Tuesday, 2018-03-13

*** tpb has joined #timvideos00:00
*** CarlFK has joined #timvideos00:36
*** ChanServ sets mode: +v CarlFK00:36
*** Kripton has quit IRC02:56
*** Kripton has joined #timvideos02:57
*** rohitksingh_work has joined #timvideos03:44
*** nancychauhan has joined #timvideos03:54
*** nancychauhan has left #timvideos04:17
*** sb0 has joined #timvideos05:10
*** sb0 has quit IRC05:14
*** CarlFK has quit IRC05:28
*** sb0 has joined #timvideos06:58
*** sb0 has quit IRC07:07
*** rohitksingh_work has quit IRC11:03
*** rohitksingh_work has joined #timvideos11:04
*** futarisIRCcloud has quit IRC11:16
*** harshgugale has joined #timvideos11:21
*** harshgugale has quit IRC11:23
*** rohitksingh_work has quit IRC12:39
mithrocr1901_modern: gah - OSError: Unable to locate Vivado directory or settings.13:19
cr1901_modernWhy is "works on my machine" rearing it's ugly head? Why now T_T?13:30
mithrocr1901_modern: That is the setting Xilinx dir issue13:31
cr1901_modernBut I fixed that in LiteX: https://github.com/enjoy-digital/litex/commit/2b00b7eba4204f22eccaedf59b257609ab99f6b313:36
tpbTitle: xilinx/vivado: Provide a fallback mechanism for using the same root f… · enjoy-digital/[email protected] · GitHub (at github.com)13:36
cr1901_modernFirst off, is it the PR that's failing or what? Because I made sure to incorporate that linked commit before sending my PR13:37
cr1901_modernmithro: ^^13:37
mithrocr1901_modern: It appears not13:38
cr1901_modernmithro: Please link me to exactly where its failing. Now there's _three_ issues that need to be fixed13:39
mithroOf course _florent_ removed all my debugging information about were it looked for the settings file13:39
cr1901_modern* the "TCL file can't be found" issue13:39
* cr1901_modern weeps internally...13:41
cr1901_modernmy fix broke T_T13:42
cr1901_modernSo issue 2 is this: ^^13:42
cr1901_modernAnd issue 3 is...13:42
* cr1901_modern searches for the link13:42
_florent_mithro: if i broke something that was probably not wanted, what have i removed?13:43
tpbTitle: build/xilinx: Fixing settings finding. · enjoy-digital/[email protected] · GitHub (at github.com)13:43
cr1901_modernmithro: Issue 3 is https://travis-ci.org/timvideos/HDMI2USB-litex-firmware/jobs/349061105#L379913:43
mithrocr1901_modern: I've fixed Issue 313:43
cr1901_modernmithro: What was wrong with it then?13:44
mithrocr1901_modern: Your looking at old build there13:45
mithrocr1901_modern: https://github.com/timvideos/HDMI2USB-litex-firmware/pull/41513:45
tpbTitle: WIP: Update to new LiteX by mithro · Pull Request #415 · timvideos/HDMI2USB-litex-firmware · GitHub (at github.com)13:45
cr1901_modernI know, but it seemed like it was my fault, and I still don't know what was wrong with it13:45
cr1901_moderni.e. _You_ fixed something that _I_ broke. :P13:46
mithrocr1901_modern: you just updated litevideo which was broken13:46
mithrocr1901_modern: https://github.com/enjoy-digital/litevideo/commit/9b4169d5d1e2c400a86ea0cbdb800730d84dc40b13:46
tpbTitle: input/clocking/S7Clocking: fix case when split_clocking=False · enjoy-digital/[email protected] · GitHub (at github.com)13:46
cr1901_modernOh yes, that would explain why I didn't explain it.13:47
cr1901_modernmithro: So the reason https://github.com/enjoy-digital/litex/commit/e07bd71b16932a504b7435a96901a3b0eb5a99cd was reverted was compatibility with migen.builf13:48
tpbTitle: build/xilinx: Fixing settings finding. · enjoy-digital/[email protected] · GitHub (at github.com)13:48
cr1901_modernI tried to fix it by providing a fallback to search for extra directories if the migen-compatible behavior failed13:48
mithrocr1901_modern: Yeah -- that change should probably go into migen13:48
cr1901_modernIt was on my todolist (after my PR got merged)13:49
mithroit should really have tests :-/13:49
cr1901_modernI'd rather see less code duplication13:50
cr1901_modernmithro: In any case, the migen-compatible behavior uses a different type signature from your solution in e07db71... looks like I didn't make the fallback behavior all-encompassing, so it broke on a platform I didn'13:53
cr1901_modernt test*13:53
mithrocr1901_modern: Yes - it is super tricky to get right13:53
mithrocr1901_modern: I'm reasonably confident your version is probably broken if you have multiple ISE and Vivado versions installed as well13:53
cr1901_modernCan you explain? I really don't want to install a second copy of ISE/Vivado to test except as a last resort13:55
cr1901_modernmaybe this behavior will just have to diverge in migen/litex for now b/c too much existing code depends on the litex behavior13:56
mithrocr1901_modern: The migen behavior is kind of broken - I think we should just fix in migen?13:56
cr1901_modernmithro: Sure I agree, but...13:57
cr1901_modernend privmsg :)13:58
*** CarlFK has joined #timvideos14:00
*** ChanServ sets mode: +v CarlFK14:00
cr1901_modernmithro: What is broken in the migen behvaior?14:07
mithrocr1901_modern: I should be able to provide the directory I have vivado/ise installed to without needing to know if Vivado or ISE is going to be used14:08
cr1901_modernYes, this is what my patch attempts to fix (by adding an alternate fallback behavior). What makes you think it breaks in the presence of multiple ISE/Vivado versions?14:09
cr1901_modernAnd I'll examine why NeTV fails14:10
cr1901_modernmithro: Check #m-labs channel when you get the chance15:30
*** MK_18 has joined #timvideos16:33
*** shivm_ has joined #timvideos16:38
*** MK_18 has quit IRC16:41
*** felix_ has quit IRC17:27
*** felix_ has joined #timvideos17:28
CarlFK[m]mithro (IRC):  have you seen this group?   Black Mesa Labs  https://blackmesalabs.wordpress.com/2018/02/28/sump2-py-as-a-vcd-waveform-viewer-for-xilinx-vivado-rtl-simulator/18:20
tpbTitle: SUMP2.py as a VCD waveform viewer for Xilinx Vivado RTL Simulator Black Mesa Labs (at blackmesalabs.wordpress.com)18:21
*** CarlFK has quit IRC18:25
*** shivm_ has joined #timvideos18:31
*** CarlFK has joined #timvideos18:36
*** ChanServ sets mode: +v CarlFK18:36
mithroCarlFK[m]: yes18:42
CarlFK[m]mithro (IRC): Peter (ps1 member) was asking what the difference was between the pmod boards you sent me and https://blackmesalabs.wordpress.com/2017/12/15/bml-hdmi-video-for-fpgas-over-pmod/18:44
tpbTitle: BML DVI digital video for FPGAs over PMOD Black Mesa Labs (at blackmesalabs.wordpress.com)18:44
mithroCarlFK[m]: mine drives things directly from the pins, his has an conversion IC in between18:45
CarlFK[m]mithro (IRC):  - k. I guess thats better?  (I'll just pass it on..)18:46
mithroeach has their advantages / disadvantages18:47
CarlFK[m]mithro (IRC): he also said "please don't use such tiny components next time. they are hard to work with."18:47
CarlFK[m]mithro (IRC): is the edid fix still blocked?  https://github.com/enjoy-digital/litevideo/commit/18b88dfee6bf6f4ab55d196747ca00c6c84c2ef219:45
tpbTitle: input/edid: fix scl polarity · enjoy-digital/[email protected] · GitHub (at github.com)19:45
mithroCarlFK[m]: correct - see the cr1901_modern thing above19:45
cr1901_modernmithro: See #m-labs chat19:50
cr1901_modernmithro: If I commit some debug info for paths, could you rerun travis for the NeTV target?19:54
mithrocr1901_modern: Sure19:54
cr1901_modernso I can see what paths litex attempted to search19:54
cr1901_modernmithro: Alright, I'll get on that in 15 mins19:55
*** Kripton has quit IRC21:51
*** Kripton has joined #timvideos21:54
cr1901_modernmithro: I'll be using this branch to stage changes/test stuff. Eventually it'll be squashed onto new-litex, where I'll make the PR https://github.com/cr1901/HDMI2USB-litex-firmware/tree/new-litex-tmp23:19
tpbTitle: GitHub - cr1901/HDMI2USB-litex-firmware at new-litex-tmp (at github.com)23:19
cr1901_modernerr, mithro: https://travis-ci.org/cr1901/HDMI2USB-litex-firmware/jobs/353104732#L362623:56
cr1901_modernwell, that's the wrong commit, nevermind23:58

Generated by irclog2html.py 2.13.1 by Marius Gedminas - find it at mg.pov.lt!