UTF safety

[robholland]

Fixes Encoding::UndefinedConversionError errors for some UTF characters.

[antstorm]

@robholland can you help me understand where this error comes from? Is it protobuf encoding complaining or GRCP? Or Temporal server?

[robholland]

It's the transcoding attempt that the protobuf code makes to try and turn the serialised JSON string from us (which contains UTF-8 characters which are valid in JSON) into ASCII-8BIT to fit in the bytes-typed data field for Temporal Payloads. UTF characters with the high bit set are not transcodable into ASCII-8BIT as they have no ascii equivalent so the transcode fails. Calling .b on the serialised JSON string escapes any high-bit-set characters which allows the string to be transcoded to ASCI-8BIT safely, and then deserialised on the other side into the correct UTF-8 characters.

[robholland]

Prior art here: protocolbuffers/protobuf#4228

[antstorm]

👍