00:36 <Mahoramape> IRC.SUPERNETS.ORG #SUPERBOWL MEMORIAL DAY PARTY PLEASE NO NIGGERS

00:36 <Mahoramape> IRC.SUPERNETS.ORG #SUPERBOWL MEMORIAL DAY PARTY PLEASE NO NIGGERS

00:36 <Mahoramape> NO NOIGGAS NO CHNIGGERS BNIGGGASBERE DMONKEYS

00:36 <Mahoramape> NO NOIGGAS NO CHNIGGERS BNIGGGASBERE DMONKEYS

00:36 <Mahoramape> YLOLLOLLOLLOLLOLLOLLOL LOLLOL R YOURMOMISAFATNIGGERYOUR MOMISA WHOREILOLDLOLLOLLOLSUPERNETS ISDDOSPROOF NICEITRYFAGGOTSLOL LOLLOLLOLLOLSLOLLOL LOLLOLKEKKEK KEK

00:36 <Mahoramape> YLOLLOLLOLLOLLOLLOLLOL LOLLOL R YOURMOMISAFATNIGGERYOUR MOMISA WHOREILOLDLOLLOLLOLSUPERNETS ISDDOSPROOF NICEITRYFAGGOTSLOL LOLLOLLOLLOLSLOLLOL LOLLOLKEKKEK KEK

00:36 <Mahoramape> STUPIDOFATNIGGERSLOL LOLLOLLOLLOL LOLLOLLOLLOL LOLLOLLOL LOLLOLILOLLLOLLOL NONIGGERSINVITED TOTHEPARTYLOLOL LOLOLLOLOLLOLOL LOLTHISNETWORKHAS FURRYFAGSIN ITKEKKEKKEK KEKKEKKEKKEK

00:36 <Mahoramape> STUPIDOFATNIGGERSLOL LOLLOLLOLLOL LOLLOLLOLLOL LOLLOLLOL LOLLOLILOLLLOLLOL NONIGGERSINVITED TOTHEPARTYLOLOL LOLOLLOLOLLOLOL LOLTHISNETWORKHAS FURRYFAGSIN ITKEKKEKKEK KEKKEKKEKKEK

00:36 <Mahoramape> YOUCAN'TEVENDDOS LOLLOLLOLLOL FUCK ON THIS

00:36 <Mahoramape> YOUCAN'TEVENDDOS LOLLOLLOLLOL FUCK ON THIS

00:36 <Mahoramape> █▀▀▀▀▀█ █ ▄▄██▀▄ █▀▀▀▀▀█ ███████╗██╗ ██╗██████╗ ███████╗██████╗ ███╗ ██╗███████╗████████╗███████╗

00:36 <Mahoramape> █▀▀▀▀▀█ █ ▄▄██▀▄ █▀▀▀▀▀█ ███████╗██╗ ██╗██████╗ ███████╗██████╗ ███╗ ██╗███████╗████████╗███████╗

00:36 <Mahoramape> █ ███ █ ██▄ ▀▄ ▄▄ █ ███ █ ██╔════╝██║ ██║██╔══██╗██╔════╝██╔══██╗████╗ ██║██╔════╝╚══██╔══╝██╔════╝

00:36 <Mahoramape> █ ███ █ ██▄ ▀▄ ▄▄ █ ███ █ ██╔════╝██║ ██║██╔══██╗██╔════╝██╔══██╗████╗ ██║██╔════╝╚══██╔══╝██╔════╝

00:36 <Mahoramape> █ ▀▀▀ █ ▄██ █▀█▀▀ █ ▀▀▀ █ ███████╗██║ ██║██████╔╝█████╗ ██████╔╝██╔██╗ ██║█████╗ ██║ ███████╗

00:36 <Mahoramape> █ ▀▀▀ █ ▄██ █▀█▀▀ █ ▀▀▀ █ ███████╗██║ ██║██████╔╝█████╗ ██████╔╝██╔██╗ ██║█████╗ ██║ ███████╗

07:26 <emilbayes> Is there anywhere I can read about vectorisation/SIMD/AVX/SSE in pony? Is that left up to llvm to figure out or can I use compiler intrinsics somehow?

07:32 <vaninwagen> afaik this is left up to llvm, emilbayes, but i might be wrong.

07:49 <emilbayes> vaninwagen: Ah ok, I see that U128 uses FFI to what I guess is llvm intrinsics? `@"llvm.bswap.i128"[U128](this)`

07:52 <vaninwagen> emilbayes: yes, correct.

09:35 <_rck> if anyone was using it, I've updated the dash docset to use the new documentation theme \o/ https://github.com/ergl/pony-docset

09:36 <_rck> the new theme looks ncier, but it broke my old parsing scripts, heh

10:02 <vaninwagen> _rck, very nice

10:48 <_rck> vaninwagen, thanks :^)

10:48 <_rck> also, re https://irclog.whitequark.org/ponylang/2018-05-24#22172901

10:50 <_rck> I was implementing a poc in-memory kv store, and trying to see how to design the application to get parallel reads while allowing writes

10:50 <_rck> so I tried to partition the store to get, at least, parallel reads for separate partitions

11:04 <vaninwagen> _rck there is still the extreme of using 1 actor per key

11:07 <vaninwagen> _rck all other solutions would require to track access in order to properly balance the actor-keyspace mapping, otherwise skewed access might lead to new bottlenecks

11:10 <vaninwagen> but this is also a nice task to tackle

12:17 <_rck> that's true, for now most of my tests have been with an uniform key distribution, but of course things will look much different under different assumptions

12:19 <_rck> still, the approach you took with pony-kv might be a good one, if your load is mostly read-only, if performs quite nicely

14:16 <SeanTAllen> _rck can you add the dash docset note to the most recent last week in pony so everyone finds out next sunday?

14:17 <SeanTAllen> vaninwagen: i dont think there is anything extreme about an actor per key. its a reasonable design in my mind.

15:04 <mwmiller> I'm starting on my first "public" pony project and I have somne questions about best practices.

15:07 <mwmiller> I am going to build a redis package, largely by wrapping the official C library. Is it expected that `stable` will be able to download and compile the source of that library? Is there a way to make the library dependency programmatically visible in `stable`? If I don't do any of the above, what will happen for the end user of "what I really want to build" when the depenencies are resolved?

15:09 <mwmiller> If there is a canonical or recommended example of a pony-wrapped C library, I'm happy to read that.

16:08 <SeanTAllen> mwmiller: wrapping a c library and how to make that work with stable is very wild west. i think your best bet is to start doing it and then come back here and discuss improvements to stable to support what you want to do or even if stable is the right way to do it or if there's some other way.

16:08 <SeanTAllen> i'm going to open an RFC to kick regex and crpyto out of the standard library because it would force this issue and get us to having to start to address.

16:42 <mwmiller> SeanTAllen> i'm going to open an RFC to kick regex and crpyto out of the standard library because it would force this issue and get

16:42 <mwmiller> Sorry, not used to bitchx. Ok, great htnaks. I'll experiment if there aren't extant best practices.

16:59 <_rck> SeanTAllen: sure, I'll add it!

17:41 <strmpnk> I don’t know if it’d be an unpopular position but doesn’t kicking out crypto and regex just displace the problem to even less certain ground (sorry Pony-stable)? I consider these things pretty important.

17:42 <strmpnk> I like the batteries included setup of the standard library and would even like to see a few things added like libponyc bindings, pony-stable integration, and such.

17:44 <strmpnk> Splitting things out is likely to give us platform support blind spots and make dependency management a bigger pain point.

17:45 <strmpnk> The support story for these things is hard but that might point towards the need to vendor these libraries rather than depend on the host.

20:15 <SeanTAllen> That's exactly what I want to create

20:15 <SeanTAllen> I want to create that pain

20:15 <SeanTAllen> Better now than later