The pipelight source code is splitted into 5-6 crates.
You can find READMEs in each modules(crates) for better crate by crate understanding.
Entry point
This crate is the smallest and has only one purpose. It is the entry point for building a single binary.
File convertion utilities
Mostly a wrapper around serde. This crate
contains the functions to read and parse pipelight.<toml/yaml/ts>
configuration files into Rust simple structs.
Beware! Those structs are not to be used as is and or converted to more comfortable structures.
Trivial convenience tooling
Contains great utility functions for some trivial things.
- git (fetch somoe info on git repositpories and generate git-hooks)
- logger (set the logger configuration on the fly)
- teleport (find a file recursively in the filesystem)
- dates (Date manipulation and serialization, Time computing)
- files (Query the filesystem)
- signal (functions to handle SIGTERM, Ctrl-C and others)
Wrapper around process execution
Functions to pipe processes output, retrieve pids, cwd and so on. Used to keep track of pipeline subprocesses.
The pipeline execution logic*
Read the Pipeline Struct extracted from tthe configuration file and set the execution workflow.
Command line structure definition
Define the command line sub-command, options and arguments. Makes extensive use of Clap.rs
A switch/case over commands line args
Determines which functions to run based on the command line arguments.
Generate config file templates
Contains the functions to generate templates when using commands like
pipelight init --template toml
When running a pipeline. This is what happens.
The entry point is pipelight/src/main.rs.
Then a switch/case function from swicth/src/case.rs to determine wich command
to execute.
The command is then detached from the standard I/O and sent to the background.
Pipelight is a single binary to execute pipelines, whach files, print logs...