# Release proposal for libdd-data-pipeline and its dependencies

This PR contains version bumps based on public API changes and commits
since last release.

## libdd-trace-utils
**Next version:** `5.0.0`
**Semver bump:** `major`
**Tag:** `libdd-trace-utils-v5.0.0`

### Commits

- feat(trace-utils)!: add from_string to span text (#2011)
- feat!: add encoder from v04 to v1 (#1896)

## libdd-trace-obfuscation
**Next version:** `3.1.0`
**Semver bump:** `minor`
**Tag:** `libdd-trace-obfuscation-v3.1.0`

### Commits

- fix(libdd-trace-obfuscation): cargo clippy fix with all lints (#1947)

## libdd-trace-stats (bumped manually)
**Next version:** ~~`3.1.0`~~`4.0.0`
**Semver bump:** ~~`minor`~~`major`
**Tag:** ~~`libdd-trace-stats-v3.1.0`~~`libdd-trace-stats-v4.0.0`

### Commits

- feat(trace-utils)!: add from_string to span text (#2011)
- feat!: add encoder from v04 to v1 (#1896)

## libdd-data-pipeline
**Next version:** `5.0.0`
**Semver bump:** `major`
**Tag:** `libdd-data-pipeline-v5.0.0`

### Commits

- feat(trace-utils)!: add from_string to span text (#2011)
- feat!: add encoder from v04 to v1 (#1896)
- fix(data-pipeline): allow old PascalCase fields in obfuscation config
scheme (#2008)
- feat(trace_buffer)!: flush based on size of chunks in bytes (#1953)

---------

Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
Co-authored-by: iunanua <[email protected]>
Co-authored-by: iunanua <[email protected]>

…ble expect (#1946)

# What does this PR do?        

Replaces the `use_v05_format`: bool parameter in `collect_trace_chunks` with the existing TraceEncoding enum, and removes `#[allow(clippy::expect_used)] + .expect("infallible...")` in the v04 msgpack encoder.

# Motivation                                              

Follow-up to [#1896](#1896), addressing review comments:                                                                                                                                   

- Passing a bare bool to `collect_trace_chunks` gave no context at the call site — you had to look up the function signature to understand what true/false meant. Using TraceEncoding makes the intent explicit.
- The `collect_and_process_traces` match had a catch-all arm that computed a bool from the format — replaced with exhaustive V04/V05 arms.
- The v04 encoder used `.expect("infallible...")` with a clippy allow to handle an error that can never occur. Replaced with `let _ = ...`, consistent with what was already done in the v1 encoder.

# Additional Notes

⚠️  Should be merged after [#1896](#1896).

# How to test the change?

Existing tests.

Co-authored-by: anais.raison <[email protected]>

# What does this PR do?

Adds the canonical V1 trace payload data model (`v1::Span`, `TracerPayload`, `TraceChunk`, typed `AttributeValue`, `SpanKind`, 128-bit `trace_id`, `error: bool`) and a matching msgpack encoder that serializes this model into the V1 wire format. New public APIs: `to_vec_from_payload`, `to_vec_from_payload_with_capacity`, `write_payload_to_slice`, `to_encoded_byte_len_from_payload`.

# Motivation

[APMSP-2810](https://datadoghq.atlassian.net/browse/APMSP-2810), M3 of the V1 trace payload rollout. Tracers will eventually produce V1 payloads natively rather than going through the v0.4→V1 conversion path implemented in [#1896](#1896), so libdatadog needs a canonical V1 data model and a corresponding encoder.

# Additional Notes

- The V1 wire format keys (`SpanKey`, `SpanLinkKey`, `SpanEventKey`, `AnyValueKey`, plus the `trace_key` and `chunk_key` modules) live in `msgpack_encoder/v1/mod.rs` and are shared between the v0.4→V1 encoder and the v1::Span encoder.
- Output is byte-identical to the M1 v0.4→V1 encoder for equivalent inputs.


[APMSP-2810]: https://datadoghq.atlassian.net/browse/APMSP-2810?atlOrigin=eyJpIjoiNWRkNTljNzYxNjVmNDY3MDlhMDU5Y2ZhYzA5YTRkZjUiLCJwIjoiZ2l0aHViLWNvbS1KU1cifQ

Co-authored-by: anais.raison <[email protected]>