10 岁, 到了打拳皇和街霸的年纪了, 中国的下一个小孩或者丁春秋没准就是你儿子😊

😎看看还行, 买是不会买的

服务器放国外不会被审查, 但是用的人多了会被墙

另外一个主要问题是没人会用

没啥时间.... 结婚生子之后就更没有时间了

你都远程了为啥还整个 mac studio ？

唱国歌或者国际歌, 国际歌效果更好

😅只有京东 plus 会员, 给泥潭拖后腿了

@xmumiffy #12 好家伙, 双十国庆是吧, 你这是想害死墨索李妮吗😅

尤主动索敌?

配置好~/.ssh/config 之后

ssh remote 登陆

sshfs remote:/remote/path /local/path 把远程文件系统挂载到本地

openssh 和 sshfs 不行吗?

@yazinnnn0 #72 看起来 quarkus 给 20m 就可以跑个基本的 crud 程序了

话说在资源敏感的情况下, 不是该上 rust 吗?

怼上数据库(postgres)后简单模拟了一下 quarkus 在有限内存的情况

implementation("io.quarkus:quarkus-hibernate-reactive")
implementation("io.quarkus:quarkus-resteasy-reactive-jackson")
implementation("io.quarkus:quarkus-hibernate-reactive-panache-kotlin")
implementation("io.quarkus:quarkus-kotlin")
implementation("io.quarkus:quarkus-reactive-pg-client")


@Entity
class Todo : PanacheEntity() {
companion object : PanacheCompanion<Todo>

var title: String? = null
var description: String? = null
var completed: Boolean? = null

@Column(name = "due_date")
var dueDate: LocalDate? = null

@Column(name = "created_at", updatable = false)
var createdAt: LocalDate? = null

@Column(name = "updated_at")
var updatedAt: LocalDate? = null
}

@Path("todo")
class TodoApi {
@GET
fun getAll() = Todo.listAll()

@POST
fun save(todo: Todo): Uni<Todo> {
if (todo.id == null) {
return todo.persistAndFlush<Todo>()
}
throw WebApplicationException("id shouldn't exist", 499)
}

@GET
@Path("{id}")
fun getOne(@RestPath id: Long) = Todo.findById(id)
}


version: '3.8'

services:
postgres:
deploy:
resources:
limits:
memory: 100m
cpus: "0.1"
image: postgres:latest
environment:
POSTGRES_DB: yazinnnn
POSTGRES_USER: yazinnnn
POSTGRES_PASSWORD: yazinnnn

sample:
image: yazi/sample:1.0
deploy:
resources:
limits:
memory: 20m
cpus: "0.1"
ports:
- "80:8080"
depends_on:
- postgres
environment:
"QUARKUS_DATASOURCE_REACTIVE_URL": vertx-reactive:postgresql://postgres/yazinnnn


➜ postgres wrk -t 12 -c 100 -d 10s http://localhost/todo 18:09:49
Running 10s test @ http://localhost/todo
12 threads and 100 connections
Thread Stats Avg Stdev Max +/- Stdev
Latency 536.18ms 143.86ms 999.36ms 75.14%
Req/Sec 16.39 10.27 80.00 71.69%
1750 requests in 10.09s, 3.31MB read
Requests/sec: 173.39
Transfer/sec: 335.78KB


➜ postgres wrk -t 12 -c 100 -d 10s http://localhost/todo/1 18:10:03
Running 10s test @ http://localhost/todo/1
12 threads and 100 connections
Thread Stats Avg Stdev Max +/- Stdev
Latency 256.09ms 78.20ms 499.31ms 84.83%
Req/Sec 31.49 17.98 191.00 60.73%
3686 requests in 10.08s, 705.52KB read
Requests/sec: 365.55
Transfer/sec: 69.97KB

要不试试 quarkus?

docker run --cpus=0.1 --memory=20m --name quarkus -p 80:8080 yazi/sample:1.0
__ ____ __ _____ ___ __ ____ ______
--/ __ \/ / / / _ | / _ \/ //_/ / / / __/
-/ /_/ / /_/ / __ |/ , _/ ,< / /_/ /\ \
--\___\_\____/_/ |_/_/|_/_/|_|\____/___/
2024-01-04 08:46:58,037 INFO [io.quarkus] (main) sample 1.0 native (powered by Quarkus 3.6.4) started in 0.284s. Listening on: http://0.0.0.0:8080
2024-01-04 08:46:58,037 INFO [io.quarkus] (main) Profile prod activated.
2024-01-04 08:46:58,037 INFO [io.quarkus] (main) Installed features: [cdi, hibernate-orm, hibernate-reactive, kotlin, reactive-pg-client, resteasy-reactive, resteasy-reactive-jackson, smallrye-context-propagation, vertx]


CONTAINER ID NAME CPU % MEM USAGE / LIMIT MEM % NET I/O BLOCK I/O PIDS
185ccb037e19 quarkus 0.00% 12.16MiB / 20MiB 60.80% 993kB / 1.55MB 0B / 0B 42

yazinnnn0@cs-679368844937-default ~ [SIGINT]> wrk -t4 -d1s -c100 http://localhost:80/hello
Running 1s test @ http://localhost:80/hello
4 threads and 100 connections
Thread Stats Avg Stdev Max +/- Stdev
Latency 127.19ms 114.59ms 430.17ms 55.63%
Req/Sec 234.82 160.64 0.87k 78.57%
765 requests in 1.01s, 79.94KB read
Requests/sec: 755.88
Transfer/sec: 78.98KB


简单试了下 demo, 容器可以给 20M 内存

如果是自由软件卖身成了专有软件, 那活该被酸

不能平替

优势是并发量大, 消耗资源少, 功能强大

劣势是编程模型复杂, 复杂点的业务你要写成 monad 地狱, 虽然并发量大,但是一般业务瓶颈在数据库, 利用不到 reactive 的最大优势

写着玩可以随便试, 用 kotlin 协程可以稍微拯救一下 monad 地狱

loom 也不是银弹, loom 是增强 blocking 的方案, 不是增强 reactive 的方案

如果作为兴趣研究的话, 研究型语言推荐 scheme/racket/haskell, 工程型语言推荐 clojure 和 fsharp

都是最潮最 in 的函数式语言