| 18 May 2022 |
 creikey | their solution is to install mysys2 | 01:38:29 |
 spoon | ah those are the ones i used, heard good things about naylib too | 01:38:30 |
| spoon | i dont remember having to install that, but i might've gotten it through scoop for something else | 01:39:26 |
| spoon | for cross compilation or native? | 01:43:11 |
|  Echo123 joined the room. | 04:03:53 |
|  mazariego joined the room. | 04:21:26 |
|  judithbnks left the room. | 10:40:15 |
|  billykin joined the room. | 12:28:26 |
| billykin left the room. | 16:20:10 |
|  Aurélien joined the room. | 16:55:49 |
|  davis_ left the room. | 17:07:33 |
 Recruit_main707 | Raylib_forever works fine in windows | 19:06:30 |
 planetis | can anyone try naylib on windows I would like to have some feedback | 19:12:59 |
| planetis | I did my part, fixed all the reported issues | 19:13:04 |
 rlipsc#3365 | Hey thanks! AFAIK it's just me using it on a website idea that's in the very early stages, and it's been a good way to feed back more realistic usage to the networking component design. I do have an end product I'd like to build with it, though. | 20:33:40 |
| rlipsc#3365 | The main goal though isn't so much making a web server with ECS, but making reusable code for TCP + HTTP + routing + JSON RPC etc that can be combined with other things without having to design for it 🙂 | 20:34:26 |
| rlipsc#3365 | It's not quite as crazy as it sounds! In some ways networking actually translates quite well to DoD & ECS: socket + buffers -> components. Then again... I did decide to do it with events that change components on the caller entity to call more events, then decide said events need to catch 'unsafe' mutation chains at compile time, then decide I also need to allow construct which runs all of the add events... within events! You may have a point... | 20:35:28 |
| rlipsc#3365 | As a proof of principle serving a website with JavaScript and images and the usual stuff it works, but I've not really taken it beyond that yet. | 20:36:07 |
| rlipsc#3365 | The basic transport design uses one entity per endpoint. So there's a TcpConnection component with a socket, and adding TcpRecv/TcpSend buffer comonents starts listening/sending immediately. Then you poll a system (much like async) to add TcpRecvComplete/TcpSendComplete components and signal completion, which you can then use to trigger further stuff. | 20:36:52 |
| rlipsc#3365 | To get HTTP, you just add ProcessHttp to your listening entities, and incoming data gets parsed to a HttpRequest. So now you've got an entity with connection, buffers, processed HTTP, and maybe your own components. You can then add a HttpResponse to send a page, which in turn adds/updates TcpSend and starts the send 🙂 | 20:37:59 |
| Recruit_main707 | if i ever start another raylib project ill try to use naylib first | 20:38:01 |
| rlipsc#3365 | The basic advantage you get from all this is lots of control at scale, as systems let you control processing rate and interval so you can balance/throttle things, and you can prioritise/chain stuff as immediate in-order responses with events. Plus cancellations are easy and closing a connection is just deleting an entity. | 20:39:17 |
 Bubblie | I should go back to to adding vulkan to boxxy | 20:40:31 |
 ajusa#3756 | No threading support though yet right? | 20:43:35 |
| ajusa#3756 | (comparing this to jester in my mind) | 20:43:46 |
| rlipsc#3365 | There is this https://github.com/rlipsc/polymorph/blob/c70b79f12ba563b790cd58ea77879ba47a943257/src/polymorph/sharedtypes.nim#L89 but all that really does is read the number of cores for you and some basic work splitting | 20:45:37 |
| rlipsc#3365 | full automatic threading is TODO, but will basically be static analysis on which systems interact and splitting on that | 20:46:13 |
| rlipsc#3365 | I wouldn't compare this to Jester though, it's more like a data oriented toolbox 🙂 | 20:46:40 |
| rlipsc#3365 | * There is this https://github.com/rlipsc/polymorph/blob/c70b79f12ba563b790cd58ea77879ba47a943257/src/polymorph/sharedtypes.nim#L89 but all that really does is read the number of cores for you and work splitting itself is up to you | 20:49:50 |
|  Arbix0501 joined the room. | 21:39:25 |
