tombl

tombl makes bash viable for DevOps-automations that involve configurations saved as .toml files.

It allows bash to read .toml files structurally, so you don't have to come up with weird ad-hoc solutions involving awk, sed, and tears as soon as it breaks in production because you didn't use an actual toml-parser.

bash $ set -euo pipefail $ tombl -e DB=databases.hmm /etc/my-config.toml declare -A DB=(["user"]="postgreker" ["password"]="super secret" ["host"]="0.0.0.0" ["port"]=5432) $ eval "$(tombl -e DB=databases.hmm /etc/my-config.toml)" $ echo "${DB[user]}" postgreker $ pg_dumpall -h "${DB[host]}" -p "${DB[port]}" -u "${DB[user]}" > out.sql

Bash is unable to store nested arrays of any kind, so any nesting will be ignored when exporting, and you'll have to adapt your -e VAR=path.to.thing to access the nested information. It is recommended that you start your scripts with set -euo pipefail in order to fail fast™.

bash $ set -euo pipefail $ cat /etc/my-config.toml [databases.hmm] user = "postgreker" password = "super secret" host = "0.0.0.0" port = 5432 thing-that-is-nested = { will-not-be-included = 123 } $ tombl -e DB=databases.hmm /etc/my-config.toml declare -A DB=(["user"]="postgreker" ["password"]="super secret" ["host"]="0.0.0.0" ["port"]=5432) $ eval "$(tombl -e DB=databases.hmm /etc/my-config.toml)" $ echo "${DB[thing-that-is-nested]}" # whoops, but this will fail fast because of `set -euo` bash: l: unbound variable