jkarneges

9 ms·

61.
by jkarneges 6y ago
Interesting to see the Wake trait introduced. It appears to make it possible to create a Waker implementation without requiring unsafe code, at the cost of requiring the use of Arc. The way the trait is used by the standard library is amusi
62.
by jkarneges 6y ago
> why to have thousands of connections instead of a single one (or a bunch) that simply forwards websocket packets with some "connection id" For anyone wanting to do this sort of thing, check out Pushpin [1]. It's a proxy
63.
by jkarneges 6y ago
Sent you a mail.
64.
by jkarneges 6y ago
Fanout founder here! I appreciate the mention, although we don't actually support sending webhooks. Well, awhile back we did, but lately we only support pushing to client-initiated connections.
65.
by jkarneges 6y ago
Good to see this being addressed! We used to offer webhooks-as-a-service at Fanout, but it wasn't very popular and last year we discontinued the feature. There are several reasons I believe we weren't successful with it: 1) There
66.
by jkarneges 6y ago
> it's hard to make an actually fun game A zillion years ago I used to do hobby games as a solo dev, and it took me a long time to understand this. I was always interested in game engine elements (sprites, tilemaps, input, menus, au
67.
by jkarneges 6y ago
It does, but it doesn’t have to. :) https://fanout.io (Disclosure: founder)
68.
by jkarneges 6y ago
> this path was both realtime & still http/resource centric There's always long-polling. :) Not quite as efficient as Server Push, but the logical behavior is similar in that the server pushes an HTTP resource at the client
69.
by jkarneges 6y ago
Sure. I believe microservice boundaries are more about development agility rather than scalability. By limiting each microservice to a minimal API surface and a "2 pizza" team, everyone can iterate faster. And if a particular micr
70.
by jkarneges 6y ago
I'd agree there is an arguable difference between sharing a server vs sharing data within the server. Bottleneck issues aside, letting two microservices connect to the same Postgres cluster but access different "databases" (c
71.
by jkarneges 6y ago
Another potential misuse of Kafka I've been wondering about is how a single Kafka instance/cluster is often shared by multiple microservices. On one hand the ability to connect multiple microservices to a central message broker is
72.
by jkarneges 6y ago
I believe the rationale goes like this: Rust doesn't have a runtime, and the designers prefer to to keep it that way. But, there was also a desire for native-feeling async programming in Rust. In order to reconcile these seemingly oppo
73.
by jkarneges 6y ago
My understanding is a Durable Object is similar to a long running JS app in a respawnable/relocatable container (imagine a Kubernetes deployment of 1 pod). There is always exactly one instance, somewhere. And because the instance is lo
74.
by jkarneges 6y ago
> the bill is perfect, except for the 100M user threshold to start demand compatibility, which I think should be lowered to 10M I kinda like the high number, as it means the spirit is to prevent monopolies, which is one of the most compe
75.
by jkarneges 6y ago
> frankly 2M in ARR after two years is not exactly a home-run for YC It's been awhile since I looked at this deeply, but I thought the path of a good startup is to raise a seed with an 18 month runway, grow to 1m ARR, and then raise
76.
by jkarneges 6y ago
> could we maybe usher in an era of more calm, reasoned, cooperation between various parties who disagree on a few specific details, but largely agree [...] on the vast majority of issues [...] ? Looking at US politics, there has been so
77.
Show HN: The cost of Rust async/await(github.com)
15 points by jkarneges 6y ago|0 comments
78.
by jkarneges 6y ago
Wow. Just a moment ago I received a PR to slightly modify a readme, in a way that seemed unusual (no insertion of links or anything, but odd punctuation choices). I couldn't understand why someone would send it, and then saw this post.
79.
Vercel and WebSockets(blog.fanout.io)
2 points by jkarneges 6y ago|0 comments
80.
by jkarneges 6y ago
Hi HN! This library makes it super easy to push data to clients from a JS server app. Just a few lines of code to declare an endpoint for clients to connect to, and to send events to them. Notably, the library works in serverless/funct
81.
Show HN: JavaScript server-sent events library(github.com)
7 points by jkarneges 6y ago|1 comments
82.
by jkarneges 6y ago
Nice to see a native app and not something Electron-based. I wonder about native app viability these days though. Years ago I used native email apps like Outlook, KMail, and Thunderbird, but lately I just use the websites made by the servic
83.
by jkarneges 6y ago
Indeed Uber allows drivers to set fares in California. I'm not sure if that's true of Lyft though? https://www.sfchronicle.com/business/article/Uber-to-Lyft-Yo...
84.
by jkarneges 6y ago
> A lot of the lambda like services don't do great with doing realtime stuff Another way to handle this is by delegating the management of long-lived connections to a separate infrastructure layer. Then it's fine if the app run
85.
Rewriting Pushpin's connection manager in Rust(blog.fanout.io)
2 points by jkarneges 6y ago|0 comments
86.
by jkarneges 6y ago
That's fair, but the point is Uber tried to meet all of these. The ball is in the state's court now to test and enforce. My feeling is (2) is overreach and has caused unnecessary collateral damage to many non-rideshare businesses.
87.
by jkarneges 6y ago
Uber actually attempted to comply with AB5, by allowing California drivers to set their prices and giving them more freedom about which rides to accept. Critics may argue Uber is still non-compliant even with the changes, but that's pr
88.
by jkarneges 6y ago
> I hate "sprint" Me too. It sounds exhausting.
89.
Show HN: High performance HTTP/WebSocket connection manager, in Rust(github.com)
4 points by jkarneges 6y ago|0 comments
90.
by jkarneges 6y ago
> but typically that involves stateful connections [...] and scaling to a large number of viewers can get extremely costly (and operationally difficult) This is why we built Pushpin, to make it easier to handle stateful connections at sc

More ›