<whitequark>
1. I don't think driving the entire sequence from cargo is a good idea
* cr1901_modern
nods
<whitequark>
2. requiring just also seems suboptimal
<whitequark>
if the firmware is in pure Rust, then it's enough to run cargo from nmigen-soc
<whitequark>
now, if the firmware has C components, that rapidly becomes a pain
<cr1901_modern>
The Makefile solution for misoc worked, but of course GNU Make isn't exactly wonderfully ergonomic to use from python (or Windows). We also don't want to create our own "command runner"/dep manager.
<whitequark>
how do you get into a situation where you have gcc (and it'll be only gcc for firmware if we're talking about C), but not gmake?
<cr1901_modern>
My own experience is that Windows builds of gcc only provide the compiler binaries and headers. A shell environment- including make and "sh" is not included. It's up to a user to get "make" and "sh" if they need it.
<cr1901_modern>
And sadly, there's multiple ways to do that. You could use Chocolately. MSYS2. The horribly out of date GnuWin32 project
<whitequark>
hmm
<cr1901_modern>
Some ppl understandably would "just" want a make binary, not the baggage of a full *nix userland
<whitequark>
well, let's see if we can get away providing just Rust
<cr1901_modern>
fair for me.
<cr1901_modern>
If I ever need C code, I'll take responsibility for it- i.e. I'm on my own.
lolsborn has quit [Quit: ZNC 1.6.3+deb1ubuntu0.2 - http://znc.in]
lolsborn has joined #m-labs
<mtrbot-ml>
[mattermost] <sb10q> someone should rewrite nix and fix the various issues it has including lack of windows support....
<cr1901_modern>
I don't see that happening
<_whitenotifier>
[nmigen] rroohhh opened issue #259: LegalizeValue can quickly exhaust available RAM - https://git.io/JeE3q
<_whitenotifier>
[nmigen] whitequark commented on issue #259: LegalizeValue can quickly exhaust available RAM - https://git.io/JeE3m
<_whitenotifier>
[nmigen] rroohhh commented on issue #259: LegalizeValue can quickly exhaust available RAM - https://git.io/JeE3O
<_whitenotifier>
[m-labs/nmigen] whitequark pushed 1 commit to master [+0/-0/±1] https://git.io/JeE33
<_whitenotifier>
[m-labs/nmigen] whitequark c1b2d2a - hdl.ir: silence UnusedElaboratable if __init__ raises an exception.
<_whitenotifier>
[nmigen] whitequark commented on issue #259: LegalizeValue can quickly exhaust available RAM - https://git.io/JeE3s
<_whitenotifier>
[nmigen] rroohhh commented on issue #259: LegalizeValue can quickly exhaust available RAM - https://git.io/JeE3G
<_whitenotifier>
[m-labs/nmigen] whitequark created branch master https://git.io/JeE3Z
Stormwind_mobile has quit [Ping timeout: 268 seconds]
Stormwind_mobile has joined #m-labs
Stormwind_mobile has quit [Ping timeout: 240 seconds]
Stormwind_mobile has joined #m-labs
Stormwind_mobile has quit [Ping timeout: 240 seconds]
Stormwind_mobile has joined #m-labs
<mtrbot-ml>
[mattermost] <sb10q> 10 years ago there were few people doing FPGA development in the open, and most of them were developing the crappy and unusable code that you can find in most repositories on opencores.org. it's only since the past few years that there's a significant amount of good open source design.
<whitequark>
they added in 2010, but yeah
<rjo>
whitequark: i haven't used it and don't know much about it. my feeling is that wishbone is more like a slightly antiquated legacy bus.
<whitequark>
okay.
<_whitenotifier>
[m-labs/nmigen] whitequark pushed 2 commits to master [+0/-0/±13] https://git.io/JeEZo
<mtrbot-ml>
[mattermost] <sb10q> @astro I've installed the packages at the system level, but if you still need nix-env I think -iA should solve the problem
Stormwind_mobile has quit [Ping timeout: 268 seconds]