Page 9 - Bash

CSV exports from ERP systems, inventory management, or legacy databases almost always need to become JSON before they can be imported into modern APIs. A CSV to JSON pipeline built from mlr, jq, and Bash glue code handles type coercion, nesting, and validation without a single line of Python.

Passing a password as a command line argument leaves it permanently in Bash history, visible to anyone who later uses the same shell. This article shows how secrets end up in Bash history, the process list, and logs, how HISTCONTROL and safe handoff methods prevent that, and how to search existing history files for forgotten credentials.

A good changelog generator reads the commit history, recognizes the type and scope of every change, and builds a structured CHANGELOG.md from it, with no manual editing required. This article shows how such a changelog generator can be built as a plain Bash script, what pitfalls appear when parsing Conventional Commits, and how the result plugs cleanly into a release pipeline.

When a production system starts behaving unexpectedly, the quality of your diagnostic toolset decides whether the outage lasts minutes or hours. ps, top, ss, lsof and journalctl together provide a complete picture, provided you know the right flags and combine them into targeted Bash helpers.

Shell-Skripte ohne Tests sind eine Blackbox in der Automatisierung. Das Bash Automated Testing System (BATS) bringt strukturierte Unit-Tests, Fixtures und Mocking in die Shell-Welt – und lässt sich nahtlos in GitHub Actions, GitLab CI und Jenkins integrieren, ohne externe Laufzeitabhängigkeiten.

A fixed path like /tmp/script.tmp in a shell script is an open invitation for race conditions, data leaks, and symlink attacks. mktemp guarantees unique, safe temporary files, and trap EXIT ensures they get cleaned up in every termination scenario.

The most frustrating pattern in CI/CD practice: a Bash script runs flawlessly locally and fails in the pipeline with a cryptic error. The cause is rarely a bug in the script itself. It is a difference in the execution environment. PATH, login shells, missing environment variables, different Bash versions, and Docker container differences are the systematic traps that break Bash in CI/CD.

Without shared rules, shell codebases end up where every script follows a different style, error handling is missing or inconsistent, and nobody knows which checks are mandatory in review. A review checklist, clear style rules, EditorConfig and ShellCheck as a CI gate lay the groundwork so teams can review shell scripts without debating every pull request from first principles.

Having a bats test suite for Bash scripts tells you nothing about which lines are actually executed. Code coverage with kcov closes that gap, shows line by line what tests really cover, and makes untested error paths visible long before they become a problem in production.

A backup that has never been tested is just a hypothesis. Bash backup scripts that combine rsync for incremental snapshots with tar for archived copies, clean up rotating backups according to a configurable retention policy, and run restore tests automatically make the difference between a data protection routine and a verified recovery process.