Fork me on GitHub
Ben Sless06:09:46

I know there's a lot of focus on performance at the moment but what's the status of json-schema and the missing bits, especially date/time?


good question. • JSON Schema -> malli, have been waiting for the PR to progress, might be stalled • Date/time, @henryw374 might have an insight on the new js/temporal lib status. If there would be a good set of abstractions for clj/s, malli should use them. Or, just a new alpha ns from the existing clj-time things • others: recursive generation, proper inferring, schematized defn, typescript-compat, map+map-of, local registries refined, ... lot of things to do, some things are started by various contributors, many just ideas. After the ast-cleanup, have clojurists-together -time to work on many things, but we (all) could do design for the missing features and mark them as "PR welcome".

Ben Sless11:09:45

I hope you don't mind, I went ahead and opened a draft PR to add time schemas in clj. I think it's 90% there and I'd rather have 90% to work with vs. 100% in time t > T

Ben Sless11:09:25

looking at json-schema -> malli gave me a headache, because the schema is versioned! do you want to support "everything" or just latest?

Ben Sless11:09:24

This has real implications, for example, when I was wrangling vega in #datavis malli could have been a great help but vega uses an old version of json schema


I tried to categorize the issues, but could add more explicit tags on those that are discussed already and would be ready to be picked up for anyone wanting to contribute.


the ast-change is fundamental, planning to finish a first version of that now. Few days of work I would guess, mainly tests.


I know of aave and snoop for instrumentation, is there a library in the malli ecosystem that does automatic generative testing of functions?