Continuing the discussion from Getting a better experience with JSON/YAML configuration:
I’d like to use it for PhpStorm / WebStorm.
Continuing the discussion from Getting a better experience with JSON/YAML configuration:
I’d like to use it for PhpStorm / WebStorm.
Can’t you just use the plugin to generate the schema yourself? What’s the problem?
There is no problem, I’m suggesting to make it easier to use for people like me - one less step. Is it a bad suggestion to make things simpler?
I think you’re saying there actually is a problem though? That the schemas are hard to get, is that right?
The point of debate is whether the proposed solutions are actually simpler. (I believe this is the more correct link for the context of this topic than the one above.)
Well, it’s not difficult, it’s just boring and pulls me away from actually solving the problem I’m trying to solve with the hosting configuration.
From the developer point of view, it looks like an extra step that would be easy to eliminate by a bit of extra effort.
From the user point of view, it makes for a bad UX for newbies who don’t know Caddy very well yet. For example, I was trying to use the default Docker container without extra build steps that install any plugins. Now in order to produce a JSON schema I’m suggested to install plugins - does not sound appealing. What’s more, I’ll have to maintain two docker images - one without the plugin (for production), and one with the plugin (for my local dev environment). From my point of view, these all are needless extra efforts required from the Caddy users.
Could I add that it should be simper form the Caddy’s user prospective. It should absolutely not matter whether it is extra effort for the Caddy developers.
Just thinking aloud… because JSON is the primary format, should this Getting a better experience with JSON/YAML configuration be adopted into the core functionality, so that there is no need to install an extra plugin to get the JSON schema?
Another idea would be to make the schema exposable via admin API, so that people who use local development environment could just feed the schema’s URL to their IDEs.
To be clear, this is a problem with Docker, not a problem with Caddy. It’s due to the limitations of Docker images and tooling.
What does this look like exactly? You mean like this:
? That’s possible, depends on how much complexity it adds to the core though.
This topic was automatically closed after 30 days. New replies are no longer allowed.