00:08
lf_ has quit [Ping timeout: 264 seconds]
00:08
lf has joined #amaranth-lang
00:33
notgull has quit [Ping timeout: 272 seconds]
00:37
notgull has joined #amaranth-lang
01:25
<
Wanda[cis] >
hmmm actually, should we have some provisions for ShapeCastable-aware muxing
01:25
<
whitequark[cis] >
that sounds tricky except in the case where both operands have exactly the same shape
01:25
<
Wanda[cis] >
it feels suboptimal that you can Mux two Views and not get back a View
01:26
<
Wanda[cis] >
but also having that as a special case feels hacky too
01:26
<
whitequark[cis] >
yep
01:31
<
Wanda[cis] >
we could require same ShapeCastable for all inputs if any of them is ShapeCastable; it's a new primitive after all
01:31
<
whitequark[cis] >
we could!
01:31
<
Wanda[cis] >
Mux and ArrayProxy will just call Value.cast before lowering to Choice
01:32
<
Wanda[cis] >
(slightly suboptimal, but eh)
01:38
<
Wanda[cis] >
I added it to unresolved questions
02:40
Degi_ has joined #amaranth-lang
02:40
Degi has quit [Ping timeout: 260 seconds]
02:40
Degi_ is now known as Degi
04:28
jjsuperpower has quit [Remote host closed the connection]
05:24
Guest14 has joined #amaranth-lang
05:29
Guest14 has quit [Quit: Client closed]
06:53
notgull has quit [Ping timeout: 256 seconds]
06:56
notgull has joined #amaranth-lang
07:28
eldiem has quit [Read error: Connection reset by peer]
07:29
eldiem has joined #amaranth-lang
07:45
notgull has quit [Ping timeout: 272 seconds]
10:17
refx_investor[m] has joined #amaranth-lang
10:20
refx_investor[m] has left #amaranth-lang [#amaranth-lang]
11:05
<
jfng[m] >
@whitequark:matrix.org what is the procedure for amendments to an RFC ? a PR then a vote on a meeting ?
11:06
<
jfng[m] >
i'd like to amend RFC #16 to introduce an alignment parameter to `csr.Builder`
11:41
<
zyp[m] >
#29 and #47 were amendments using that procedure, yes
11:42
<
zyp[m] >
but adding another parameter sounds like an addition that could very well be a normal RFC
12:31
<
_whitenotifier-7 >
[amaranth-lang/amaranth] wanda-phi 161b014 - hdl._ast, hdl._ir: Deduplicate shape unification logic. NFC
12:32
<
_whitenotifier-5 >
[amaranth-lang/amaranth-lang.github.io] github-merge-queue[bot] 2092cf8 - Deploying to main from @ amaranth-lang/amaranth@161b01450ede96cf3d7b6999732f057465c2b7bb 🚀
12:59
notgull has joined #amaranth-lang
17:43
<
Wanda[cis] >
should we have some way to convert a Format object to string at simulation time? I think yield Format(...) would work for the old sim interface, perhaps await sim.format(Format(...)) for the new interface?
17:49
<
whitequark[cis] >
I don't really have any arguments for or against that
18:07
<
zyp[m] >
if yield Format(…) works, yield sim.get(Format(…)) would too
18:07
<
zyp[m] >
* if yield Format(…) works, await sim.get(Format(…)) would too
18:07
<
whitequark[cis] >
how so? Format is not a `Value
18:07
<
whitequark[cis] >
* how so? Format is not a Value
18:08
<
zyp[m] >
right, I'm not thinking
18:29
<
Wanda[cis] >
(not yet, at least)
20:10
skipwich has quit [Ping timeout: 255 seconds]