Resolve conflicts with upstream main for timeline undo

.changeset/eight-seas-matter.md

@@ -0,0 +1,5 @@
+---
+'@nanocollective/nanocoder': patch
+---
+
+Defer ink and @/app loading in the CLI entry point until the interactive TUI branch, so --acp, --plain and auth paths no longer pay the Ink/App module-graph cost at startup.

.changeset/feat-vscode-settings-dashboard.md

@@ -0,0 +1,5 @@
+---
+"@nanocollective/nanocoder": minor
+---
+
+Added Settings Tab to the VS Code extension webview for configuring providers and behavior directly from the UI.

.changeset/fix-cancel-stops-execution.md

@@ -0,0 +1,5 @@
+---
+"@nanocollective/nanocoder": patch
+---
+
+Fixed the VS Code extension's stop button leaving a request running. Two holes: `AcpSession.cancel()` aborted the current controller and immediately replaced it with a fresh one, so a cancel that landed before the turn read the signal — the window while the agent is still resolving the prompt's file references — handed the turn an unaborted controller and the stop was lost. The controller is now rotated when a turn begins instead. Separately, the extension never answered the agent's pending permission request when you hit stop: the tool card kept its spinner and Allow/Deny buttons, and because the request stayed on the pending list every later message was refused with "Please approve or deny the pending tool before sending a new message" until the window was reloaded. Stopping (or reconnecting after the agent process restarts) now resolves those requests as cancelled. Thanks to @akramcodez. Closes #864.

.changeset/fix-vscode-tool-aggregator-reset.md

@@ -0,0 +1,11 @@
+---
+"@nanocollective/nanocoder": patch
+---
+
+Fixed new tool calls landing back in an earlier card instead of a fresh one when a thought, reply, edit card, or plan update came in between. Closes #856.
+
+Fixed a manually collapsed tool card re-expanding on its next update.
+
+Reused one footer per agent turn instead of creating one per text segment.
+
+Fixed a turn's copy button sometimes copying a newer turn's text instead of its own.

.changeset/loud-plums-invite.md

@@ -0,0 +1,5 @@
+---
+"@nanocollective/nanocoder": patch
+---
+
+Setup wizard's config location picker now shows the resolved path next to each option instead of a bare label.

.changeset/sunset-setup-commands.md

@@ -0,0 +1,5 @@
+---
+"@nanocollective/nanocoder": minor
+---
+
+Sunset `/setup-providers` and `/setup-mcp` in favour of `/settings`. Both retired names still work for now — they open the matching `/settings` tab with a notice instead of erroring. `/settings` now takes a tab argument (`/settings providers`, `/settings mcp`), MCP has its own settings tab, and provider edits made from settings apply to the running session instead of waiting for the next launch. Selecting a provider or MCP server in settings now opens that entry's edit/delete choice directly, with a separate row for adding a new one.

.changeset/vscode-code-lens-actions.md

@@ -0,0 +1,5 @@
+---
+"@nanocollective/nanocoder": minor
+---
+
+Added editor code lenses to the VS Code extension: every function, method, constructor and class now carries `Explain Code` and `Generate Tests` links, and clicking one reveals the chat view and sends that symbol - instruction, `file:startLine-endLine` and the source, fenced with the document language - as a prompt. Symbols come from the language server, so no per-language parsing is involved, and the lenses can be turned off with `nanocoder.codeLens`. Long symbols are capped before being inlined, so a lens click on a large class cannot spend a whole context window on one turn. Also fixes a pre-existing hang where sending a message while a tool approval was still pending left the composer spinning forever. Closes #750.

.github/workflows/pr-checks.yml

@@ -62,6 +62,11 @@ jobs:
       - name: Run TypeScript type checking
         run: pnpm test:types
 
+      # The root tsconfig only includes source/**, so the extension's own
+      # sources and specs are checked by their own project or not at all.
+      - name: Run TypeScript type checking (VS Code extension)
+        run: pnpm test:types:vscode
+
   test-format:
     name: Format Checks
     runs-on: ubuntu-latest

benchmarks/measure.ts

@@ -122,34 +122,30 @@ function measureModuleCount(args: string[]): number {
 }
 
 /**
- * Boots the CLI in interactive mode under the counting loader, waits until
- * the module-resolution count has been stable for `stableMs`, then kills the
- * process and returns the final count.
+ * Boots the CLI under the counting loader, waits until the module-resolution
+ * count has been stable for `stableMs`, then kills the process and returns
+ * the final count plus approximate wall-clock boot time.
  *
- * This measures the full app graph (Ink, hooks, tool manager, command
- * registry, MCP loader, etc.) rather than just the `--help` fast path.
+ * This measures the full resolved graph for the given args (interactive TUI,
+ * ACP server, etc.) rather than just a oneshot `--help` fast path.
  * Polling-until-stable keeps the result deterministic regardless of machine
  * speed: we wait for steady state, not a fixed wall-clock window.
- */
-/**
- * Same polling-until-stable pattern as `measureInteractiveModuleCount`, but
- * returns both the module count AND the wall-clock time (in ms) from process
- * spawn to the moment the module count reached its final value.
  *
  * The wall-clock number is labelled **approximate** because it's
  * machine-dependent and subject to I/O / scheduler noise. It's not a
  * deterministic metric like module count — it's a sanity check that
  * optimizations translate into real boot-time savings on the local machine.
  */
-async function measureInteractiveStartup(
+async function measureStartupUntilStable(
 	args: string[] = [],
+	label = 'interactive',
 ): Promise<{moduleCount: number; bootMs: number}> {
 	const loaderPath = path.join(__dirname, 'count-loader.mjs');
 	const loaderUrl = new URL(`file://${loaderPath}`).href;
 	const countFile = path.join(
 		repoRoot,
 		'benchmarks',
-		`.module-count-interactive-${process.pid}-${Date.now()}`,
+		`.module-count-${label}-${process.pid}-${Date.now()}`,
 	);
 
 	const child = spawn(
@@ -356,7 +352,7 @@ async function runWithUrlCapture(
 					CI: '1',
 					// Keep the user's global MCP config out of the graph
 					// so the explain breakdown only reflects nanocoder's own
-					// boot cost. Matches `measureInteractiveStartup`.
+					// boot cost. Matches `measureStartupUntilStable`.
 					NODE_ENV: 'test',
 				},
 			},
@@ -709,13 +705,23 @@ export async function measure(): Promise<Measurements> {
 	// clock boot number; the module count is deterministic so we only keep
 	// the first one.
 	const interactiveRuns = [
-		await measureInteractiveStartup(),
-		await measureInteractiveStartup(),
-		await measureInteractiveStartup(),
+		await measureStartupUntilStable([], 'interactive'),
+		await measureStartupUntilStable([], 'interactive'),
+		await measureStartupUntilStable([], 'interactive'),
 	];
 	const interactiveModuleCount = interactiveRuns[0].moduleCount;
 	const bootSamples = interactiveRuns.map(r => r.bootMs).sort((a, b) => a - b);
 	const interactiveBootMsMedian = bootSamples[1];
+	// ACP server is long-lived (JSON-RPC over stdin/stdout). Poll until the
+	// module graph is stable, then terminate — same pattern as interactive.
+	const acpRuns = [
+		await measureStartupUntilStable(['--acp'], 'acp'),
+		await measureStartupUntilStable(['--acp'], 'acp'),
+		await measureStartupUntilStable(['--acp'], 'acp'),
+	];
+	const acpModuleCount = acpRuns[0].moduleCount;
+	const acpBootSamples = acpRuns.map(r => r.bootMs).sort((a, b) => a - b);
+	const acpBootMsMedian = acpBootSamples[1];
 	// Time from spawn to first stdout byte — what the user actually perceives.
 	const firstRenderMs = await measureTimeToFirstRender();
 	const distStats = measureDir(path.join(repoRoot, 'dist'));
@@ -810,6 +816,18 @@ export async function measure(): Promise<Measurements> {
 				value: interactiveBootMsMedian,
 				warnOnDecrease: false,
 			},
+			// ACP server steady-state module graph (no Ink / App). Deterministic
+			// proxy for editor-integration startup cost.
+			acp_module_count: {
+				kind: 'numeric',
+				value: acpModuleCount,
+				warnOnDecrease: false,
+			},
+			acp_boot_ms_approx: {
+				kind: 'numeric',
+				value: acpBootMsMedian,
+				warnOnDecrease: false,
+			},
 			// Time from spawn to first stdout byte — what the user sees.
 			// Measured without the ESM loader so there's no instrumentation
 			// overhead. Median of 3 runs.

docs/configuration/mcp-configuration.md

@@ -38,7 +38,7 @@ Create a `.mcp.json` file in your project root:
 }
 ```
 
-Use `/mcp` to view connected servers and their tools. Use `/setup-mcp` for interactive setup.
+Use `/mcp` to view connected servers and their tools. Use `/settings mcp` for interactive setup.
 
 ## Optional: Local-First Cross-Session Memory
 
@@ -241,7 +241,7 @@ Supported syntax: `$VAR`, `${VAR}`, `${VAR:-default}`
 
 ## Setup Wizard
 
-Run `/setup-mcp` for interactive configuration with:
+Run `/settings mcp` for interactive configuration with:
 
 - Pre-configured templates for popular servers (Filesystem, GitHub, Brave Search, Context7, DeepWiki, Playwright, etc.)
 - Custom server setup for stdio, HTTP, and WebSocket