ROS2: Normalize byte as uint8 instead of int8

[achim-k]

Changelog

Description

In ROS2, byte fields should be treated as uint8 types:

Type name	C++	Python	DDS type
byte	uint8_t	builtins.bytes*	octet

Prior to this change we were treating byte fields as int8 which causes errors when parsing a byte field with a default value (e.g. 255) which is out of the range of a int8.

[jtbandes]

Seems fine.

Is it correct to normalize char to uint8? Is char really meant to be platform-dependent? 😭 (iirc @snosenzo dealt with this in IDL/CDR support?)

[jtbandes]

Maybe I'm just thinking of wchar: https://github.com/foxglove/omgidl/blob/fa6ff16507b9737aa896a4e7a8b5cf14cfd02a24/packages/omgidl-serialization/src/DeserializationInfoCache.ts#L205-L209

[jtbandes]

Actually just noticed this differs from how we handle the IDL primitives: https://github.com/foxglove/omgidl/blob/fa6ff16507b9737aa896a4e7a8b5cf14cfd02a24/packages/omgidl-parser/src/primitiveTypes.ts#L11-L12

Is that correct?

[snosenzo]

Actually just noticed this differs from how we handle the IDL primitives: https://github.com/foxglove/omgidl/blob/fa6ff16507b9737aa896a4e7a8b5cf14cfd02a24/packages/omgidl-parser/src/primitiveTypes.ts#L11-L12

@jtbandes I honestly think when I wrote that I was looking at what this repo was doing. So I think if we change it here we should probably also change it there