dvargas92495 / roamjs-scripts Goto Github PK

I'm trying to start learning how roam extensions are built but I am trying to keep the dev environment in docker containers. I have managed to build and launch the dev server for my test extension as per the instructions. However, because this is executed inside a docker container, and the host is hardcoded to 127.0.0.1, I cannot connect to it from outside the container, even with the right port bindings in place.

For this scenario, I'd like to set the host to 0.0.0.0 as instructed here. Ideally without having to patch these dependencies.

Would you consider adding a new command-line option or even env var to allow overriding the webpack dev server host? or even even a flag to set it to 0.0.0.0?

Thanks.

It's one of the requirements to be able to serve the "depot" extension from URL: https://roamresearch.com/#/app/developer-documentation/page/n8ECay0Wl

I am in the process of hacking the Google Calendar extension to support a custom formatter, and as part of my experimental work, I want to write some tests to limit the extent of the destruction I'd be causing 💥 . This being my first exposure to TS, Jest, and the whole JS dev ecosystem (I come from a world of C++ and python) means I am getting stuck every two steps, but that is to be expected :)

Anyway, I have seen that this repo contains a test and it runs! 😅 but when I have tried to extract the minimal elements that will allow me to run a test on the Gcal extension I have not been successful (yet). This is the motivator for this request. It would be of great help if the init command, together with the other files it generates, could also output a test file and the appropriate config that will make it run.

Some of my ideas around it:

1. It should import something from the main script, even if dummy
2. There should be a test in the test file
3. The test should fail by default (to encourage devs to actually touch the tests)
4. This should be the default behaviour, but there should be a flag to disable it, e.g. init --no-test

I'm creating this issue to see if you like the idea. I'm happy to contribute the changes if you do. When I figure out how to get that dummy test to run, that is.