02:25 <mithro> Kind of a long shot, but does migen/misoc have some type of "road map" or other document which has the kind of "here is what we plan to do" type thing?

06:50 <sb0_> mithro, in a few words:

06:50 <sb0_> * rewrite the migen simulator in python to get rid of iverilog problems and improve the API (Python introduced nice generator features since I initially wrote the migen simulator)

06:51 <sb0_> * make migen very small, move any non-basic core to misoc

06:51 <sb0_> * stabilize migen api, improve docs/unittests

06:51 <sb0_> * release

06:51 <sb0_> misoc, I'm not totally sure yet

06:52 <sb0_> regarding the "python" simulator, here's what I have in mind: https://gist.github.com/sbourdeauducq/93bc62ee51c3b9cbfcc6

06:52 <sb0_> (this does not show the API improvements that I want to make, just how to simulate in python and how fast/slow it runs)

22:35 <whitequark> wow, microsoft tools are not just disgusting, they keep getting more disgusting with every release

22:57 <cr1901_modern> Looks like you're having a genuinely bad time with Win7...

23:27 <cr1901_modern> whitequark: You poor soul, you're building LLVM on Windows XD?

23:29 <cr1901_modern> Last time I tried it w/ 2012, cl.exe complained std::tie doesn't take more than 2 arguments. Wonder if they fixed that.