More documentation on best practices around parsing output from scripts

[svermeulen]

The readme states the following:

The output of all the commands is much nicer for humans than it is for computers. If you'd like the output of commands, or to be able to load a config file in your own tooling, we provide a small api.

I'd like to obtain things like the list of errors/warnings and the files/line numbers. I could parse the output from cyan build but I think this is probably not good practice. The text above suggests there is an API I could use for this, but I'm unclear on what this refers to exactly?

Alternatively - maybe we have a command line option to control whether to execute in 'plumbing' mode vs 'porcelain' mode, similar to git. So all commands have stable, parseable output, in addition to the default that is optimized for reading by user

[euclidianAce]

This is definitely something I'd like to have and the api I was referring to was more about loading configs than running commands, but it would definitely be helpful to dump json or some structured format.

Not saying this to anyone specific but I think the way to do this would be to have a generic message emitting api (similar to the logging api) that knows what mode it's in to dump json or redirect to some logging function or user provided callback if they are using the api.

[vlaaad]

I'm looking forward to this feature, it would be very useful for integrating cyan into IDEs 👍