05:37 <conventia[m]> Should it work with python 3.12? I just tried following the docs and building failed due to yosys not understanding read_ilang.

05:37 <whitequark[cis]1> which version are you using?

05:37 <conventia[m]> Which version of?

05:38 <whitequark[cis]1> of amaranth

05:38 <whitequark[cis]1> the read_ilang thing is a yosys deprecation that we have missed but it should be fixed in the latest release and main branch

05:40 <conventia[m]> pdm.lock says 0.5.0.dev152

05:41 <whitequark[cis]1> try fixing it to 0.5.4?

05:52 <conventia[m]> That worked or well, I got a sentinel.v file with verilog in it and no errors.

05:52 <conventia[m]> Thanks!

05:53 <conventia[m]> I guess I'm using a newer oss-cad-build.

05:56 <whitequark[cis]1> yes, that's fine; the bug is on us

05:56 <whitequark[cis]1> you can always use amaranth[builtin-yosys] to use a known good version of yosys

05:57 <whitequark[cis]1> s/use/install/, s//`/, s//`/

05:57 <conventia[m]> I just went with the first option in the quick start.

05:59 <conventia[m]> Though, some feedback for cr1901, the docs suggest that the Arty A7 35T, Cmod S7 and ICEBreaker v1.0 are supported, but the -p arg to attosoc only supports the icestick and hx8k.

05:59 <whitequark[cis]1> oh the sentinel quickstart

06:00 <conventia[m]> Yeah, sorry, I should've made that more explicit. 🙂

06:00 <conventia[m]> I was looking at sentinel last week and so when the docs were posted, I decided to try them.

06:32 <conventia[m]> I'm guessing that `ERROR: Failed to expand region (0, 0) |_> (13, 17) of 1303 ICESTORM_LCs` means that the ICEStick ran out of space?

06:32 <whitequark[cis]1> yes

06:32 <conventia[m]> I guess more input for the docs. 🙂

06:57 <conventia[m]> Got it working on the icebreaker

12:29 <cr1901> conventia[m]: Yes, sorry, the online docs are not tracking the "main" or even "next" branch right now ("do as I say, not as I do")

12:29 <cr1901> The online docs track the "docs" branch, but I will soon be fast-forwarding "next" to match "docs", and changing the readthedocs config to track "next"

12:30 <cr1901> You've unfortunately seen an intermediate state that ppl weren't supposed to see b/c I was just so happy the damn docs were done

12:30 <cr1901> That explains the "Arty A7 35T, Cmod S7 and ICEBreaker v1.0 are supported" lines

12:31 <cr1901> And the pdm.lock being bad on an initial checkout

12:32 <cr1901> As for the demo not fitting on icestick, that broke even for me after https://github.com/amaranth-lang/amaranth/commit/1d2b9c309e5a0e06487105679eec361472978c5f. I mention it in the CHANGELOG.md, but oops! I haven't pushed that yet :)

12:33 <cr1901> Nothing I can really do about it except look for small optimizations elsewhere to pare down the design.

12:36 <cr1901> Also, IME the demo never really consistently fit on OSS CAD suite toolchains, but it did consistently fit for me with a yosys I compiled myself until 1d2b9c3. I guess different compiler toolchains diverge just enough to lose some opts: https://github.com/cr1901/sentinel/issues/2

12:38 <cr1901> Anyways, glad to hear you got it working on the icebreaker :). Did you check out the "docs" branch for that or modify the source from a fresh checkout?

13:33 <conventia[m]> I just added my own icebreaker config, but I didn't match the pins very well, I suspect.

15:12 <conventia[m]> On the block diagram, the colors are pretty hard to distinguish. I think making the arrows thicker would help.

16:11 <cr1901> I'll keep that in mind for next release; I do think the block diagram needs improvement. In any case, both main and next branches have been updated to use amaranth 0.5.4, so please pull and try again. Also docs have been updated for icestick: https://sentinel-cpu.readthedocs.io/en/latest/usage/quickstart.html#a-full-example-soc-in-amaranth

16:13 <cr1901> Also thanks for the grammar fix :P

17:24 <cr1901> https://github.com/cr1901/sentinel/compare/v0.1.0-alpha.1..v0.1.0-beta 9,765 additions, 2,767 deletions (incl. docs). But it was worth it. Will announce later.

19:12 <conventia[m]> Are you using the latest pdm? It seems to have some warnings but I don't know if there is some reason to not upgrade?

19:13 <conventia[m]> And no problem, glad to help.