How Does It Work

Summary:
Technical overview of Bash Dev Env architecture and internals
Categories:
Tags:
Created on:
Last updated:

This chapter provides a technical overview of the architecture and internals of Bash Dev Env. It covers the main components, their interactions, and how they work together to provide a seamless installation and configuration experience.

1. Install command

this project consists of one simple command install that will

  • load the profile provided in argument
  • rotate logs
  • execute each install/configure/test scripts (in order including the dependencies of each profile selected scripts)
  • recapitulate duration, actions and important information

Some scripts are mandatory and are always executed, no matter what profile chosen:

  • _Test - do some mandatory checks before running the process
  • _Upgrade (first) - upgrade apt packages
  • MandatorySoftwares (second) - softwares mandatory for this whole system to work properly
  • _Clean (last) - clean apt, temp files, …

2. Engine

2.1. lib/loadAndCheckConfig.sh - loadAndCheckConfig

The most important function is loadAndCheckConfig from lib/loadAndCheckConfig.sh file.

This function is responsible for:

  • loading .env file

    • checks .env file exists, exit on error
    • loads .env file
    • checks validity of each .env variables

  • deducing some global variables

    • WSL only
      • WINDOWS_PROFILE_DIR
      • LOCAL_APP_DATA
      • WINDOW_PATH : converted windows PATH variable to linux compatible PATH

  • exporting the following variables from .env:

    • USERGROUP
    • USERNAME
    • GIT_USER_MAIL
    • GIT_USERNAME
    • SSH_LOGIN
    • POWERSHELL_BIN

  • deducing and export the following variables

    • USERID => computed from USERNAME provided by .env file
    • USERGROUP_ID => computed from USERGROUP provided by .env file
    • BASH_DEV_ENV_ROOT_DIR => <full path to this project directory>
    • IPCONFIG => heavy computation of the right command to use for ipconfig depending environment

  • following variables (with default values that could be overridden by .env file)

    • CONF_OVERRIDE_DIRS => defaults to (${BASH_DEV_ENV_ROOT_DIR}/conf_override)
    • LOGS_DIR => defaults to ${BASH_DEV_ENV_ROOT_DIR}/logs
    • PROJECTS_DIR => defaults to ${HOME}/projects
    • BACKUP_DIR => defaults to ${BASH_DEV_ENV_ROOT_DIR}/backup
    • INSTALL_SCRIPTS_ROOT_DIR => defaults to “${BASH_DEV_ENV_ROOT_DIR}”

  • setting sudoer without password temporarily

    • create a file in /etc/sudoers.d folder to avoid asking sudo password when it expires
    • create /etc/sudoers.d/bash-dev-env (check refactoring needed section)

  • installing and configuring the file ${HOME}/.bash-dev-env

    • the aim of this file is to provide common configuration variables for both ShellZsh and ShellBash
    • the main advantage of this file is that it can be automatically updated without risking to overwrite .bashrc file
    • the variable OVERWRITE_CONFIG_FILES is ignored
    • it provides the following environment variables, among others: PATH, COMPOSER_HOME, AWS_REGION, AWS_PROFILE, SSH_LOGIN, …
    • all these variables can then be reused by all the different script configurations like ssh aliases, Saml2Aws, …
    • this file loads this project .env file so variables like SHOW_FORTUNES and SHOW_MOTD can be taken into account by .bashrc, …

Note: this function has too much responsibilities and has to be refactored (cut in more specialized functions).

2.2. lib/loadAndCheckConfig.sh - loadProfile

This function :

  • checks that no files are missing for each script configuration (install, configure, test)
  • checks that each script dependency exists
  • computes the order to run the scripts depending on CONFIG_LIST order and dependencies of each config script.

Loads the file profile.default.sh if PROFILE is default (-p parameter). Providing CONFIG_LIST array of scripts to load.

3. Libraries

Some of the functions you can find:

  • lib/utils/Assert.sh provides utility functions to do some checks.
  • Functions::aptInstall use this function in order to be sure to use the correct default parameters it includes retry.
  • Functions::aptAddRepository
  • getGithubLatestRelease very useful in order to get last release version of a github project
  • upgradeGithubRelease
  • Assert::wsl
  • NetFunctions::addHost add host in /etc/hosts and windows hosts file if wsl (backup file before updates)
  • validator functions
  • Version::checkMinimal and Version::compare: version comparison function
  • log and logRotate functions
  • backup function
  • installFile/installDir install file or dir doing backup before and OVERRIDE_CONF variable aware
  • gitPullIfNoChanges/gitCloneOrPullIfNoChanges

4. Configurations

Configurations are folders in the scripts directory. when you create a new configuration, please be sure it sources the file lib/common.sh, as it initializes correctly the environment, load the main libraries, load the profile and set bash flags like errexit and pipefail.

Configuration should have the following files:

  • install
  • configure
  • test

It could have also the following optional files:

  • fortunes
  • dependencies
  • breakOnConfigFailure if present and config is failing, the whole process is stopped
  • breakOnTestFailure if present and test is failing, the whole process is stopped