Struct CoderRun 

pub struct CoderRun {
    pub emission_shape: String,
    pub model_id: String,
    pub files_written: Vec<String>,
    pub raw_reply: String,
    pub first_emission: String,
}

Outcome of one Coder::run turn. Surfaced via the ACP worker’s TaskReply.emission_shape so the foreman’s scorecard can distinguish T0a / T0b / T0c instead of lumping them as “empty diff”.

Fields

emission_shape: String

Wire-stable shape label: “whole_files”, “unified_diff”, “prose”.

model_id: String

Model id the inference backend returned.

files_written: Vec<String>

Relative paths of files the run wrote (empty for prose / diff — the diff path doesn’t tell us which files it touched without re-parsing).

raw_reply: String

The raw model reply. Useful for audit logs and post-mortem. NOTE: when the whole-file re-prompt fallback fires this becomes a composite first+retry transcript — use Self::first_emission when you need just the model’s initial output.

first_emission: String

The model’s first raw emission, before any re-prompt fallback. Always the initial reply (never a composite), so the eval scorecard can judge it with git apply --check (#30B) to tell a clean diff from a sloppy one the fuzzy worker merely rescued.

Trait Implementations

impl Clone for CoderRun

fn clone(&self) -> CoderRun

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for CoderRun

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

Formats the value using the given formatter.