07:55 <_whitenotifier-5> [amaranth] vk2seb commented on pull request #1578: [WIP] Implement RFC 41: `lib.fixed` - https://github.com/amaranth-lang/amaranth/pull/1578#issuecomment-2921529121

08:44 <_whitenotifier-5> [amaranth] whitequark commented on issue #1603: Using `sim.add_clock(Period())` (`Period` without any arguments) leads to an empty VCD file - https://github.com/amaranth-lang/amaranth/issues/1603#issuecomment-2921640947

09:07 <whitequark[cis]> zyp: so, i wanted to throw around some ideas around streams and FIFOs

09:08 <whitequark[cis]> in glasgow right now i have a StreamFIFO with this signature:

09:08 <whitequark[cis]> def __init__(self, *, shape, depth, w_domain="sync", r_domain="sync", buffered=True):

09:08 <whitequark[cis]> this converts the four options of sync/async buffered/unbuffered into one component

09:11 <whitequark[cis]> i'm not yet sure if it should behave this specific way or not. in particular, i think sync/async FIFOs should be different components, because specifying equal domains to StreamFIFO changes its implementation in a way applying DomainRenamer does not

09:12 <whitequark[cis]> and that in turn changes other properties such as latency, so that even if the StreamFIFO.elaborate method could find out whether the two domains are actually the same or not (functionality we wanted for a while), it should probably not switch the implementation based on that either

09:13 <whitequark[cis]> on the other hand, i'd like to have ratio conversion built-in, both "individual element to array of element" and "array of elements to bigger array of elements"

11:01 <zyp[m]> I'm inclined to agree, the sync/async difference should probably be more explicit than just whether the clock domains happens to be the same

11:07 <zyp[m]> and yeah, ratio conversion ties right into RFC 56

15:58 <_whitenotifier-5> [rfcs] wanda-phi opened pull request #78: Add an RFC for `InternalPort`. - https://github.com/amaranth-lang/rfcs/pull/78

15:59 <_whitenotifier-5> [rfcs] wanda-phi edited pull request #78: Add an RFC for `InternalPort`. - https://github.com/amaranth-lang/rfcs/pull/78

15:59 <_whitenotifier-5> [rfcs] wanda-phi edited pull request #78: Add an RFC for `InternalPort`. - https://github.com/amaranth-lang/rfcs/pull/78

16:03 <_whitenotifier-5> [rfcs] whitequark reviewed pull request #78 commit - https://github.com/amaranth-lang/rfcs/pull/78#discussion_r2116190019

16:43 <_whitenotifier-5> [rfcs] wanda-phi opened pull request #79: Add an RFC for `PortLike.with_direction`. - https://github.com/amaranth-lang/rfcs/pull/79

16:44 <_whitenotifier-5> [rfcs] wanda-phi edited pull request #79: Add an RFC for `PortLike.with_direction`. - https://github.com/amaranth-lang/rfcs/pull/79

16:44 <_whitenotifier-5> [rfcs] wanda-phi edited pull request #79: Add an RFC for `PortLike.with_direction`. - https://github.com/amaranth-lang/rfcs/pull/79

21:36 <_whitenotifier-5> [amaranth] goekce commented on issue #1603: Using `sim.add_clock(Period())` (`Period` without any arguments) leads to an empty VCD file - https://github.com/amaranth-lang/amaranth/issues/1603#issuecomment-2923563417

21:54 <_whitenotifier-5> [amaranth] whitequark commented on issue #1603: Using `sim.add_clock(Period())` (`Period` without any arguments) leads to an empty VCD file - https://github.com/amaranth-lang/amaranth/issues/1603#issuecomment-2923586957

23:20 <whitequark[cis]> adamgreig: how's post-conference recovery?