No support for bytes

[ndmitchell]

From the spec:

A bytes is an immutable sequence of values in the range 0-255. The type of a bytes is "bytes".

We don't yet support bytes.

[nataliejameson]

Probably good to double check with this that we support byte string literals with this as well, then.

[thesayyn]

would it be really tricky to add support for bytes?

[ndmitchell]

Probably not very hard at all. It's just no one has yet had a use case.

[thesayyn]

We are investing in starlark-rust, i might be able contribute this myself.

[thesayyn]

I am trying to implement bytes type for starlark-rust and bumped into this limitation.

So the spec says;

str(bytes) would decode bytes to UTF-K, replacing invalid sequences by U+FFFD.
bytes(str) (that is, a new predeclared function called "bytes") would encode a text string to UTF-8, replacing invalid code units with the encoding of U+FFFD. (If K=8 and str is a valid encoding, the result is the same as the input.)
repr(bytes) would return a byte string literal such as b"ABC".

But starlark-rust implemetation of str(bytes) and repr(bytes) is basically a call to collect_repr on StarlarkValue trait which makes it not possible to distinguish between the two.

Would it make sense to add a to_string method to StarlarkValue trait that calls collect_repr by default and allow types to return a different string for two cases?

[thesayyn]

#177 is an attempt at fixing this. Can maintainers PTAL?