GitHub Actions と pytest を組み合わせると、コードをプッシュするたびに自動でテストが走る CI 環境を無料で構築できます。 本記事では、ゼロから始める workflow ファイルの書き方から、matrix build・pip キャッシュ・artifact 保存まで、実務で使える設定をまとめました。 1. GitHub Actions で Python テストを自動化するメリット CI/CD パイプラインを整備すると、以下の効果が得られます。 プッシュ・PR のたびにテストが自動実行され、バグの混入を早期に検知できる 複数の Python バージョンへの対応状況を並行確認できる テスト結果がリポジトリ上に保存され、チームで共有できる GitHub Actions はパブリックリポジトリであれば無料で利用できる 2. ディレクトリ構成と pytest 基本設定 本記事で想定するディレクトリ構成は以下のとおりです。 my_project/ ├── .github/ │ └── workflows/ │ └── test.yml ├── src/ │ └── calculator.py ├── tests/ │ ├── __init__.py │ └── test_calculator.py ├── pyproject.toml └── requirements.txt …

If you're starting with Rust, you've probably heard about Cargo. It’s the official package manager and build system for Rust, and it handles almost everything you need to get your Rust project up and running—from compiling code to managing dependencies. In this article, we’ll take a closer look at how dependency management works using `Cargo.toml`. ## What is `Cargo.toml`? When you create a new Rust project using the command: cargo new my_project Cargo sets up a new project directory with a file called `Cargo.toml`. This file is the heart of your project configuration. It contains metadata about your project—like its name, version, and authors—and, more importantly, a list of dependencies. Here’s a simplified example of what a `Cargo.toml` might look like: [package] name = "my_project" version = "0.1.0" edition = "2021" [dependencies] Right now, it's empty under `[dependencies]`, but this is where the magic of dependency management begins. ## Adding Dependencies Dependencies are other Rust packages (called _crates_) that your project uses. Most of these crates live on crates.io, Rust's central package registry. To use a crate, you simply add it under the `[dependencies]` section with the name and version. For example, if you want to use the popular HTTP client crate `reqwest`, you add: [dependencies] reqwest = "0.11" When you run `cargo build`, Cargo will: 1. Download the `reqwest` crate (if not already cached), 2. Download any other crates that `reqwest` depends on, 3. Compile everything, and 4. Link them into your project. ## Understanding Version Numbers Cargo uses Semantic Versioning, a convention for versioning that helps manage updates and compatibility. Here’s a quick breakdown: * `1.2.3` means: * `1` is the **major** version (breaking changes), * `2` is the **minor** version (new features, backward-compatible), * `3` is the **patch** version (bug fixes). In `Cargo.toml`, if you write: serde = "1.0" Cargo interprets this as `^1.0`, meaning it will accept any `1.x.y` version that’s compatible with `1.0.0` (but not `2.0.0`). You can pin a version more strictly if needed: serde = "=1.0.130" Or specify a range: serde = ">=1.0.100, <1.1.0" ## Adding Local or Git Dependencies Besides crates.io, you can also use: ### Local crates my_utils = { path = "../my_utils" } This is useful for code you’ve written yourself but want to share between projects. ### Git repositories my_crate = { git = "https://github.com/username/my_crate.git" } Cargo will fetch the crate directly from the Git repository. ## Features and Optional Dependencies Some crates support optional features. You can specify them like this: tokio = { version = "1", features = ["full"] } Or use optional dependencies and conditionally enable them: [dependencies] serde = { version = "1.0", optional = true } [features] default = [] use_serde = ["serde"] ## Final Thoughts `Cargo.toml` is a powerful and flexible way to manage dependencies in Rust. By simply editing this file and running `cargo build`, you can pull in any library you need, keep your code modular, and avoid reinventing the wheel. As your projects grow, you’ll find more advanced use cases for `Cargo.toml`—like workspace management, build scripts, and custom profiles—but for now, understanding dependencies is a great first step. If you're a software developer who enjoys exploring different technologies and techniques like this one, check out LiveAPI. It’s a super-convenient tool that lets you generate interactive API docs instantly. LiveAPI helps you discover, understand and use APIs in large tech infrastructures with ease! So, if you’re working with a codebase that lacks documentation, just use LiveAPI to generate it and save time! You can instantly try it out here! 🚀

When you want full control over your C++ project without fighting with CMake boilerplate or inconsistent dependency setups, meet Zyn — a powerful tool designed for modern C++ workflows with Git-level dependency tracking, semantic versioning, and a clean project structure. 🔧 Quick Start Create your project in seconds: `zyn new my_project cd my_project` Add dependencies directly from Git: `zyn install https://github.com/fmtlib/fmt.git@9.1.0` Run it fast, run it clean: `zyn run --release` 🧠 Why Zyn? Unlike legacy build systems, Zyn keeps things predictable, clean, and modular: * Git-based dependencies with version locking * Semantic versioning support * Clean project layout with predictable src/, include/, and .zyn/ folders * C++ standards support out of the box * Automatic compiler flag generation 🛠️ Commands ## Command What it does `zyn new <name>` Create a new project `zyn install <url>` Install dependency from Git `zyn add <path>` Add a local dependency `zyn run [--debug/--release]` Run app `zyn update` Update all dependencies `zyn clean` Wipe build artifacts 🗂 Project Structure Zyn generates a structure that makes sense: project/ ├── .zyn/ │ ├── deps/ # Git dependencies │ ├── build/ # Compiled outputs │ └── lock/ # Version locks ├── src/ # C++ source files ├── include/ # C++ headers └── zyn.toml # Central project config 🔩 Configuration Example ` [project] name = "my_app" version = "1.0.0" language = "cpp" standard = "c++20" compiler = "clang++" [directories] sources = "src" include = "include" build = ".zyn/build" [dependencies] fmt = { git = "https://github.com/fmtlib/fmt.git", tag = "9.1.0" } boost = { path = "/opt/boost" } [libraries] lib_dirs = ["/usr/local/lib"] libraries = ["pthread", "dl"] ` 🔁 Dependency Management Add a Git repo directly: `[dependencies] json = { git = "https://github.com/nlohmann/json.git", tag = "v3.11.2" }` Or use a local path: `[dependencies] mylib = { path = "../mylib" }` Zyn will lock versions, build automatically, and ensure reproducibility. ✅ Best Practices * Always commit zyn.toml and .zyn/lock/ * Use tagged releases instead of branches * Run zyn update regularly to keep deps fresh * Prefer minimal, focused dependencies — Zyn makes it easy 🧨 Conclusion Zyn is what happens when you combine modern package management with low-level power. If you're serious about C++, and want a real workflow, try Zyn. Forget clutter. Forget hacks. Just build. Zyn — When Control Begins.

Bash is a must-know tool for developers, but many only use basic commands. Here are 10 powerful Bash tricks to speed up your workflow and make you feel like a terminal ninja. ## 1. Use `!!` to Repeat the Last Command Typed a command but forgot `sudo`? Instead of retyping, just run: sudo !! This reruns the last command with `sudo`. ## 2. Use `!$` to Reuse the Last Argument Instead of retyping the last argument, use `!$`: mkdir my_project cd !$ This changes into the newly created directory. ## 3. Find Large Files Quickly Need to free up space? Find large files with: find . -type f -size +100M This lists all files larger than 100MB in the current directory. ## 4. Extract Any Compressed File with One Command No need to remember different extract commands. Just use: extract() { if [ -f "$1" ]; then case "$1" in *.tar.bz2) tar xjf "$1" ;; *.tar.gz) tar xzf "$1" ;; *.zip) unzip "$1" ;; *.rar) unrar x "$1" ;; *.7z) 7z x "$1" ;; *) echo "Unknown file format" ;; esac else echo "'$1' is not a valid file" fi } Save this in your `.bashrc` or `.bash_profile` and run `extract filename.zip`. ## 5. Monitor Real-Time File Changes To see live changes in a file, use: tail -f filename.log Great for monitoring logs or debugging. ## 6. Rename Multiple Files at Once Easily rename `.txt` files to `.md` using: for file in *.txt; do mv "$file" "${file%.txt}.md"; done This renames all `.txt` files to `.md`. ## 7. Kill a Process by Name Instead of finding and killing a process manually, run: pkill -f process_name This kills any process matching `process_name`. ## 8. Copy Files Over SSH Like a Pro Instead of `scp` every time, mount a remote folder using SSH: sshfs user@remote:/path/to/folder /local/mount Now, you can access remote files like they’re local! ## 9. Create a Quick and Simple HTTP Server Serve the current directory over HTTP with: python3 -m http.server 8000 Now, open `http://localhost:8000` in your browser. ## 10. Make Your `cd` Command Smarter Tired of typing long paths? Add this to `.bashrc`: shopt -s autocd Now, typing a directory name alone will `cd` into it. ### Final Thoughts These Bash tricks will speed up your workflow and help you work like a pro. Got your own favorite Bash trick? Share it in the comments! 🚀 Like, share, and support this blog if you found it useful!

Hi, We are facing some issue on our project running `cargo build` or `cargo test` can cause the following error error: failed to build archive at `C:\Users\lucianopa\Documents\my_project\target\x86_64-pc-windows-msvc\debug\deps\libmyproject.rlib`: failed to remove temporary directory: The process cannot access the file because it is being used by another process. (os error 32) at path "C:\Users\lucianopa\Documents\ my_project\target\x86_64-pc-windows-msvc\debug\deps\\.tmpB3EUxx.temp-archive" The issue is flaky but less likely to happen if we specify `--jobs 1` It does looks like some issue with cargo but my question would be what are those .temp-archive for and if we are doing something in our configuration that makes this error happen or more likely? Also if there would be any tips on how to debug this further? Thanks, Luciano Almeida

I’m new to Rust and trying to understand the practical differences between: * `cargo new my_project` * `cargo new --lib my_project`

Problem I'm having is to include module_a into module_b. Asking chatgpt it gave this solution my_project/ │── src/ │ │── main.rs (Entry point) │ │── module_a.rs (Contains functions and structs) │ │── module_b.rs (Depends on module_a) │── Cargo.toml // src/module_a.rs pub struct Data { pub value: i32, } pub fn show_data(data: &Data) { println!("Data value: {}", data.value); } // src/module_b.rs use crate::module_a::Data; // Correct way to import from module_a pub fn process_data() -> Data { Data { value: 42 } } Well, it doesn't work. Keep getting "Could not find `module_a` in crate root."

I have a workspace package ("virtual manifest" organization) containing two library crates. $ find . -name lib.rs -o -name Cargo.toml ./Cargo.toml ./my_project/Cargo.toml ./my_project/src/lib.rs ./test_support/Cargo.toml ./test_support/src/lib.rs The `test_support` crate is a private testing support library for the other crate; it won't be published. As such, I haven't bothered to give it complete metadata, and I have [lints.clippy] all = { level = "deny" } cargo = { level = "deny" } cargo-common-metadata = { level = "allow", priority = 1 } in its `Cargo.toml`. The `my_package` crate, on the other hand, has just [lints.clippy] all = { level = "deny" } cargo = { level = "deny" } because I do want to get dinged if I leave something out of that one. ... But this doesn't work. Clippy rejects the `test_support` crate for not having complete metadata, unless I add the `cargo-common-metadata` allow line to `my_package/Cargo.toml` as well. Why doesn't it work to have `cargo-common-metadata = { level = "allow", priority = 1 }` in just one of the two crates' `Cargo.toml`? Is this not the appropriate way to handle a crate with a private test-only dependency that won't be published, and if so, what _is_ the appropriate way to handle such a situation? complete (redacted) Cargo.toml (click for more details) complete (redacted) my_project/Cargo.toml (click for more details) complete (redacted) test_support/Cargo.toml (click for more details)

I wrote a script that creates a new QGis project and loads a vector layer into it: my_project = QgsProject() my_project_path = ./folder/Test.qgz my_project.addMapLayer(vlayer) my_project.write(my_project_path) Now I want to set the extent of the canvas…