pi-openai-toolkit 0.6.0 → 0.7.1

package/src/extension-runtime.ts

@@ -6,7 +6,7 @@ import type {
 	SessionBeforeCompactEvent,
 } from "@earendil-works/pi-coding-agent";
 import { loadToolkitConfig } from "./config";
-import { writeDebugArtifact } from "./debug";
+import { writeDebugArtifact, writeReplayFailureArtifact } from "./debug";
 import { resolveLatestNativeCompactionEntry } from "./details-store";
 import { runNativeFallbackCompaction } from "./native-fallback";
 import {
@@ -15,6 +15,7 @@ import {
 } from "./payload-rewrite";
 import { getCompactionRequestExtras, rememberRequestContext } from "./request-context-cache";
 import { executeRemoteV2Compaction } from "./remote-v2-client";
+import { registerInlineCompactionRuntime } from "./inline-compaction";
 import {
 	resolveNativeCompactionEnvironment,
 	resolveRemoteCompactionExecution,
@@ -25,6 +26,7 @@ import {
 	createNativeCompactionDetails,
 	createNativeCompactionResult,
 	COMPACTION_EXTENSION_ID,
+	getLatestDeferredToolCarryover,
 	isNativeCompactionDetails,
 	type CompactionConfig,
 	type NativeCompactionDetails,
@@ -99,7 +101,8 @@ async function runResponsesNativeCompact(
 ): Promise<ResponsesCompactOutcome> {
 	const { consumer, compactor } = execution;
 	const instructions = buildCompactionInstructions(ctx.getSystemPrompt(), event.customInstructions);
-	const branchEntries = ctx.sessionManager.getBranch();
+	const branchEntries = event.branchEntries ?? ctx.sessionManager.getBranch();
+	const deferredToolCarryover = getLatestDeferredToolCarryover(branchEntries);
 	const latestNativeCompaction = resolveLatestNativeCompactionEntry(branchEntries, {
 		baseUrl: consumer.baseUrl,
 	});
@@ -217,6 +220,7 @@ async function runResponsesNativeCompact(
 				model: compactor.model,
 				baseUrl: compactor.baseUrl,
 			},
+			deferredToolCarryover,
 			compactedWindow: compactResult.compactedWindow,
 			compactResponseId: compactResult.compactResponseId,
 			createdAt: compactResult.createdAt,
@@ -504,6 +508,30 @@ async function handleBeforeProviderRequest(event: BeforeProviderRequestEvent, ct
 			config,
 			ctx,
 		);
+
+		// Fail loud instead of letting Pi send the sentinel-only payload: the
+		// compacted history would be silently lost while the request still succeeds.
+		// A forced redacted failure record is written even when logProviderPayloads
+		// is disabled so the incident is diagnosable without leaking content.
+		writeReplayFailureArtifact(
+			{
+				reason: rewrite.reason,
+				parity: rewrite.parity,
+				compactionEntryId: latestNativeCompactionEntry.id,
+				provider: runtime.provider,
+				api: runtime.api,
+				model: runtime.model,
+			},
+			config,
+			ctx,
+		);
+		if (ctx.hasUI) {
+			ctx.ui.notify(
+				`${COMPACTION_EXTENSION_ID}: native compaction replay failed (${rewrite.reason}); provider request aborted`,
+				"error",
+			);
+		}
+		ctx.abort();
 		return undefined;
 	}
 
@@ -537,6 +565,8 @@ async function handleBeforeProviderRequest(event: BeforeProviderRequestEvent, ct
 }
 
 export default function (pi: ExtensionAPI) {
+	const inlineCompaction = registerInlineCompactionRuntime(pi, loadToolkitConfig);
+
 	pi.on("session_start", (_event, ctx) => {
 		const { config: toolkitConfig, source, warnings } = loadToolkitConfig();
 		const config = toolkitConfig.compaction;
@@ -570,4 +600,6 @@ export default function (pi: ExtensionAPI) {
 
 	pi.on("session_before_compact", handleSessionBeforeCompact);
 	pi.on("before_provider_request", handleBeforeProviderRequest);
+	pi.on("turn_end", (event, ctx) => inlineCompaction.handlePublicTurnEnd(event, ctx));
+	pi.on("session_shutdown", (_event, ctx) => inlineCompaction.dispose(ctx));
 }