Clarify README usage and platform support (#43)

* docs: clarify readme usage and platform support

* docs: split config from advanced workflows

* test: stabilize stdin tty smoke timing

* test: slow pager smoke key timing on ci

Author: benvinegar

README.md

@@ -23,15 +23,16 @@ npm i -g hunkdiff
 Requirements:
 
 - Node.js 18+
+- Currently supported on macOS and Linux
 - Git is recommended for most workflows
 
-## Basic usage
+## Usage
+
+### Basics
 
 ```bash
-hunk                                # show help
-hunk --version                      # show the installed version
-hunk diff before.ts after.ts        # compare two files directly
-git diff --no-color | hunk patch -  # review a patch from stdin
+hunk           # show help
+hunk --version # get version
 ```
 
 ### Working with Git
@@ -43,6 +44,13 @@ hunk show         # review the latest commit
 hunk show HEAD~1  # review an earlier commit
 ```
 
+### Working with raw files/patches
+
+```bash
+hunk diff before.ts after.ts        # compare two files directly
+git diff --no-color | hunk patch -  # review a patch from stdin
+```
+
 ## Feature comparison
 
 | Capability | hunk | difftastic | delta | diff |
@@ -90,24 +98,28 @@ Ready-to-run demo diffs live in [`examples/`](examples/README.md).
 
 Each example includes the exact command to run from the repository root.
 
-## Advanced features
+## Config
 
-- `hunk patch [file|-]` opens patch files or patch stdin
-- `hunk pager` opens Hunk for diff-like stdin and falls back to plain-text paging otherwise
-- `hunk diff --agent-context <file>` loads inline agent rationale from a JSON sidecar
-- `hunk mcp serve` runs the local MCP daemon for agent-to-diff communication
-- Hunk reads config from `~/.config/hunk/config.toml` and `.hunk/config.toml`
+Hunk reads config from:
+
+- `~/.config/hunk/config.toml`
+- `.hunk/config.toml`
 
-Minimal config example:
+Example:
 
 ```toml
-theme = "midnight"
-mode = "auto"
+theme = "midnight" # midnight, graphite, paper, ember
+mode = "auto"      # auto, split, stack
 line_numbers = true
 wrap_lines = false
 agent_notes = false
 ```
 
+## Advanced workflows
+
+- `hunk diff --agent-context <file>` loads inline agent rationale from a JSON sidecar
+- `hunk mcp serve` runs the local MCP daemon for agent-to-diff communication
+
 ## Performance notes
 
 Hunk spends more startup time than plain diff output tools because it launches an interactive UI with syntax highlighting, navigation state, and optional agent context. In exchange, it is optimized for reviewing a full changeset instead of printing static diff text and exiting.

test/tty-render-smoke.test.ts

@@ -152,8 +152,8 @@ async function runStdinPagerSmoke(options?: { input?: string; inputCommand?: str
   const transcript = join(fixture.dir, "stdin-pager-transcript.txt");
   const subcommand = options?.command === "pager" ? "pager" : "patch -";
   const patchCommand = `cat ${shellQuote(fixture.coloredPatch)} | bun run ${shellQuote(sourceEntrypoint)} ${subcommand}`;
-  const scriptCommand = `timeout 5 script -q -f -e -c ${shellQuote(patchCommand)} ${shellQuote(transcript)}`;
-  const inputCommand = options?.inputCommand ?? `(sleep 1; printf ${shellQuote(options?.input ?? "q")})`;
+  const scriptCommand = `timeout 7 script -q -f -e -c ${shellQuote(patchCommand)} ${shellQuote(transcript)}`;
+  const inputCommand = options?.inputCommand ?? `(sleep 2; printf ${shellQuote(options?.input ?? "q")})`;
   const proc = Bun.spawnSync(["bash", "-lc", `${inputCommand} | ${scriptCommand}`], {
     cwd: fixture.dir,
     stdin: "ignore",
@@ -243,7 +243,7 @@ describe("TTY render smoke", () => {
 
     const output = await runStdinPagerSmoke({
       lines: 40,
-      inputCommand: `(sleep 1; printf ' '; sleep 1; printf q)`,
+      inputCommand: `(sleep 2; printf ' '; sleep 2; printf q)`,
     });
 
     expect(output).toContain("before_23");