fix(session): type turn/end error as one structured failure
TurnEndReasonMap.error now carries a single `error: LlmFailure` field: an LlmError keeps its structured facts, any other error flattens to errorChain text under the UNKNOWN code. Consumers read message/code directly instead of defending against an unknown union — this also fixes errorChain() rendering structured failures as '[object Object]' in the TUI and ACP error paths. Document the turn-stopping contract: a concludesTurn result never short-circuits already-submitted next-step work (same-step additionalContexts or racing steering still runs), data decides.
This commit is contained in:
@@ -4,6 +4,7 @@ import type {
|
||||
CallId,
|
||||
LlmCallConfig,
|
||||
LlmCallConfigAdapterDefaults,
|
||||
LlmFailure,
|
||||
StreamChunk,
|
||||
TokenUsage,
|
||||
ToolResultMessage,
|
||||
@@ -109,9 +110,11 @@ export interface TurnEndReasonMap {
|
||||
|
||||
blocked: { kind: 'blocked' }
|
||||
/**
|
||||
* The turn failed.
|
||||
* The turn failed. `error` is always a structured failure: the `LlmError`
|
||||
* facts verbatim, or `{ message: errorChain(error), code: 'UNKNOWN' }`
|
||||
* flattened from any other error.
|
||||
*/
|
||||
error: { kind: 'error'; error: unknown }
|
||||
error: { kind: 'error'; error: LlmFailure }
|
||||
/** At least one step reached its output-token ceiling, even if a plugin continued the turn. */
|
||||
'max-tokens': { kind: 'max-tokens' }
|
||||
/**
|
||||
|
||||
Reference in New Issue
Block a user