You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Npm, pnpm and yarn can import packages from URLs meaning we should be able to dynamically generate packages on the server and show some UI on how to install them in your projects.
Some factors to consider:
How will we handle versioning.
Should we only use minor versions so the user doesn't have to update manually (only run npm i) or do we increment major versions (as there could be breaking changes).
When do we even bump the version number, currently changes in an ontologies are saved automatically so do we want to change the version for every single small commit?
I think the versions are embeded in the url so we could just generate a new url for the newer packages and have the user manually update their package.json. We have to keep hosting older versions tho.
Do we generate one package per ontology or group them together per drive?
How do we deal with the settings currently in the atomic.config.json file
Some settings like module alias and 'use namespaces' are configured in the atomic.config.json file. These settings are project specific and so it doesn't make sense to set these on a more global package level. Maybe integrate this into the url and generate the package at request time?
The text was updated successfully, but these errors were encountered:
Now that i'm thinking about it, generating the package at request time seems like a good option. We could embed everything into the URL. e.g. https://my-atomicserver.com/package/<commithash>?features=namespaces&alias=@tomic/react
Npm, pnpm and yarn can import packages from URLs meaning we should be able to dynamically generate packages on the server and show some UI on how to install them in your projects.
Some factors to consider:
Should we only use minor versions so the user doesn't have to update manually (only run npm i) or do we increment major versions (as there could be breaking changes).
When do we even bump the version number, currently changes in an ontologies are saved automatically so do we want to change the version for every single small commit?
I think the versions are embeded in the url so we could just generate a new url for the newer packages and have the user manually update their package.json. We have to keep hosting older versions tho.
Do we generate one package per ontology or group them together per drive?
How do we deal with the settings currently in the atomic.config.json file
Some settings like module alias and 'use namespaces' are configured in the
atomic.config.json
file. These settings are project specific and so it doesn't make sense to set these on a more global package level. Maybe integrate this into the url and generate the package at request time?The text was updated successfully, but these errors were encountered: