09:05 <ktemkin> hmm -- it might be nice to have a clean idiom for passing around clock attributes (like the frequency of a domain's clock signal); that'd allow for nice, clean declarations like `u = UARTReceiver(baud=115200)`

09:08 <ktemkin> though I imagine that'd be tricky to implement without also creating a confusing mechanic, given the way clock domains are created

21:30 <awygle> ugh still can't get diamond backend to work

21:30 <awygle> probably shoulda just set up a red hat/fedora VM

21:39 <emily> what are you trying to run it on?

21:43 <cr1901_modern> https://ycnrg.org/lattice-diamond-on-ubuntu-16-04/ This worked for me for Ubuntu. If you are on windoze, you will need to set up a script to add diamond to your path

21:43 <cr1901_modern> because lattice conveniently doesn't provide one for you on Windoze

21:45 <awygle> ubuntu

21:45 <awygle> i got it _installed_

21:45 <awygle> but all the paths are wrong

21:45 <awygle> for dynamic libraries

21:45 <awygle> so i fixed that, and now it's failing to read the "execute" shell script properly

21:46 <awygle> ubuntu 18.04 specifically

21:46 <cr1901_modern> execute shell script?

21:46 <cr1901_modern> I've only ever run nmigen w/ Diamond on Windows

21:46 <awygle> /usr/local/diamond/3.11-x64/pnbase/bin/execute or something liket hat

21:47 <cr1901_modern> There's no pnbase directory on 3.10 on my machine

21:47 <awygle> sorry, "synpbase"

21:48 <cr1901_modern> No execute under /usr/local/diamond/3.10_x64/synpbase/bin/ on my machine

21:50 <cr1901_modern> awygle: Try this

21:50 <awygle> it's probably a change in 3.11 then, maybe i should try 3.10

21:51 <cr1901_modern> set bindir env variable to /usr/local/diamond/3.11_x64/bin/lin64

21:51 <cr1901_modern> source /usr/local/diamond/3.11_x64/bin/lin64/diamond_env

21:51 <cr1901_modern> err export* bindir

21:52 <cr1901_modern> This is what I had to do to get pnmainc working

21:53 <awygle> yosys is picking up the wrong libc++ now i think

21:54 <cr1901_modern> Okay, that's a problem then :/

21:55 <cr1901_modern> These are my notes for dealing w/ diamond paths. I was trying to add autodetection to omigen, but eventually they dropped all the autodetection stuff: http://ix.io/1rj5

21:57 <cr1901_modern> diamond_env sets LD_LIBRARY_PATH, and yosys will need to run to convert the RTLIL to Verilog

21:57 <cr1901_modern> Prob worth opening an issue for this

21:59 <cr1901_modern> awygle: Could you generate the build inputs but not run diamond, and then paste the shell script for me?

22:01 <awygle> i can try, gimme a sec

22:07 <cr1901_modern> awygle: Actually I found the lines I want

22:08 <cr1901_modern> https://github.com/nmigen/nmigen/blob/master/nmigen/vendor/lattice_ecp5.py#L191-L194 I guess my question is: why is yosys choking due to libc++ on your system, but not necessarily others?

22:11 <awygle> i probably fucked everything somehow

22:11 <awygle> https://gist.github.com/awygle/2c0df14aea8c005e1dd43fd695a8df7b here are the three errors that i get

22:16 <cr1901_modern> awygle: FWIW, I now have the same error after trying it :P

22:18 <awygle> Well that's encouraging, sort of

22:27 <cr1901_modern> I've never run yosys when diamond_env was active, that's why I never saw this issue

22:29 <whitequark> awygle: ugh, so, you shouldn't set bindir while you're running nmigen

22:29 <whitequark> one moment

22:29 <whitequark> try running nmigen with `export NMIGEN_ENV_Diamond=/usr/local/diamond/3.11_x64/bin/lin64/diamond_env`

22:42 <awygle> whitequark: so, that prevents the yosys explosion, but there's still the parse problem with the execute shell script

22:43 <whitequark> which parse problem?

22:43 <whitequark> can you show the log?

22:44 <awygle> which log?

22:45 <awygle> https://gist.github.com/awygle/2c0df14aea8c005e1dd43fd695a8df7b this is what ig et on the terminal

22:45 <whitequark> hm

22:46 <whitequark> okay, let me grab diamind

22:46 <whitequark> *diamond

22:47 <_whitenotifier-3> [nmigen] rroohhh opened issue #320: Unclear error message on direction mismatch between Instance in submodule and toplevel usage - https://git.io/Jv8fh

22:48 <_whitenotifier-3> [nmigen] whitequark commented on issue #320: Unclear error message on direction mismatch between Instance in submodule and toplevel usage - https://git.io/Jv8fj

23:03 <awygle> the o->n conversion of Signal(max=N) should be Signal(range(N)), right?

23:04 <whitequark> yes

23:04 <whitequark> (it says that in the deprecation warning, no?)

23:05 <awygle> i just got an explosion, not a deprecation warning. but i'm messing around in nmigen.compat, so that's probably why

23:05 <awygle> (by explosion i mean "unexpected keyword argument" error)

23:06 <whitequark> awygle: ok, i have diamond working

23:06 <awygle> oh sweet

23:06 <whitequark> `fakeroot alien --scripts x.rpm` is just fine

23:06 <whitequark> not sure what broke for you

23:08 <awygle> on what linux distro?

23:08 <whitequark> debian

23:08 <awygle> i didn't use fakeroot i just used Actual Root, lemme try with fakeroot and see if that helps...

23:08 <whitequark> mostly packages from stable (everything diamond depends on should be from stable, anyway_

23:09 <whitequark> the diamond GUI runs

23:09 <whitequark> let me check nmigen now

23:17 <whitequark> awygle: i can reproduce your crash

23:21 <whitequark> awygle: sigh

23:27 <whitequark> awygle: for x in /usr/local/diamond/3.11_x64/synpbase/bin/*; do [ -f $x ] && sudo sed 's,#!/bin/sh,#!/bin/bash,' -i $x; done

23:32 <emily> hey come on at least do /usr/bin/env bash

23:33 <awygle> Jesus. Arright lol

23:35 <whitequark> emily: pointless

23:35 <whitequark> the rest of it is still full of absolute paths

23:36 <whitequark> awygle: applies to 3.10 too iirc

23:48 <ktemkin> oh gods; it's 2020 and we're still getting bitten by the "everyone links /bin/sh to bash except debian who links to dash by default" thing?

23:50 <ktemkin> life hack: put the blinders back on and `dpkg-reconfigurure dash` your way to `bash` being your `/bin/sh`~