohitksingh1mithro: Yeah, they are on my Repo
tpbTitle: rohit91/HDMI2USB-vmodvga · GitHub (at
mithrorohitksingh1: so if we could organise to get me a R1 board during that time, I can give it a good whack too12:02
rohitksingh1mithro: yeah. I have got hot air soldering station and soldered the AD9984A, AD3304 and Connectors. Waiting for the AD3334s, Resistor Networks and RF Beads from Element1412:04
rohitksingh1Also, I'd have to do that power supply fix again for this PCB, for it to be working.12:07
mithrorohitksingh1: how long until you get the parts from Element14?12:11
rohitksingh1mithro: They had issues with my Virtual Credit Card. Now, I've re-ordered using my father's Credit Card. The Order status is showing "Received". I hope I get them this week.12:14
mithrorohitksingh1: did you end up chatting with tija about the clocking infrastructure?12:16
rohitksingh1mithro: I figured out it myself :)12:16
mithrorohitksingh1: is it documented somewhere now?12:17
mithrorohitksingh1: could you add it to the developer document?12:17
mithroshenki: flights all booked now12:17
mithrorohitksingh1: another idea, break the pins we don't think we'll use but could still be useful out into a separate header which we could connect to a PMOD if needed?12:18
rohitksingh1mithro: No, they are in my notebook (paper one). I'll add it to the document12:18
mithrorohitksingh1: please do12:18
rohitksingh1mithro: Yeah that seems plausible idea! Will add headers for COAST, CLAMP, OE and PWRDWN. (These are ones we'll not be uing)12:19
mithrorohitksingh1: you should pull them to the correct voltage too12:20
rohitksingh1mithro: Do you mean pulling up the pins to some voltage or adding Voltage Pins and GND to header too?. Here only PWRDWN requires pul-up, rest can be used as-is12:22
mithrorohitksingh1: the second part12:23
mithrorohitksingh1: If nothing was connected to the header, you want things to still work12:23
rohitksingh1mithro: Okay. Got that!12:23
mithroshenki / Joelw:
tpbTitle: Parallella Quick Start Guide | Parallella (at
shenkimithro: to adelaide?12:27
mithroshenki: ?12:27
shenki21:47 <@mithro> shenki: flights all booked now12:27
mithroshenki: yeah12:27
mithroshenki: TripIt should have already shared the flight with you?12:28
shenkimithro: yep, see it now12:30
mithroshenki: great!12:31
* shenki has to decide about pycon12:32
shenkican still get fairly decently priced flights12:32
mithroCARAM_: I got added to the AppEngine Managed VM stuff12:33
*** cfelton_work has joined #timvideos12:58
*** CarlFK has joined #timvideos13:01
*** ChanServ sets mode: +v CarlFK13:01
mithroshenki: Ajit hasn't been doing snippers for almost a week now13:05
mithroshenki: can you poke him?13:05
shenkimithro: i reminded him today, yes13:11
shenkimithro: he took your advice about the copyright header literally13:12
tpbTitle: Debugging program · b91d9f9 · ajitmathew/HDMI2USB · GitHub (at
shenki(note that it's a c file)13:12
shenkimithro: that licence header doesn't correspond to the text of the licence on opensource.org13:13
mithrowhy does that have -- in front of it?13:13
shenkibecause we told him to cut and paste it from the vhdl :D13:13
shenki(i imagine)13:13
mithroha, that won't compile though?13:14
shenkicorrect, it will not13:14
mithrooccasionally I come across things in C which surprise me though :)13:14
mithroshenki: the copyleft URL is borked too13:15
mithrohow does it compare to ?13:15
mithroI can never keep the MIT/BSD licenses straight in my head13:16
mithrothere also appears to be13:16
mithroand ?13:16
mithroshenki: C code should pass "lint"13:16
mithroshenki: you also want to wrap any code comments in ```c\n\n```13:17
shenkiit's simliar to the 2bsd one13:19
shenkii dislike using this web-based review thing. i would prefer a mailing list13:19
mithroshenki: I'm 100% sure it's copied from somewhere13:20
shenkiyes, the 2bsd one13:20
mithroshenki: can you create an issue for us to fix that13:20
mithroshenki: is "kernel style" the same as Rusty's CCAN style?13:34
mithroshenki: do I need to review something?13:36
mithroshenki: pokg?13:43
mithrorohitksingh1: do you want me to review the PCB at the moment?13:43
shenkimithro: an issue to fix what?13:45
shenkimithro: yeah, ccan is pretty much kernel style13:45
shenkimithro: we don't need to follow it to the letter, but something that looks vaguely similar means our code will be readable13:46
mithroshenki: I'm happy to enforce CCAN style13:46
mithroshenki: does Rusty have a "lint" configuration that we can just use?13:46
shenkimithro: not sure13:47
shenkimithro: ccan has a lint tool, but that's about linting a ccan module13:48
rohitksingh1mithro: you can review the schematic. PCB is not routed yet.13:49
mithroshenki: a lot of those tests make sense for us too13:50
mithroI'm reading <-
tpbTitle: Ccanlint · rustyrussell/ccan Wiki · GitHub (at
tpbTitle: Man Page for lint (freebsd Section 1) - The UNIX and Linux Forums (at
shenkiwe could get people to run it through the kernels' script13:54
shenki~/dev/kernels/linus/scripts/ --no-tree -f debugger.c13:55
mithroshenki: SGTM?13:56
mithroshenki: how strict are they? I assume they have all the normal white space checks?13:57
mithrobtw I was looking through all my stuff and found a bunch of ULPI USB modules in my draw13:59
mithromust have ordered them a long time back13:59
mithroshenki: I'm happy for you to drive our C style, but I do want something people can just run and then fix errors (and possibly something we make part of a "presubmit check")13:59
mithroI'm running home now14:00
shenkimithro: it's pretty strict14:05
shenkii think it would catch the major violations, and direct people in the right direction14:05
mithroI <3 code consistency14:07
shenkithis code is awesome14:18
shenkitija is a real hardware dude :)14:18
* shenki is accidently rewriting it14:19
mithroshenki: don't hardware dude's normally write terrible code? :P14:42
mithroanyway, bed time for me14:44
mithrorohitksingh1: sorry didn't get to the schematic today14:44
rohitksingh1mithro: No problem. We can discuss it in VC tomorrow!14:45
mithroaps-sids: did you get anywhere with the possibility of porting the dvswitch plugin to gstreamer1.0 ?14:46
*** tija has joined #timvideos16:12
tijashenki/mithro :
shenkitija: haha16:16
shenkitija: don't take offence at the comments, it's all about learning16:17
tijashenki: none taken :)16:18
shenkitija: no rush to address the issues straight away, but keep he feedback in mind as you keep developing the tool16:18
shenkitija: i have been awake for 20 hours now, im going to bed. damn tour de france. but if you could send me a bitstream that works with the tool, i will test it on my fpga tomorrow16:19
tijashenki: okay will do that16:19
*** rohitksingh has joined #timvideos16:36
*** rohitksingh1 has quit IRC16:39
*** slomo has quit IRC18:02
CARAM_mithro: cool18:53
CARAM_wanna test out an encoder?18:53
CARAM_and see how well it works?18:54
*** slomo has joined #timvideos19:09
*** slomo has joined #timvideos19:09
*** tija has quit IRC19:28
*** cfelton_work has quit IRC19:50
*** mparuszewski has joined #timvideos20:18
*** slomo has quit IRC21:09
*** FeltonChris has joined #timvideos22:34
*** rohitksingh1 has joined #timvideos22:37
*** rohitksingh has quit IRC22:37
*** rohitksingh1 is now known as rohitksingh22:37
Joelwmithro: Yeah! I got my Parallella the other day. Haven't looked at it yet though.23:23
mithroJoelw: Cool23:25
mithroJoelw: are you free the weekend of 26th/27th?23:28
JoelwProbably not at this stage - our project launch has been pushed back a week, so it'll probably be happening that weekend.23:28
mithroWorking weekends, huh?23:32
JoelwNot normally, but the migration process takes a few days and users can't do their work while it's happening so that we can roll back without data loss, so we have to do part of it on the weekend :(23:33

Generated by 2.12.1 by Marius Gedminas - find it at!