Description
Does it make sense to divide docker-entrypoint.sh into multiple files, so that the functionality encompassed can be used piecewise? Issues like #242 highlight some of the problems I've had with it doing deployments, and the proliferation of a nearly infinite series of articles on how to develop and deploy wordpress with docker indicates the pain points with the current system are affecting many people.
Converting docker-entrypoint.sh into a dispatcher script which would be shorter and easier to replace with a custom configuration could help with this. Both docker-entrypoint.sh and a customized configuration can then source and/or execute helper scripts as needed. Entrypoint becomes more readable and still obeys the principle of consistency.
The scripts would then (hopefully) be more maintainable because they focus on a single responsibility. For example, the logic which copies wordpress over can be moved to entrypoint/copy_wordpress.sh. Similarly the logic which overwrites wp-config.php with environment variable support can be entrypoint/enable_env_vars.sh.
Configuring behavior for these scripts can be implemented. If the scripts are segmented they're also easier to replace. I can provide my own entrypoint (or my own copy_wordpress) which does what I want without obliterating the rest of the entrypoint code.
Preserving as much of the configuration as possible is also helpful. Some of what docker-entrypoint.sh is doing is security focused. I'd like to be able to keep as much of that as possible while opting out of what I don't want.
Thoughts? Hate mail? 🍌s?