Break down cloud tool

[spbnick]

The cloud tool's source code is a single file 2312 lines long, which is shameful, and makes it hard to read, navigate, and maintain.

Break it down into smaller modules (.sh include files). Put modules into the kcidb directory. Group functions and variables by prefix, and put them into correspondingly-named files. E.g. all functions and variables with prefix psql_/PSQL_ should go into a file named kcidb/psql.sh, and so on.

The original cloud file must add the directory it's located in in front of the PATH environment variable (following its syntax, i.e. using the : path separator), and then include the modules using the . (source) command. E.g. like this:

export PATH="$(dirname "$(realpath "$0")")/kcidb:$PATH"

. psql.sh

Each module file should start with a short description of its contents, and should have anti-recursive inclusion protection around its code and definitions. E.g. the kcidb/psql.sh file should have:

# PostgreSQL database deployment definitions and management

if [ -z "${_PSQL_SH+set}" ]; then
declare _PSQL_SH=

# Code and definitions go here

fi # _PSQL_SH

Various definitions reused by multiple files, that are too few for a separate module, should go into kcidb/misc.sh, but don't have to have a misc_ prefix added to their names.

[spbnick]

Ah, this is for kcidb, not kcidb-io.