Ian Lyttle

Now fixed in R-devel r83370 (https://github.com/wch/r-source/commit/593df7418e92a771904da730ace2dcb8bf5d221e).

@HenrikBengtsson just a quick note to say thanks for nudging this issue along with R-Core, and to R-Core for fixing it! Also, of course, thanks for hosting this repo as...

Awesome, thanks for the heads-up!

Thinking of making this issue dormant for now - still interested in being able to validate a spec natively but it is not (yet) critical to be able to do...

note to self - think about implementing this as an internal generic. We could have an internal generic `.validate` with methods `.validate.vegaspec_vega` and `.validate.vegaspec_vegalite`. Our constructor `as_vegaspec()` would have an...

Wondering now if `validate()` should not be it's own function, separate from `as_vegaspec()` or any of the others. Thinking this because I also think that **V8** should be a suggested...

Some good news - jsonvalidate is moving to ajv, and there is a branch! https://github.com/ropensci/jsonvalidate/issues/2

new release of jsonvalidate removes the roadblock. Just as a note to myself, but it seems like one small error can lead to an avalanche of validation feedback. This might...

This is all implemented in #661 and #681. Big thanks to @juliasilge and @machow for being so patient and generous with your talents!

I think we are thinking along the same lines here; that *building* a spec and *performing surgery* on a spec are two different things. I know I am bring lazy...