add request_builder_body_binary function.

[Tom-McQuillan]

The existing request_builder_body function only works with text data because it:

• Expects a C string (*const c_char): assumes null-terminated text
• Validates UTF-8 encoding: calls CStr::from_ptr(body).to_str() which fails on non-UTF-8 bytes
• Rejects binary data: returns error "Body contains invalid UTF-8" when it encounters bytes like 0x00, 0xFF, etc.

What request_builder_body_binary Provides:

• Accepts raw bytes (*const u8 + length): no assumption about text format
• No UTF-8 validation: treats data as pure binary
• Works with any file type: videos, images, PDFs, executables, etc.

[Tom-McQuillan]

On second thoughts, I think this is a bug. The request_builder_body should add the string as-is and not attempt to validate it.

My use case is that I need to upload video files that contain non-UTF8 values.

The request_builder_body_binary method I created works, but I think request_builder_body should pass binary data directly. This would work if we removed the UTF8 checking and stripped out the .to_str() methods.

[jimkring]

@Tom-McQuillan I think that James McNally was mentioning something similar about passing bytes instead of strings. Text encoding is a bit outside my area of experience so I don't have any off-hand advice right his moment. That said, I agree that we should not be trying to validate UTF8 -- especially since LabVIEW does not use UTF8 strings (see ni_lib_unicode).

Probably we need some unit+integration tests that define real-world use cases for different body payloads to ensure we've got the right approach and full coverage of user/API needs.

[jimkring]

@Tom-McQuillan the new multipart support has support for binary fields. See #32.