Struct TaskReply 

pub struct TaskReply {
    pub model_id: String,
    pub content: String,
    pub diff: String,
    pub empty_diff: bool,
    pub diff_applied: bool,
    pub emission_shape: Option<String>,
    pub raw_emission: Option<String>,
}

Structured reply for one prompt turn.

Contract

Per workspace memory feedback_drake_patch_not_prose and feedback_empty_diff_is_a_crash:

• model_id is mandatory. It is a non-Option String so the field cannot be silently omitted from the wire format. Use TaskReply::new for the validated constructor that rejects an empty model_id — foreman uses this field to attribute the patch and update the model’s scorecard, so a missing id is non-recoverable.
• empty_diff: true means the worker produced no real edits and foreman should disqualify it pre-arbiter.
• diff_applied: true means a unified diff was found in content and newt_tools::apply_patch accepted it.

Fields

model_id: String

MANDATORY — the model that produced this reply.

content: String

Assistant content (typically the unified diff).

diff: String

Captured workspace diff after the turn.

empty_diff: bool

True if the captured diff is empty (no real changes).

diff_applied: bool

True if a unified diff was detected in content and applied successfully.

emission_shape: Option<String>

Set by the newt-coder plugin: “whole_files”, “unified_diff”, or “prose” (the wire-stable constants in plugins_protocol::emission_shape). None when the legacy newt-flat path produced the reply.

Lets the foreman’s scorecard distinguish failure modes T0a / T0b / T0c instead of lumping them together as “empty diff”.

raw_emission: Option<String>

The model’s first raw emission — newt-coder’s CoderRun.first_emission, or the flat path’s reply content.

The eval diff_applies evaluator runs git apply --check against this (when it is diff-shaped) rather than the post-hoc captured diff, so a model that emits an unappliable diff the fuzzy worker only rescued is scored honestly (#30B). None for legacy payloads.

Implementations

impl TaskReply

pub fn new( model_id: impl Into<String>, content: impl Into<String>, diff: impl Into<String>, diff_applied: bool, ) -> Result<Self>

Validated constructor. Rejects an empty model_id so a buggy backend can’t silently produce an unattributable reply.

pub fn with_emission_shape(self, shape: impl Into<String>) -> Self

Builder: attach the emission shape label the newt-coder plugin produced. The legacy newt-flat path leaves this None.

pub fn with_raw_emission(self, raw: impl Into<String>) -> Self

Builder: attach the model’s first raw emission so the eval diff_applies evaluator can judge it with the strict oracle (#30B).

Trait Implementations

impl Clone for TaskReply

fn clone(&self) -> TaskReply

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for TaskReply

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl<'de> Deserialize<'de> for TaskReply

fn deserialize<__D>(__deserializer: __D) -> Result<Self, __D::Error>

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl PartialEq for TaskReply

fn eq(&self, other: &TaskReply) -> bool

Tests for self and other values to be equal, and is used by ==.

fn ne(&self, other: &Rhs) -> bool

Tests for !=. The default implementation is almost always sufficient, and should not be overridden without very good reason.

impl Serialize for TaskReply

fn serialize<__S>(&self, __serializer: __S) -> Result<__S::Ok, __S::Error>

where __S: Serializer,

Serialize this value into the given Serde serializer.

impl Eq for TaskReply

impl StructuralPartialEq for TaskReply