Hacker News Reader

Comment by sshine | Hacker News Reader

I wonder how widespread YAML 1.1 is.

If you assume that YAML 1.2 is the default, you don't need that nasty %YAML header.

This doesn't translate to arbitrary, open environments, but you can make that choice in closed environments.

While JSON numbers are grammatically simple, they're almost always distinct from how you'd implement numbers in any language that has JSON parsers, syntactically, exactness and precision-wise.

So while YAML is a lot more complex, you always need to limit yourself from what kinds of numbers you actually try to express in JSON. This is especially true for scientific numbers, big numbers, and numbers exact down to many digits.

View on HN

Among ecosystems based on YAML-formatted configuration defaulting to YAML 1.1 is nearly universal. The heyday of YAML was during the YAML 1.1 era, and those projects can't change their YAML parsers' default version to 1.2 without breaking extant config files.

By the time YAML 1.2 had been published and implementations written, greenfield projects were using either JSON5 (a true superset of JSON) or TOML.

  > While JSON numbers are grammatically simple, they're almost always distinct
  > from how you'd implement numbers in any language that has JSON parsers,
  > syntactically, exactness and precision-wise.

For statically-typed languages the range and precision is determined by the type of the destination value passed to the parser; it's straightforward to reject (or clamp) a JSON number `12345` being parsed into a `uint8_t`.

For dynamically-typed languages there's less emphasis on performance, so using an arbitrary-precision numeric type (Python's Decimal, Go's "math/big" types) provide lossless decoding.

The only language I know of that really struggles with JSON numbers is, ironically, JavaScript -- its BigInt type is relatively new and not well integrated with its JSON API[0], and it doesn't have an arbitrary-precision type.

[0] See https://developer.mozilla.org/en-US/docs/Web/JavaScript/Refe... for the incantation needed to encode a BigInt as a number.

> If you assume that YAML 1.2 is the default, you don't need that nasty %YAML header.

Indeed the YAML 1.2 spec says a document without a YAML directive should be assumed to be 1.2[1]:

> A version 1.2 YAML processor must accept documents with an explicit “%YAML 1.2” directive, as well as documents lacking a “YAML” directive. Such documents are assumed to conform to the 1.2 version specification.

It's only the YAML 1.2 spec that says it's a superset of JSON. The YAML authors weren't aware of JSON when publishing version 1.1[2]:

> The YAML 1.1 specification was published in 2005. Around this time, the developers became aware of JSON. By sheer coincidence, JSON was almost a complete subset of YAML (both syntactically and semantically).

> In 2006, Kyrylo Simonov produced PyYAML and LibYAML. A lot of the YAML frameworks in various programming languages are built over LibYAML and many others have looked to PyYAML as a solid reference for their implementations.

> The YAML 1.2 specification was published in 2009. Its primary focus was making YAML a strict superset of JSON. It also removed many of the problematic implicit typing recommendations.

The middle paragraph there is the reason this is a problem people keep running into: Most implementations are based on LibYAML, which is an implementation of YAML 1.1 that does not really support 1.2[3]. Indeed the last example from the post doesn't actually work for me on Ruby 3.4.4 with LibYAML 0.2.5. It produces the exact same output as the one before it.

[1]: https://yaml.org/spec/1.2.2/#681-yaml-directives

[2]: https://yaml.org/spec/1.2.2/#12-yaml-history

[3]: https://github.com/yaml/libyaml/issues/20

Reply on news.ycombinator.com