nes: fix: use speculativeRequestDelay for cached speculative results (#4985)

* test: add failing test for cached speculative result using wrong delay

When a speculative request completes and its result is cached before the
user accepts, the subsequent getNextEdit() takes the cache path where
isFromSpeculativeRequest stays false. This causes computeMinimumResponseDelay
to use the normal cacheDelay instead of the speculative-specific
speculativeRequestDelay (typically 0ms).

The test sets cacheDelay=5000ms and speculativeRequestDelay=0ms, then
verifies that a cached speculative result returns fast. Currently it
times out at 5000ms, confirming the bug.

* fix: use speculativeRequestDelay for cached speculative results

When a speculative request completes and caches its result before the
user accepts, the subsequent getNextEdit() takes the cache path. In
this path, isFromSpeculativeRequest was not being set from the cached
edit's source, causing computeMinimumResponseDelay to apply the normal
cacheDelay instead of the speculative-specific speculativeRequestDelay.

Set isFromSpeculativeRequest from cachedEdit.source.isSpeculative in
the cache path so that cached speculative results use the correct
(typically 0ms) delay.

Author: ulugbekna

src/extension/inlineEdits/node/nextEditProvider.ts

@@ -346,6 +346,7 @@ export class NextEditProvider extends Disposable implements INextEditProvider<Ne
 			}
 			isRebasedCachedEdit = !!cachedEdit.rebasedEdit;
 			isSubsequentCachedEdit = cachedEdit.subsequentN !== undefined && cachedEdit.subsequentN > 0;
+			isFromSpeculativeRequest = cachedEdit.source.isSpeculative;
 			req = cachedEdit.source;
 			logContext.setIsCachedResult(cachedEdit.source.log);
 			currentDocument = documentAtInvocationTime;

src/extension/inlineEdits/test/node/nextEditProviderSpeculative.spec.ts

@@ -1311,4 +1311,63 @@ describe('NextEditProvider speculative requests', () => {
 			await statelessProvider.calls[0].completed.p;
 		});
 	});
+
+	describe('cached speculative result delay', () => {
+		it('uses speculativeRequestDelay (not cacheDelay) when speculative result is served from cache', async () => {
+			const CACHE_DELAY_MS = 5_000;
+			await configService.setConfig(ConfigKey.TeamInternal.InlineEditsSpeculativeRequests, SpeculativeRequestsEnablement.On);
+			await configService.setConfig(ConfigKey.TeamInternal.InlineEditsCacheDelay, CACHE_DELAY_MS);
+			await configService.setConfig(ConfigKey.TeamInternal.InlineEditsSpeculativeRequestDelay, 0);
+
+			const statelessProvider = new TestStatelessNextEditProvider();
+			statelessProvider.enqueueBehavior({ kind: 'yieldEditThenNoSuggestions', edit: lineReplacement(1, 'const value = 2;') });
+			statelessProvider.enqueueBehavior({ kind: 'yieldEditThenNoSuggestions', edit: lineReplacement(2, 'console.log(value + 1);') });
+			const { nextEditProvider, workspace } = createProviderAndWorkspace(statelessProvider);
+
+			const doc = workspace.addDocument({
+				id: DocumentId.create(URI.file('/test/spec-cache-delay.ts').toString()),
+				initialValue: 'const value = 1;\nconsole.log(value);',
+			});
+			doc.setSelection([new OffsetRange(0, 0)], undefined);
+
+			// First request (fresh)
+			const firstSuggestion = await getNextEdit(nextEditProvider, doc.id);
+			assert(firstSuggestion.result?.edit);
+
+			// Show → triggers speculative request; wait for it to complete and cache
+			nextEditProvider.handleShown(firstSuggestion);
+			await statelessProvider.waitForCall(2);
+			await statelessProvider.calls[1].completed.p;
+
+			// Accept and apply — doc now matches speculative request's postEditContent
+			nextEditProvider.handleAcceptance(doc.id, firstSuggestion);
+			doc.applyEdit(firstSuggestion.result.edit.toEdit());
+
+			// Next getNextEdit hits the cache path (speculative result already cached).
+			// With enforceCacheDelay=true, it should use speculativeRequestDelay (0ms),
+			// NOT the normal cacheDelay (5000ms).
+			const context: NESInlineCompletionContext = {
+				triggerKind: 1,
+				selectedCompletionInfo: undefined,
+				requestUuid: generateUuid(),
+				requestIssuedDateTime: Date.now(),
+				earliestShownDateTime: Date.now(),
+				enforceCacheDelay: true,
+			};
+			const logContext = new InlineEditRequestLogContext(doc.id.toString(), 1, context);
+			const telemetryBuilder = new NextEditProviderTelemetryBuilder(gitExtensionService, mockNotebookService, workspaceService, nextEditProvider.ID, undefined);
+			const start = Date.now();
+			try {
+				const cachedSuggestion = await nextEditProvider.getNextEdit(doc.id, context, logContext, CancellationToken.None, telemetryBuilder.nesBuilder);
+				const elapsed = Date.now() - start;
+				assert(cachedSuggestion.result?.edit);
+
+				// The result comes from a speculative request's cache, so it should
+				// use the speculative delay (0ms) rather than the cache delay (5000ms)
+				expect(elapsed).toBeLessThan(100);
+			} finally {
+				telemetryBuilder.dispose();
+			}
+		});
+	});
 });