no more _2 _3 in your flake.lock.
usage:
cat flake.lock | python nixfollows.pyno more _2 _3 in your flake.lock.
usage:
cat flake.lock | python nixfollows.py| # :schema https://docs.jj-vcs.dev/latest/config-schema.json | |
| # some contents are from https://gist.github.com/thoughtpolice/8f2fd36ae17cd11b8e7bd93a70e31ad6 | |
| [user] | |
| name = "lxl66566" | |
| email = "lxl66566@gmail.com" | |
| [ui] | |
| # 当只输入 `jj` 时,默认执行的命令设为 `jj log` (查看日志) | |
| default-command = "log" |
If you don't use idea, it's hard to maintain the same code style with your colleagues. This script solves this problem, it uses idea cli and the code style XML of your project to format your code.
| import http.server | |
| import os | |
| import re | |
| import socketserver | |
| PORT = 9000 | |
| UPLOAD_DIR = "./" | |
| # 一个简单的辅助函数,用于解析 multipart/form-data 的头部 |
| // ==UserScript== | |
| // @name 划词高亮页面内相同文字 | |
| // @namespace http://tampermonkey.net/ | |
| // @version 2.1 | |
| // @description 当在网页上选中一段文字后,自动高亮所有其他相同文字。点击或选择新内容后自动清除旧高亮。当选中的文本是在输入框或可编辑区域内的时候,不会触发高亮。 | |
| // @author lxl66566 (Gemini 2.5 pro) | |
| // @match *://*/* | |
| // @grant GM_addStyle | |
| // @license MIT | |
| // ==/UserScript== |
| [package] | |
| name = "aes_simd_rust" | |
| version = "0.1.0" | |
| edition = "2021" | |
| [dependencies] | |
| openssl = "*" | |
| rand = "0.9" # For generating test data | |
| [dev-dependencies] |
| //! The entire build.rs does one thing: applies schema.sql to | |
| //! target/sqlx_schema.db for sqlx to use during compilation. | |
| use std::process; // For panic | |
| use std::{env, fs, path::PathBuf}; | |
| // build.rs runs in a sync context, so we need a runtime to run sqlx async functions | |
| use fuck_backslash::FuckBackslash; | |
| use path_absolutize::Absolutize; | |
| // We need the execute trait from sqlx prelude |
| #![feature(test)] | |
| extern crate test; | |
| use std::collections::HashMap; | |
| use counter::Counter; | |
| use dashmap::DashMap; | |
| use rayon::prelude::*; | |
| use test::black_box; |
我一直不明白,第 k 大的数的正解为什么时间复杂度是 O(n)。看快排代码,二分递归代码一眼就是 O(nlog n),一看题解,全都说证明在算法导论,自己看书。 我看不懂书,因此想尝试做一个 benchmark,通过数据规模增长和 benchmark 用时来判断其时间复杂度。
注:find_kth_largest 函数本身没有 clone;我在 benchmark 之前就进行了数据 clone。
测试结果为(已排序,否则 cargo test 输出的排序为 10 100 1000 10000 50 500 5000):
| import { Bench } from "tinybench"; | |
| const bench = new Bench({ | |
| time: 1000, | |
| }); | |
| // This will not change the origin array | |
| function partition<T>( | |
| arr: readonly T[], | |
| predicate: (item: T) => boolean, |