Wednesday, 2014-07-30

*** tpb has joined #timvideos00:00
mparuszewskimithro: I did what you asked me for: created design documentation for Feedbook (,, deployed prose gatekeeper,, and updated document about deployment process. Now I am going to00:38
mparuszewskisleep, have a nice day! :)00:38
tpbTitle: Feedbook design - Google Docs (at
*** FeltonChris has joined #timvideos00:52
*** Niharika has joined #timvideos01:58
mithroJoelw: don't worry I fixed it. Is there a good tutorial for using chipscope?02:00
mithroJoelw: I'm also not sure what the best way to work with IOBUF and chipscope, it complains when I try and connect to the output of it but doesn't really present the input in an easy to find manner02:01
Joelwmithro: Not sure! I think I used a basic Xilinx one to get started02:02
JoelwYou can't correct it directly to a pad, and internal net names aren't always what you expect them to be02:02
JoelwIf you can't figure out what an internal signal is, one option is to route it to a register and then an external pin (so that it doesn't get trimmed)02:02
JoelwThough you can usually figure out what the thing is actually called, eventually.02:03
JoelwFrom memory, combinatorial outputs will have funny names02:03
JoelwThe other option is to use the CORE generator rather than the inserter, since this lets you just wire it up to signals in HDL02:03
JoelwThat only really works if your signals are all in one spot (e.g. if you route debug signals from all of your modules back to the top level) - bit annoying in practice02:04
JoelwI vaguely recall that they added some sort of pragma that let you give hints to the inserter, but I might be wrong02:04
mithroJoelw: so I want to tap an I2C bus plus a couple of other misc low speed lines02:30
FeltonChrisin VHDL you should be able to use the keep attribute to prevent a net from being renamed02:39
FeltonChrissignal diagnostic_sig : std_logic;02:39
FeltonChrisattribute keep : string;02:39
FeltonChrisattribute keep of diagnostic_sig : signal is "true";02:39
FeltonChrisin Verilog things have changed from /*synthesis keep*/ to (*KEEP=TRUE*) wire foo;02:39
*** mparuszewski has quit IRC02:42
*** Niharika1 has joined #timvideos03:27
*** Palash has joined #timvideos03:28
*** Niharika1 has quit IRC03:28
*** Niharika has quit IRC03:29
*** tija has joined #timvideos04:11
tijashenki: git help?04:14
shenkitija: hello04:16
tijashenki: Hi04:16
*** aps-sids has quit IRC04:18
*** hyades_ has quit IRC04:18
tijashenki: I know how to squash n commits from head and make them 1 using rebase -i. But how to rebase specific commits.04:18
shenkitija: here's the way i do it:04:19
shenkigit rebase -i HEAD~504:19
shenkithen you mark the one(s) you want to fix up with "e" for edit04:19
shenkiwhen you quit the editor, it will stop at the commit(s) you said you wanted to edit04:19
*** mparuszewski has joined #timvideos04:19
shenkiat taht point, you edit the code, git add myfile.vhd, git commit --amend, git rebase --continue04:20
tijashenki: thanks! Even my veteran gsoc friends did not know this :P04:22
mithroafternoon people, at the airport now04:22
tijashenki: Btw I had posted the pipelined image buffer commit yesterday. Have a look04:23
shenkitija: yep! i'm doing some small tidyups to the commit and then i'll push it :)04:23
shenkitija: good work04:23
shenkimithro: hello04:23
shenkimithro: all packed?04:23
mithromithro: I hope so, as I'm at the airport :P04:24
shenkimithro: you're talking to yourself, i take it you didn't get much sleep?04:25
mithroshenki: true04:25
tijashenki: Are you planning to run hdl pretty on all the codes. It is difficult to add exact whitespaces than to remove it.04:27
mithrotija: Yes, onday soon we'll run hdlpretty on all the code and then enable a git-commit-hook which makes sure all new code conforms to that format04:32
shenkitija: yes. once we get all your changes merged in, we will do that04:35
shenkiand then we will add the hook as mithro said04:36
shenkitija: did you open a pull request for these changes? I don't think you did...04:36
shenki(I don't need one, i just want to be sure to close it if there was one)04:36
mithrotija: btw, the plural of computer code is code, like sheep - we don't have "computer codes" for some reason04:36
*** tvCommitBot has joined #timvideos04:37
tvCommitBot[HDMI2USB] shenki pushed 1 new commit to master:
tvCommitBotHDMI2USB/master 8284540 ajitmathew: Memory Read and Write Pipeline...04:37
*** tvCommitBot has left #timvideos04:37
shenkibam. moar fps for everyone!04:38
shenkitija: well done!04:38
*** Palash has quit IRC04:45
tijamithro: point noted04:48
mithrotija: it seems a very common Indian thing to do, to talk about "computer codes"04:49
*** Niharika has joined #timvideos04:58
shenkimithro: what is travis trying to do?
tpbTitle: Travis CI - Free Hosted Continuous Integration Platform for the Open Source Community (at
shenkiit's a bit confused04:58
mithroshenki: I didn't upload a private key to the timvideos/HDMI2USB repo04:59
mithroshenki: but that isn't what those logs how04:59
shenkiit's pulling some strange ref05:10
shenkitija: what code are you rebasing?05:10
tijashenki: the ones committed to ajit-debug-v305:11
shenkitija: okay05:11
tijashenki: you had put some comments05:11
shenkitija: im looking at your heartbeat patch05:11
shenkitija: did you need/want to make updates to that, or is it ready to go?05:11
tijashenki: I am yet to commit that. Everything is working as required by carlFK but haven't commited it.05:12
shenkitija: okay, can you comment that on the pull request?05:13
tpbTitle: Issue #31 : Module for Heart Beat on static images added by ajitmathew · Pull Request #39 · timvideos/HDMI2USB · GitHub (at
tijashenki: comment that "I will be commiting changes soon?"05:14
shenkisomething like that05:15
shenkijust so i know not to fix it up and merge it05:16
shenkitija: there's a line that's commented out in the patch too, make sure you remove that05:16
tijashenki: aye aye captain!05:17
shenkihrm, maybe not. i swear i saw it just before05:17
mithroshenki: travis seems to be pretty confused at the moment05:19
tpbTitle: Travis CI Status (at
shenkimithro: looks like it was broken, and is now fixed?05:22
mithroshenki: pubkey failure now05:23
*** tvCommitBot has joined #timvideos05:29
tvCommitBot[HDMI2USB] mithro opened pull request #79: Adding a ssh step to before_install. (master...travis-fix)
*** tvCommitBot has left #timvideos05:29
mithrogetting on a plane now05:30
mithrowon't probably be back till tomorrow05:30
shenkimithro: k05:32
shenkimithro: safe flight05:32
shenkimithro: it's building, but we're not getting the output on travis.ci05:34
shenki(i know it's building as i can see it running on the host)05:34
mithroshenki: that could have been my pull request.05:42
mithroshenki: we should get your receive hook to tag the directory with the source repo and Travis build number05:43
shenkithe git push gets the ref05:46
shenkiplus the old head, and the new head05:46
mithroTravis sets some environment values, I wonder is there is a way to get them to pass through ssh?05:48
mithroI should get the script to do a ssh-add -L after the key is added.05:49
mithroAnyway departing now05:49
tijashenki: rebasing done06:18
shenkitija: cool06:18
shenkitija: do you have a link?06:18
tijashenki: One problem. Since I wasn't sure about the rebasing method you told me I did it on a test branch. How to bring changes to ajit-debug-v3?06:19
tpbTitle: Commits · ajitmathew/HDMI2USB · GitHub (at
shenkitija: that's fine. perhaps next time just push it to a branch called ajit-debug-v4, but this is fine06:20
shenkitija: can you submit a pull request for that?06:20
tijashenki: the branch is not up to date with all the changes that you have made to master like the memory pipelining. Will a pull request still work?06:22
shenkitija: i think so06:22
*** tvCommitBot has joined #timvideos06:39
tvCommitBot[HDMI2USB] ajitmathew opened pull request #80: Debug Exposed Via CDC (master...testing)
*** tvCommitBot has left #timvideos06:39
tijashenki: Hopefully I have done it correctly.06:41
tijagoing for lunch.06:47
shenkitija: your git client doesn't have your name set07:04
shenkigit config --global "Ajit Matthew"07:04
shenkitija: in these patches i'll fix it for you, but be sure to run that command so future patches are ok07:05
shenkitija: the most recent commit reverts another one; was that intentional?07:06
*** aps-sids has joined #timvideos07:06
*** hyades has joined #timvideos07:11
*** mparuszewski has quit IRC07:22
mithroIn Sydney now!07:27
*** slomo has joined #timvideos07:37
mithroHa, I already had done a ssh-add -L07:43
shenkimithro: that was quick07:45
mithroHad a nap on the plane and dreamt about how to do the debug infra structure07:48
mithroI had an idea about how to do address auto assignment but need to check how well the compiler deals with signals which are determined to be constant.07:51
shenkitija: ping07:59
*** Palash has joined #timvideos08:02
tijashenki: pong08:10
tijashenki: yes that reverts an previous one.08:11
shenkitija: okay. why do we need to do that?08:11
tijashenki: there was a conflict between connect counter to system and Debug module connected to USB commits08:14
shenkitija: okay. i dont understand why your latest commit would be a revert thought08:15
shenkiwouldn't you have had to revert first, and then apply your commits?08:15
*** slomo has quit IRC08:19
shenkiyou branched off of a work in progress branch08:20
shenkithat's kind of my fault for rebasing stuff to fix up your patches08:20
tijado you want me to put my commits in a new branch or want me to force push to existing one?08:22
shenkitija: ive fixed it up, i'll push a fixed branch up in a minute08:23
shenkitija: then if you can reivew that my changes are okay, we will merge08:23
shenkitija: did you see mithro's changes to make the build process output in colour?08:37
tijashenki: No08:39
shenkitija: if you pull the latest master branch down, and type make, you will see it08:41
*** slomo has joined #timvideos08:44
*** slomo has joined #timvideos08:44
shenkitija: once we commit these changes, you should discard your current branches and create a new one on master08:45
shenkitija: that is why we ran into trouble with this set - the commit they were based on was modified in review, so git couldn't merge the branch in08:45
tijashenki: okay I was working on v3 of you branch which is behind your v4 branch.08:48
shenkitija: right, you should work off of hdmi2usb/master08:53
shenkitija: consider any of the branches under my account as temporary08:54
tpbTitle: Commits · shenki/HDMI2USB · GitHub (at
shenkitija: most of it was just whitespace fixes.08:56
shenkitija: i did change this line
tpbTitle: Support 5 addresses for debug data · 852eba5 · shenki/HDMI2USB · GitHub (at
shenkito use or_reduce from ieee.std_logic_misc.all08:57
shenkijust to make it easier to read08:57
shenkitija: also, i fixed a lot of coding style issues with
tpbTitle: Debug Program with CDC · b48eec9 · shenki/HDMI2USB · GitHub (at
shenkitija: we can work together to fix the structure of the program too, but for now i just fixed style issues08:57
shenkitija: i also added a makefile08:57
shenkitija: if you can look through the changes, and okay them, then i will commit08:58
tijashenki: yeah I will look at them08:58
shenkitija: cool :) lets try to get it merged today08:58
*** Palash has quit IRC09:01
tijashenki: I cloned the branch and ran make and I am getting this error.09:11
tija/bin/sh: 7: Bad substitution09:11
tijamake: *** [syn] Error 209:11
*** Niharika has left #timvideos09:13
shenkitija: hrm09:32
tijashenki: I was able to make it using an older makefile09:32
shenkitija: try this from the hdmi2usb dir with the new makefile09:33
shenki$ echo "blah" | tools/colormake.pl09:33
tijashenki: I am getting a "blah" :P09:35
shenkiokay, it works09:36
shenkitija: ls -la /bin/sh09:37
shenkitija: what does that output?09:38
shenki"dash" or "bash"?09:38
*** tija has quit IRC09:44
*** tija has joined #timvideos09:45
*** tija has quit IRC10:00
aps-sidsslomo: I'm not able to understand how to port this piece of code. Can you please take a look?  --
tpbTitle: flumotion-orig/flumotion/component/ at porting-to-gst1.0 · aps-sids/flumotion-orig · GitHub (at
aps-sidsslomo: flag_is_set(), BUFFER_FLAG_IN_CAPS and set_caps() are no longer there10:40
slomoi don't know what the point of that code is10:40
aps-sidsslomo: What I understand is that this is a callback when a Buffer probe is added on sinkpad peer. Then it seems that they check if caps are provided in the buffer and if they are, caps are set on the buffer. Buffers have changed a lot in 1.0 and I don't know the core functionality clearly10:45
slomoi don't know what this code is supposed to solve :)10:46
slomoso i don't know how it should look now10:46
*** tija has joined #timvideos11:04
tijashenki: debug CDC is working, debug UART is working, pipelined image buffer is working, makefile for fpga_debug.c is working, fpga_debug.c is working.11:16
shenkitija: sweet!11:17
shenkitija: so you're happy for your commits to go in with the changes I made?11:17
tijashenki: the makefile for fpga firmware throws an error.11:18
shenki19:07 <+shenki> tija: ls -la /bin/sh11:18
shenki19:08 <+shenki> tija: what does that output?11:18
tijashenki: dash11:19
shenkitija: okay. we can work around that for now, but we will have to get mithro to fix it properly11:20
shenkitija: sudo dpkg-reconfigure dash11:20
shenkitija: then answer "no"11:21
tijashenki: okay11:22
shenkitija: did that fix it?11:30
tijashenki: another error poped up11:30
tija/opt/Xilinx/14.7/ISE_DS/ISE/bin/lin64/unwrapped/map: symbol lookup error: /opt/Xilinx/14.7/ISE_DS/ISE//lib/lin64/ undefined symbol:11:30
tijamake: *** [map] Error 12711:30
shenkitija: hrm. did you start a new shell?11:31
*** tvCommitBot has joined #timvideos11:31
tvCommitBot[HDMI2USB] shenki pushed 5 new commits to master:
tvCommitBotHDMI2USB/master 583699d Ajit Mathew: Debug via CDC Added...11:31
tvCommitBotHDMI2USB/master 852eba5 Ajit Mathew: Support 5 addresses for debug data...11:31
tvCommitBotHDMI2USB/master e9a515b Ajit Mathew: Debug Module connected to USB...11:31
*** tvCommitBot has left #timvideos11:31
tijashenki: with a new shell means bash?11:32
shenkiyeah, a new terminal window11:33
tijaoh new terminal window, no11:34
shenkitija: can you try that?11:35
tijaI thought you were asking did I change from dash to bash11:35
shenkiyeah, do that11:35
shenkibut then open a new window11:35
shenkinot sure why it's necessary. could guess, but it doesn't make a heap of sense11:36
tpbTitle: MAP 13.1 linking error on runtime (at exit) - Xilinx User Community Forums (at
tijashenki: failed again11:37
shenkitija: did you try the workaround suggested?11:37
shenkiexport LD_PRELOAD=/opt/Xilinx/13.1/ISE_DS/ISE/lib/lin64/libAntlr.so11:38
shenkiobviously adjusting that for the version and location of ISE11:38
tijashenki: trying that11:41
shenkitija: ok11:44
tijashenki: failed again11:44
shenkitija: strange. and if you checkout the old makefile it works again?11:44
*** slomo has quit IRC11:45
tijashenki: the new makefile works for you?11:46
shenkitija: yep11:49
shenkitija: it works on the build server too11:49
shenkitija: i assume it works on mithro's laptop, as that's where he tested it11:49
shenkimy laptop and the server are ubuntu 14.04 64 bit11:49
shenkiand i xilinx 14.711:49
tijashenki: now I am getting error for this makefile too11:50
shenkitija: do you have a version of the code that you know works?11:50
tijai mean the old  makefile11:50
shenkitija: do you have an older version laying around?11:51
shenkithat you know has worked in the past11:51
tijayou mean to say older version of makefile?11:51
shenkithe makefile and the rest of the repository, yeah11:52
tijashenki yes11:52
tijashenki: just a while I generated a xsvf using an older makefile for the clone of the branch. Now the same makefile is not working.11:54
tpbTitle: hdmi2usb.xsvf - Google Drive (at
shenkitija: try rebooting your laptop11:58
shenkiperhaps it's unhappy about changing the shell11:58
shenkii can't see why, but it's worth a shot11:59
tijashenki: let me try12:00
*** tija has quit IRC12:00
*** tija has joined #timvideos12:13
tijashenki: after reboot the colourful makefile seems to be working. Atleast it cleared the map phase.12:14
tijashenki: the makefile worked, rebooting did the trick.12:25
*** tija has quit IRC12:29
*** rohitksingh has joined #timvideos12:35
*** FeltonChris has quit IRC13:09
*** slomo has joined #timvideos13:22
*** Niharika has joined #timvideos13:51
*** mparuszewski has joined #timvideos14:52
*** slomo has quit IRC16:14
*** Niharika has left #timvideos17:19
*** Niharika has quit IRC17:19
*** Niharika has joined #timvideos17:30
aps-sidsCarlFK: Can you tell me how does fgdp-producer creates the stream?
tpbTitle: streaming-system/encoder.xml at master · timvideos/streaming-system · GitHub (at
CarlFKaps-sids: um. no clue.  never really looked to see how the code works17:51
aps-sidsCarlFK: ok, but this gets input from collector (present at venue) right?17:52
CarlFKsounds plausible17:52
CarlFKI have looked at chunks of code last year trying to help port it, but didn't pay much attention to how things fit together17:53
aps-sidsCarlFK: actually, I'm not able to find a proper way to test the streaming system components17:54
CarlFKyeah, that would be nice ;)17:54
*** rohitksingh1 has joined #timvideos18:10
*** rohitksingh has quit IRC18:12
aps-sidsshenki: around?18:54
*** Niharika has quit IRC18:54
aps-sidsOh, he too is in Australia :/18:57
CarlFKaps-sids: given there is nothing else, use dvswitch to create a stream18:59
aps-sidsyeah, that's what I'm trying now19:04
aps-sidsCarlFK: how can we use dvsource-file command?19:16
aps-sidsCarlFK: I mean what do we need to install? I remember creating some link with dvsource-v4l2-other but can't seem to figure out now.19:18
aps-sidsnevermind, figured that out.19:30
aps-sidsCarlFK: dvswitch is crashing for me on trusty, when I use dvsource-file and feed a DV from file19:34
CarlFKI would use the VM...  if it doesn't work, prolly easier to fix that than dvswitch on trusty20:29
*** CarlFK has quit IRC23:15

Generated by 2.12.1 by Marius Gedminas - find it at!