Sanity checking an idea for editing yaml without wanting to throw my laptop out the window
Often find myself getting frustrated editing yaml, and it seems to be used everywhere for some reason I cannot fathom
I have an idea to write an editor plugin that will, when opening a yaml file, convert it to json (or some other less painful configuration language), then convert back on save. I don't know enough about yaml syntax to know if that's possible or if there's some quirk that makes them not completely cross compatible
Or alternatively if it exists a better CLI tool for editing yaml than just a normal text editor because I'm getting sick of pasting in a block of yaml and then having to fix the 8 indentation errors that somehow spawn from that
YAML to JSON is probably doable, JSON back to YAML not so much.
There are multiple ways to mark multiline strings in YAML. Then there are anchors, like bionicjoey mentioned. Also comments, YAML has them. You'd have to have some way to retain the extra information, if you want to make the full round trip.
Here's an example:
def-db: &def-db
# here be dragons
login: admin
passwd: nimda
prod:
db: *def-db
desc: |
I'm a teapot
short and stout
dev:
db:
<<: *def-db
passwd: pass
desc: "I'm a teapot\nshort and stout\n"
Comments are an issue I'd have to think about. Would prebuilt libraries for importing/exporting data from/to these languages not handle the multiline strings for me?
What do anchors do in yaml I've never heard of them before
I think the difference is that it sounds they are just looking for something JSON-like, just enough to edit and save a change. It might not need to be valid.
So, a new not-a-markup-language, only human readable and editable, and objectively better than its predecessor? Well, it's all according to tradition. I believe YAML got its start the same way.
Not a terrible idea to add bits into the JSON but it makes it much harder to program, going for a quick and easy fix here rather than writing my own parser
Maybe a 3rd file would work? You could add all of the relevant data there and when translating between one language or the other it would prune any comments or unsupported features as the output is generated.
My plan is not to have multiple files at all and only have the JSON be in memory, or at least if there is a second one have it be temporary and deleted on exit
The problem with the approach of having an intermediary format is I need to create the intermediate format, and given yaml starts for yet another markup language that doesn't seem like a good idea (also a ton of effort to reinvent the wheel)