workspace activation

You activate a workspace by placing environment variables into your shell and modifying the PATH.

Shell environment variables represent the root config layer and all config resolving must happen from this base layer. If any code in your codebase must resolve config info it MUST have access to environment variables directly or indirectly or at the very least to the root path of the workspace so the environment variables may be established by loading and parsing the raw files.

Any config resolution logic must in its entirety be seeded by the environment variables and may then branch out with arbitrary logic. Ideally further config resolution layers follow rules and patterns that me be shared across implementations.

pio conventions

Given a file tree such as:

/project   <-   clone from github
/project/bin/activate.sh
/project.activate.sh
/project.profile.json

With file content such as:

/project.activate.sh

#!/bin/bash -e

# Digital Ocean - https://digitalocean.com/
export DIGIO_EMAIL="xxx"
export DIGIO_CLIENT_ID="xxx"
export DIGIO_API_KEY="xxx"

# pio credentials
export PIO_SEED_SALT="xxx"
export PIO_SEED_KEY="xxx"
export PIO_USER_ID="xxx"
export PIO_USER_SECRET="xxx"

# pio config
export PIO_PROFILE_PATH="/project.profile.json"

/project.profile.sh

{
    "config": {
        "pio.vm": {
            "adapter": "digitalocean"
        }
    }
}

/project/bin/activate.sh

#!/bin/bash

# TODO: Relocate into helper module.
# @credit http://stackoverflow.com/a/246128/330439
SOURCE="${BASH_SOURCE[0]:-$0}"
while [ -h "$SOURCE" ]; do
  DIR="$( cd -P "$( dirname "$SOURCE" )" && pwd )"
  SOURCE="$(readlink "$SOURCE")"
  [[ $SOURCE != /* ]] && SOURCE="$DIR/$SOURCE"
done
BASE_PATH="$( cd -P "$( dirname "$SOURCE" )" && pwd )"


echo "[pio] Switching environment ..."

export PATH=$BASE_PATH:$PATH

PS1="\[\033[1;34m\]\[\033[47m\](OS)\[\033[0m\] \[\033[1;35m\]$(basename $(dirname $BASE_PATH))\[\033[0m\] \[\033[33m\]\u\[\033[1;33m\]$\[\033[0m\] "


if [ ! -f "$BASE_PATH/../../$(basename $(dirname $BASE_PATH)).activate.sh" ]; then
  # Create /project.activate.sh file
fi

if [ -f "$BASE_PATH/../../$(basename $(dirname $BASE_PATH)).activate.sh" ]; then
  . $BASE_PATH/../../$(basename $(dirname $BASE_PATH)).activate.sh
fi

The following can be done:

The shell environment variables modified when calling source bin/activate.sh
The /project.activate.sh created by a script if it does not exist
The /project.activate.sh holding private credentials loaded into the shell environment
Establishing a root config layer based on shell environment variables
Arbitrary config resolution logic on top of root config layer
Shared config resolution logic across implementations by adhering to patterns

cadorn/WorkspaceActivation.md

pio conventions