vipqualitypost0[ has quit [Quit: Idle timeout reached: 172800s]
jjsuperpower_ has quit [Remote host closed the connection]
jjsuperpower_ has joined #amaranth-lang
coltonp has quit [Quit: Client closed]
<Wanda[cis]>
regarding RFC 36: is there any reason why sim.critical() is an async context manager instead of a plain context manager? it seems a plain context manager (with sim.critical(): instead of async with sim.critical():) would make more sense
<Wanda[cis]>
also didn't we want sample() on edge() triggers too?
<Wanda[cis]>
as it stands, the DDR buffer example is unimplementable?
<whitequark[cis]>
oh, that's probably just an omission
<Wanda[cis]>
it needs syntax and it needs a way to return things
<Wanda[cis]>
neither of those is obvious
<Wanda[cis]>
.edge(signal, val, sample=(a, b, c)) returning either None (if not hit) or (val_a, val_b_, val_c)?
<whitequark[cis]>
ugh
<whitequark[cis]>
i think i'm too tired to think about right now
zyp[m] has joined #amaranth-lang
<zyp[m]>
<Wanda[cis]> "regarding RFC 36: is there any..." <- I think the idea of making it async was that that'd let it be initially implemented as a shim on top of the old API and fit in with async set/get, and since that's out, it does indeed make more sense for it to be a plain context manager, yes
<Wanda[cis]>
right
<Wanda[cis]>
I'm basically preparing to do a major refactor of the sim stuff right now, so yeah
<Wanda[cis]>
hm, having a Period would be convenient indeed
<Wanda[cis]>
<del>said a catgirl</del>
<Wanda[cis]>
I do like the idea of making it not sim-only
<Wanda[cis]>
ie. also usable for clock constraints and (when we have them) input/output delay constraints, or stuff like IODelay wrapper parameters
<whitequark[cis]>
yes
<whitequark[cis]>
amaranth.hdl.Period is ... fine I guess?
<whitequark[cis]>
it's sort of weird. it could be re-exported from sim maybe. but also I don'
<whitequark[cis]>
* it's sort of weird. it could be re-exported from sim maybe. but also I don't like having three places where we have Period with no completely obvious indication of which one is the main one
<whitequark[cis]>
the prelude and amaranth.hdl already sort of have that and it's already not ideal,
<whitequark[cis]>
so maybe making it worse isn't the best ide
<whitequark[cis]>
s/ide/idea/
<galibert[m]>
As long as there isn't an amaranth.hdl.Endometriosis
<Wanda[cis]>
zyp: thoughts?
<zyp[m]>
sounds good to me
<zyp[m]>
as long as the fact that everything needs to be representable as whole femtoseconds isn't causing any issues, but I don't expect that to be any worse than float rounding
<Wanda[cis]>
it's strictly preferable to float rounding
<Wanda[cis]>
looks at the RFC 36 classes not quite sure how much of this will end up being backend-specific
<Wanda[cis]>
I can probably get away with stuffing a bunch of that to core?
<whitequark[cis]>
it's probably not super important to do a split between core and backend yet
<Wanda[cis]>
mhm
<Wanda[cis]>
I'll just do something that feels reasonably right then
<whitequark[cis]>
sgtm
Raito_Bezarius has joined #amaranth-lang
Raito_Bezarius has quit [Read error: Connection reset by peer]
Raito_Bezarius has joined #amaranth-lang
<_whitenotifier-5>
[amaranth] github-merge-queue[bot] created branch gh-readonly-queue/main/pr-1343-d2b2c9b06a3d8072a0705fd8bd38393294c8c4da - https://github.com/amaranth-lang/amaranth
<_whitenotifier-6>
[amaranth-lang/amaranth-lang.github.io] github-merge-queue[bot] 8322c78 - Deploying to main from @ amaranth-lang/amaranth@9201cc3179d53c9fcfb6443e571533d64dbdd417 🚀
adamse has joined #amaranth-lang
_whitenotifier-6 has quit [Remote host closed the connection]
_whitenotifier-5 has quit [Remote host closed the connection]