root/fusion
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

test(FN-WF): complete every pending step in the smoke executor mock

Browse Source 
The mock marked only step 0 done. A review gate that appends named remediation
work (`review-remediation-steps`) adds steps beyond the first, and a workflow whose
parse node disables trailing-step reopening depends on exactly that mechanism to
give a bounced card something to execute — so those steps stayed pending forever.
The mock now completes every pending step, as a real executor does.

This is necessary but not sufficient for S05 on builtin:coding-ideas-v2: the
remediation step IS appended and still ends pending
(`["Implement deterministic pipeline output:done",
   "Fix: The disposable fixture needs the scripted remediation commit.:pending"]`),
so the bounced card is not re-dispatched through an executor session that can
complete it. S05 and S07 therefore stay on their proven workflows rather than
shipping red.

Lane green: 6 files, 81 tests, 19/19 scenarios.
This commit is contained in:
Fusion Agent
2026-08-25 01:30:20 +00:00
parent cfe65527ca
commit d866617f40
2 changed files with 19 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1
packages/engine/src/__tests__/pipeline-smoke/_pipeline-harness.ts
View File

@@ -640,6 +640,7 @@ export class PipelineSmokeHarness {
behavior,
state: entry.state,
observeMockRuntime: () => this.guard.assertMockRuntime("mock/scripted"),
readTaskSteps: async () => (await this.freshTask(taskId).catch(() => undefined))?.steps?.map((step) => step.status) ?? [],
});
}

19
packages/engine/src/__tests__/pipeline-smoke/_pipeline-mock-scripts.ts
View File

@@ -95,6 +95,8 @@ export function installPipelineMockScripts(input: {
readonly behavior?: PipelineScriptedMergeBehavior;
readonly state?: PipelineMockScriptState;
readonly observeMockRuntime: () => void;
/** Current step statuses, read live so appended remediation work is completed too. */
readonly readTaskSteps: () => Promise<string[]>;
}): void {
const behavior = input.behavior ?? {};
const state = input.state ?? { planReviewIndex: 0, codeReviewIndex: 0, implementationCommitted: false };
@@ -268,7 +270,22 @@ export function installPipelineMockScripts(input: {
state.implementationCommitted = true;
}
if (hasTaskUpdateTool) {
await context.invokeTool("fn_task_update", { step: 0, status: "done" });
/*
FNXC:PipelineSmoke 2026-08-24-19:00:
Complete EVERY pending step, not only step 0. A review gate that appends named remediation
work (`review-remediation-steps`) adds steps beyond the first, and a workflow whose parse
node disables trailing-step reopening depends on exactly that mechanism to give the bounced
card something to execute. Marking only step 0 left those remediation steps pending forever,
so the card could never satisfy the completed-implementation projection and S05
("REVISE twice, then approve") stalled with
"did not persist completed implementation-step projection". A real executor finishes the
steps it was handed; the mock must too.
*/
const live = await input.readTaskSteps();
for (let index = 0; index < Math.max(live.length, 1); index += 1) {
if (live[index] && live[index] !== "pending" && live[index] !== "in-progress") continue;
await context.invokeTool("fn_task_update", { step: index, status: "done" });
}
}
/*
FNXC:PipelineSmoke 2026-08-23-20:23: