| *** tpb has joined #timvideos | 00:00 | |
| *** Kripton has quit IRC | 00:20 | |
| *** zaratustra15 has joined #timvideos | 00:23 | |
| *** Kripton has joined #timvideos | 00:25 | |
| *** deltab25 has joined #timvideos | 00:29 | |
| *** zaratustra15 has quit IRC | 00:30 | |
| *** annieslmaos has joined #timvideos | 00:38 | |
| *** d__b has joined #timvideos | 00:41 | |
| *** pj1 has joined #timvideos | 00:43 | |
| *** annieslmaos has quit IRC | 00:44 | |
| *** pj1 has quit IRC | 00:45 | |
| *** GrapeNinja has joined #timvideos | 00:49 | |
| *** GrapeNinja is now known as Guest87060 | 00:50 | |
| *** pilottage has joined #timvideos | 00:52 | |
| *** pilottage has quit IRC | 00:54 | |
| *** oldschool^19 has joined #timvideos | 01:04 | |
| *** xfxf has joined #timvideos | 01:04 | |
| xfxf | / | 01:04 |
|---|---|---|
| *** oldschool^19 has quit IRC | 01:09 | |
| *** deltab3 has joined #timvideos | 01:16 | |
| *** deltab3 has quit IRC | 01:17 | |
| *** musician_pro has joined #timvideos | 01:22 | |
| *** musician_pro has quit IRC | 01:27 | |
| *** ascheel4 has joined #timvideos | 01:41 | |
| *** niko7 has joined #timvideos | 01:41 | |
| *** ascheel4 has quit IRC | 01:42 | |
| *** niko7 has quit IRC | 01:43 | |
| *** Pugabyte1 has joined #timvideos | 02:20 | |
| *** Pugabyte1 has quit IRC | 02:21 | |
| *** sekizo25 has joined #timvideos | 02:43 | |
| *** sekizo25 has quit IRC | 02:48 | |
| xobs | Cautious optimism: Much quieter today? | 03:03 |
| *** Turbo-Folker has joined #timvideos | 03:05 | |
| *** rogue2 has joined #timvideos | 03:05 | |
| *** rogue2 has quit IRC | 03:12 | |
| CarlFK | I may have set some stuff so only registered nics can talk | 03:18 |
| CarlFK | which isn't great for new people. but oh well... | 03:18 |
| xobs | I did something similar on #kosagi on OFTC. Also set it +s so it doesn't show up in /list | 03:18 |
| xobs | IRC flags are weird and confusing | 03:19 |
| CarlFK | I just cut/pasted some random stuff that someone told me on the internet | 03:20 |
| *** NightMonkey20 has joined #timvideos | 03:23 | |
| *** danmackay24 has joined #timvideos | 03:23 | |
| *** HollyW00d28 has joined #timvideos | 03:25 | |
| *** Elwell has quit IRC | 03:28 | |
| *** NightMonkey20 has quit IRC | 03:28 | |
| *** danmackay24 has quit IRC | 03:30 | |
| *** HollyW00d28 has quit IRC | 03:31 | |
| xobs | What could possibly go wrong | 03:37 |
| mithro | xobs: I would love to chat about litex-buildenv ideas and thoughts sometime | 03:43 |
| xobs | mithro: Sure. I was just responding to that bug. | 03:43 |
| xobs | We can chat here or there. | 03:43 |
| mithro | xobs: I need to go find dinner first | 03:43 |
| mithro | Should be back in like 15-20 minutes | 03:47 |
| xobs | Alright, enjoy-dinner | 03:47 |
| *** Elwell has joined #timvideos | 03:53 | |
| *** diogenese0 has joined #timvideos | 03:54 | |
| *** diogenese0 has quit IRC | 03:55 | |
| *** rohitksingh_work has joined #timvideos | 03:56 | |
| *** Jovan26 has joined #timvideos | 04:00 | |
| *** Jovan26 has quit IRC | 04:01 | |
| mithro | xobs: Back now | 04:09 |
| *** mww113 has joined #timvideos | 04:10 | |
| *** mww113 has quit IRC | 04:11 | |
| xobs | Welcome back | 04:11 |
| mithro | xobs: FYI -- I wrote this a while back - https://docs.google.com/document/d/1SLRMCTETNKA0Cfy8dExF8QAn8j0FVbUEgorsicJmoeU/edit#heading=h.h2wbgumo1dxi | 04:13 |
| tpb | Title: Type of "litex" tests - Google Docs (at docs.google.com) | 04:13 |
| mithro | https://docs.google.com/document/d/11uzjWRWk9-KuBFc7chUNUluL5ajysY2qfXvt1vttl7k/edit#heading=h.2gyfmvssus6p is also interesting | 04:15 |
| tpb | Title: LiteX Firmware - Google Docs (at docs.google.com) | 04:15 |
| *** cfields19 has joined #timvideos | 04:36 | |
| *** cfields19 has quit IRC | 04:37 | |
| mithro | xobs: So what is your plans around that Python script? | 04:59 |
| *** Elwell has quit IRC | 04:59 | |
| xobs | mithro: The goal of the script was to make it easy to get new users onboard who know hardware but not Python. i.e. bunnie. | 05:00 |
| xobs | It started out as a bash script, then turned into a Python script when I realized it could work on Windows as well. | 05:00 |
| xobs | I want to avoid any modules that aren't part of Python's standard library. | 05:00 |
| mithro | xobs: Yeap | 05:01 |
| mithro | xobs: That is part of the aim around the whole litex-buildenv but we went a little bit to far down the shell script rabbit hole | 05:01 |
| xobs | I added dependency checking because lately I've had to jump around systems a lot, and I would frequently forget to install dependencies such as Vivado or a compiler. | 05:02 |
| xobs | (In fact, I'm getting a replacement machine today, so I'll get to go through the whole exercise again.) | 05:02 |
| mithro | xobs: FYI - Anaconda can take a conda environment and create a single downloadable installer | 05:02 |
| mithro | xobs: Including Python and everything | 05:03 |
| mithro | https://github.com/conda/constructor | 05:03 |
| tpb | Title: GitHub - conda/constructor: tool for creating installers from conda packages (at github.com) | 05:03 |
| *** neptune has joined #timvideos | 05:03 | |
| mithro | xobs: Have you seen apio plus what TinyFPGA has been doing? | 05:04 |
| xobs | mithro: That might be, but I was really turned off by https://github.com/ContinuumIO/anaconda-issues/issues/8203#issuecomment-372984539 -- their dismissive nature insisting that "There is nothing to be done here" when it's really broken. | 05:05 |
| tpb | Title: Due to glibc 2.12 limitation, static executables that use time(), cpuinfo() and maybe a few others cannot be run on systems that do not support or use `vsyscall=emulate` · Issue #8203 · ContinuumIO/anaconda-issues · GitHub (at github.com) | 05:05 |
| mithro | xobs: Well, it's because we are statically linking the compiler | 05:06 |
| mithro | xobs: Which is a bit unusual.... | 05:06 |
| mithro | xobs: https://github.com/FPGAwars/apio | 05:07 |
| tpb | Title: GitHub - FPGAwars/apio: Experimental open source ecosystem for open FPGAs (at github.com) | 05:07 |
| mithro | xobs: https://tinyfpga.com/bx/guide.html | 05:09 |
| tpb | Title: TinyFPGA BX User Guide (at tinyfpga.com) | 05:09 |
| mithro | xobs: I don't know another good way to get a self contained Python environment on Linux | 05:10 |
| xobs | mithro: Why do you need a self-contained Python environment? | 05:11 |
| mithro | Because people's system Python tends to be screwed up in all types of fascinating ways -- plus up until recently Python 3.5 wasn't available on all Linux systems and 3.6 wasn't compatible with misoc | 05:13 |
| mithro | xobs: Plus you still need a way to get distribute the binary cross compiler toolchain | 05:16 |
| xobs | I could see system python being screwed up. On Ubuntu, at least, the dependency order goes site-packages, $PYTHONPATH, /usr/share/python | 05:17 |
| xobs | In lxbuildenv I treat the cross-compiler as a thing that the user needs to install. If they've done riscv stuff for other platforms, then they won't need to install an extra compiler on their system. They already need to manually install Vivado. | 05:19 |
| mithro | The worse case is when they have previously installed litex or similar dependency | 05:19 |
| mithro | And you get weird things like "why are they getting this error despite installing the newer version" | 05:19 |
| xobs | Agreed. Python seems to prioritize globally-installed packages over manually-specified packages, which is why I first did a site.py that undoes that, and now I have lxbuildenv muck with sys.path to also undo that. | 05:20 |
| mithro | xobs: By using the conda Python we have a bit more control - for example we could try going to Python 3.6+ which for example shouldn't need the hash seed thingy. https://stackoverflow.com/questions/39980323/are-dictionaries-ordered-in-python-3-6 | 05:24 |
| tpb | Title: Are dictionaries ordered in Python 3.6+? - Stack Overflow (at stackoverflow.com) | 05:24 |
| mithro | xobs: the ideal setup would be "git clone xyz; cd xyz; make" then your done | 05:30 |
| xobs | mithro: That would be ideal. That's what lxbuildenv is, assuming you already do work with riscv and therefore have a compiler. | 05:32 |
| mithro | xobs: Thats a pretty big _if_ for most people :-P -- also at some point I expect to start seeing issues with people using older riscv compilers | 05:33 |
| xobs | I have concerns with conda. I don't know how well it's maintained, and I remember it was extremely difficult to figure out how to build packages for it, and in fact in looking at it now it seems as though you can't necessarily get the package source. | 05:33 |
| mithro | xobs: If you know of alternatives - happy to see if they are better, the advantage of conda is that it's used very heavily by the scientific Python crowd, so we kind of ride on those coat tails | 05:34 |
| mithro | xobs: Another alternative might be something docker based or similar | 05:34 |
| mithro | FYI - Our conda packages are here -> https://github.com/timvideos/conda-hdmi2usb-packages | 05:35 |
| tpb | Title: GitHub - timvideos/conda-hdmi2usb-packages: Conda build recipes for the toolchains needed by LiteX / MiSoC firmware (at github.com) | 05:35 |
| mithro | xobs: I was playing with changing our toolchain builds to use their more standard way of building compilers here -> https://github.com/mithro/conda-hdmi2usb-packages/tree/crosstool-ng | 05:36 |
| tpb | Title: GitHub - mithro/conda-hdmi2usb-packages at crosstool-ng (at github.com) | 05:36 |
| mithro | https://github.com/mithro/conda-hdmi2usb-packages/tree/crosstool-ng/crosstool-ng-feedstock/recipe | 05:36 |
| tpb | Title: conda-hdmi2usb-packages/crosstool-ng-feedstock/recipe at crosstool-ng · mithro/conda-hdmi2usb-packages · GitHub (at github.com) | 05:36 |
| xobs | mithro: Ah, okay. That makes more sense! I didn't see any version of "conda source", so I didn't realize that you were building your own conda packages. | 05:37 |
| mithro | xobs: At the moment conda seems to fall into the "conda is crappy, but it seems to be the least crappy of all the options" | 05:37 |
| mithro | xobs: They are slowly relearning a lot of things debian learnt years ago | 05:38 |
| xobs | Them and everyone else. I'm pretty sure the Node group has similar issues. | 05:39 |
| mithro | ha yeah | 05:39 |
| mithro | xobs: Atleast the conda people seem to be learning them /faster/ than debian did | 05:39 |
| xobs | They have a corpus of knowledge to build on. | 05:40 |
| mithro | Only took them 3 years to start thinking about reproducible builds... | 05:40 |
| mithro | xobs: Yeah - same way the FOSS community caught up to Google's bigtable and gfs in a much faster timeline then it took google to originally develop them | 05:40 |
| mithro | Random thing of the day, Spent the morning drawing this -> https://docs.google.com/drawings/d/1qREImoaUjWDSsnbimDu-Mig3_M9hTr-6q-zbPd2VPEM/edit | 05:41 |
| xobs | The lxbuildenv approach with a "deps/" directory has the nice property that one given project repo points to a known-good set of submodules, which makes those buidls reproducible. | 05:41 |
| tpb | Title: Verilog to Routing (VtR / VPR) - Carry Chain + Pack Patterns - Google Drawings (at docs.google.com) | 05:41 |
| mithro | xobs: Assuming you have the same version of risc-v and vivado installed? | 05:41 |
| xobs | Naturally. | 05:42 |
| mithro | https://crosstool-ng.github.io/docs/build/ <-- that actually looks like a pretty decent way to produce crosstools | 05:42 |
| tpb | Title: Building the Toolchain (at crosstool-ng.github.io) | 05:42 |
| mithro | (And what conda has started using...) | 05:43 |
| xobs | I know more about Python than I did back in January at LCA, but there are still some blind spots. | 05:44 |
| mithro | The _real_ solution is openembedded / yocto / bitbake thingy -- but that is biting a pretty huge thing | 05:45 |
| xobs | I thought the version of Python used for hdmi2usb was special, which isn't particularly the case. | 05:45 |
| xobs | I also didn't know which packages were important and which weren't. Or what dependencies there were. | 05:45 |
| mithro | xobs: That was somewhat partly the goal -- let people ignore all the Python details | 05:46 |
| xobs | I'm hoping that by avoiding the need to install something special, and then perform additional setup steps inside of that environment, it can make the build system seem less complex and more approachable by giving someone less to focus on. | 05:46 |
| mithro | xobs: Of course everyone seems to have a /different/ set of things they think are easy / hard to do | 05:48 |
| xobs | Also, you're right. Toolchains are hard. The crosstool-ng stuff looks useful. | 05:48 |
| xobs | mithro: It's true! But I was surprised to find that the whole migen/litex stuff really has no extra dependencies. | 05:48 |
| mithro | xobs: We have issues like you can't use gcc 4.7 or gcc 4.4 for lm32 but 5.1 and 4.9 where fine, etc | 05:50 |
| tumbleweed | defaultdict(<class 'list'>, {'s': ['cherubin'], 'l': ['ajfus']}) | 05:51 |
| tumbleweed | oopspaste | 05:51 |
| mithro | tumbleweed: oh - I thought you were making some type of comment about Python 3.6 having ordered dictionaries :-P | 05:51 |
| xobs | It was apropos | 05:52 |
| tumbleweed | it was a t-shirt assignment dump | 05:52 |
| mithro | xobs: BTW - when I sad "git clone x; cd x; make" that is kind of literally what I meant - but your makefile could just have a single target which runs the lxbuildenv.py script... | 05:53 |
| mithro | xobs: For example Verilog to Routing has a makefile which runs cmake with the default configuration and then builds it -- made me not care about the fact the project was using cmake | 05:54 |
| xobs | mithro: Your "all:" target would be "$(PYTHON) top.py" | 05:54 |
| mithro | xobs: Something like that | 05:54 |
| mithro | I'm going to walk home now, since it's almost 11pm | 05:55 |
| xobs | That's why the dependency checker has solutions for when dependencies are missing -- how to download Vivado and where to get the riscv toolchain from. | 05:55 |
| mithro | xobs: Yeah | 05:55 |
| mithro | xobs: btw - easy_install was pretty terrible in many ways - but it had a really nice "stub" feature, you committed the small stub which would then download and setup a proper easy_install environment when one wasn't available | 05:57 |
| mithro | xobs: Maybe it would be a good idea to start a Google doc with some type of "wishlist" ? | 05:57 |
| mithro | xobs: Would be a shame for you to accidently relearn a bunch of things we have already seen previously :-P | 05:58 |
| mithro | better to be making all new mistakes! ;-) | 05:59 |
| mithro | xobs: it would be worth checking out of apio and stock python is maybe a better fit - Its likely that tinyfpga has opinions there | 06:03 |
| mithro | Btw, actually walking home now - so typing is slow | 06:04 |
| *** Elwell has joined #timvideos | 06:04 | |
| *** Elwell has quit IRC | 06:05 | |
| xobs | I'm curious to know what other sorts of gremlins lurk in the notion of using system python. | 06:05 |
| *** Elwell has joined #timvideos | 06:06 | |
| mithro | xobs: People do all types of strange things to their system python | 06:31 |
| mithro | xobs: I dumped some random thoughts here -> https://docs.google.com/document/d/1V47Zc0eyjUUGBNsJuYORdb4TqH4C1WjzrX9odX0YdZE/edit | 06:40 |
| tpb | Title: Google Docs - create and edit documents online, for free. (at docs.google.com) | 06:41 |
| mithro | Anyway, sleepy time for me! | 06:46 |
| mithro | have a good night | 06:46 |
| xobs | I'll have a look at the doc. | 06:46 |
| xobs | Goodnight! | 06:46 |
| *** Kripton has quit IRC | 06:51 | |
| *** NeddySeagoon25 has joined #timvideos | 06:52 | |
| *** NeddySeagoon25 has quit IRC | 06:52 | |
| *** ktechmidas has joined #timvideos | 06:57 | |
| *** Kripton has joined #timvideos | 07:00 | |
| *** Remco4 has joined #timvideos | 07:36 | |
| *** Nineteen has joined #timvideos | 07:49 | |
| *** thaytan has quit IRC | 07:51 | |
| *** thaytan has joined #timvideos | 07:51 | |
| *** ChanServ sets mode: +v thaytan | 07:52 | |
| *** bluszcz7 has joined #timvideos | 08:26 | |
| *** Torgeir has joined #timvideos | 08:41 | |
| *** rohitksingh_work has quit IRC | 08:54 | |
| *** Guest11047 has joined #timvideos | 09:16 | |
| *** Guest11047 has quit IRC | 09:18 | |
| *** Razesdark23 has joined #timvideos | 10:00 | |
| *** tasse24 has joined #timvideos | 10:09 | |
| *** tasse24 has quit IRC | 10:14 | |
| *** Arokh21 has joined #timvideos | 10:24 | |
| *** __idiot__ has joined #timvideos | 10:38 | |
| *** deeprave has quit IRC | 11:09 | |
| *** _florent_ has quit IRC | 11:09 | |
| *** deeprave has joined #timvideos | 11:10 | |
| *** _florent_ has joined #timvideos | 11:12 | |
| *** poutine24 has joined #timvideos | 11:22 | |
| *** poutine24 has quit IRC | 11:22 | |
| *** tinyfpga has quit IRC | 11:57 | |
| *** tinyfpga has joined #timvideos | 11:57 | |
| *** Zapy11 has joined #timvideos | 12:08 | |
| *** Phex has joined #timvideos | 12:18 | |
| *** Phex has quit IRC | 12:24 | |
| *** infernix16 has joined #timvideos | 12:27 | |
| *** infernix16 has quit IRC | 12:32 | |
| *** fsamareanu22 has joined #timvideos | 12:32 | |
| *** fsamareanu22 has quit IRC | 12:33 | |
| *** Iota3 has joined #timvideos | 12:47 | |
| *** Iota3 has quit IRC | 12:52 | |
| *** tinyfpga has quit IRC | 12:56 | |
| *** tinyfpga has joined #timvideos | 12:57 | |
| *** rohitksingh has joined #timvideos | 13:03 | |
| *** clarjon116 has joined #timvideos | 13:09 | |
| *** evil has joined #timvideos | 14:02 | |
| *** evil has quit IRC | 14:02 | |
| *** mcintosh14 has joined #timvideos | 14:14 | |
| *** mcintosh14 has quit IRC | 14:15 | |
| *** _f15h has joined #timvideos | 14:19 | |
| *** geot has joined #timvideos | 14:19 | |
| *** _f15h has quit IRC | 14:20 | |
| *** jrslepak24 has joined #timvideos | 14:20 | |
| *** geot has joined #timvideos | 14:20 | |
| *** jrslepak24 has quit IRC | 14:26 | |
| *** jwhisnant0 has joined #timvideos | 14:27 | |
| *** jwhisnant0 has quit IRC | 14:30 | |
| *** Kronuz11 has joined #timvideos | 14:52 | |
| *** rohitksingh has quit IRC | 15:00 | |
| *** dwC-- has joined #timvideos | 15:42 | |
| *** rohitksingh has joined #timvideos | 15:42 | |
| *** rdococ4 has joined #timvideos | 15:54 | |
| *** rdococ4 has quit IRC | 15:56 | |
| *** GeekDude23 has joined #timvideos | 16:04 | |
| *** GeekDude23 has quit IRC | 16:05 | |
| *** Kripton has quit IRC | 16:13 | |
| *** Kripton has joined #timvideos | 16:16 | |
| *** geot has quit IRC | 16:22 | |
| *** rohitksingh has quit IRC | 16:23 | |
| *** connection has joined #timvideos | 16:39 | |
| *** connection has quit IRC | 16:40 | |
| *** rohitksingh has joined #timvideos | 16:43 | |
| *** mdoep25 has joined #timvideos | 17:04 | |
| *** geot has joined #timvideos | 17:06 | |
| *** NightMonkey21 has joined #timvideos | 17:20 | |
| *** NightMonkey21 has quit IRC | 17:26 | |
| *** geot has quit IRC | 17:34 | |
| *** samsagaz has joined #timvideos | 17:34 | |
| *** geot has joined #timvideos | 17:39 | |
| *** rohitksingh has quit IRC | 17:45 | |
| *** rohitksingh has joined #timvideos | 17:49 | |
| *** cr1901 has joined #timvideos | 17:58 | |
| *** CarlFK has quit IRC | 18:09 | |
| *** geot has quit IRC | 18:38 | |
| *** CarlFK has joined #timvideos | 18:46 | |
| *** ChanServ sets mode: +v CarlFK | 18:46 | |
| *** waldo323 has quit IRC | 20:34 | |
| *** waldo323 has joined #timvideos | 21:16 | |
| *** rohitksingh has quit IRC | 22:18 | |
| mithro | xobs: Added a few more things to that doc -- bunch of things I have seen go wrong with System Python and more of a proper wish list... | 23:58 |
Generated by irclog2html.py 2.13.1 by Marius Gedminas - find it at mg.pov.lt!