fix(docker): use target binary build param as name of image entrypoint

[SuperFluffy]

Summary

Fixes container images to use the $TARGETBINARY build param as the name of the binary run in the image entrypoint.

Background

Before this patch, the containerfile entrypoint used a symlinked /usr/local/bin/entrypoint as the entrypoint, which caused the name of the binary to be "entrypoint". This itself was to workaround to the ENTRYPOINT command not allowing for variables in exec form. This patch instead creates an ad-hoc script which itself executes $TARGETBINARY, and uses this script as the entrypoint.

Together with the script using exec and the exec form of ENTRYPOINT the expected binary (and process) name are used.

Taking astria-cli's help message as an example:

Before:

Get the balance of a Sequencer account

Usage: entrypoint sequencer balance get [OPTIONS] <ADDRESS>

After:

Get the balance of a Sequencer account

Usage: astria-cli sequencer balance get [OPTIONS] <ADDRESS>

Changes

• Create an adhoc /usr/local/bin/entrypoint.sh script in containerfiles/Dockerfile which is used as a new entrypoint and executes $TARGETBINARY

Testing

Built and ran images using the dockerfile. No functionality was changed but the binary is $TARGETBINARY and not entrypoint now.

Related Issues

closes #1571