feat: export cliFetch util

[pi0]

Summary by CodeRabbit

• Bug Fixes

  • CLI now exits with non-zero codes on failures and surfaces descriptive errors for non-OK responses.

• Improvements

  • CLI options renamed for consistency (dir instead of cwd/base, entry instead of url); downstream mappings updated.
  • Verbose mode provides clearer request/response debugging using provided I/O streams.
  • Response streaming preserves binary/text encoding behavior.

• New Features

  • Centralized CLI fetch path to standardize request construction, streaming, and error propagation.

_{✏️ Tip: You can customize this high-level summary in your review settings.}

[coderabbitai]

Caution

Review failed

The pull request is closed.

📝 Walkthrough

Walkthrough

Replaces inline CLI fetch with a new exported cliFetch that uses LoadOptions; renames loader options (url→entry, base→dir); CLI now delegates fetch to cliFetch, uses try/catch to route stdout/stderr, and exits with code 0/1 based on success or thrown errors.

Changes

Cohort / File(s)	Summary
CLI entrypoint & fetch src/cli.ts	Adds exported cliFetch(options: LoadOptions & { url?: string; method?: string; headers?: string[]; data?: string; verbose?: boolean; hostname?: string; stdin?: NodeJS.ReadStream; stdout?: NodeJS.WriteStream; stderr?: NodeJS.WriteStream }). Replaces inlined fetch with delegation to cliFetch, adds try/catch-driven exit codes, streams response, and surfaces non-OK responses as thrown errors.
Loader API & types src/loader.ts	Renames LoadOptions fields (url → entry, base → dir), updates entry resolution/discovery to use entry and dir, and exports the updated LoadOptions type for public consumption.

Estimated code review effort

🎯 3 (Moderate) | ⏱️ ~25 minutes

Possibly related PRs

• feat: loadServerEntry #168: Introduces or modifies the loader API and LoadOptions type used here; likely the predecessor or companion change to these renames and the loadServerEntry usage.

Poem

🐰 I hopped through code to shift a path,
entry and dir now do the math,
cliFetch carries header and stream,
verbose prints while night owls dream,
exit codes dance — a tidy craft.

🚥 Pre-merge checks | ✅ 2 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Docstring Coverage	⚠️ Warning	Docstring coverage is 0.00% which is insufficient. The required threshold is 80.00%.	Write docstrings for the functions missing them to satisfy the coverage threshold.

✅ Passed checks (2 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title accurately captures the primary change: exporting a new cliFetch utility function. This is the main public API addition in the changeset.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing touches

• 📝 Generate docstrings

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[coderabbitai]

Actionable comments posted: 0

Caution

Some comments are outside the diff and can’t be posted inline due to platform limitations.

⚠️ Outside diff range comments (2)

src/cli.ts (1)

157-185: Add duplex: "half" for streaming request bodies

When --data @- or `--data `@file is used, the body is a WHATWG ReadableStream. Node.js's fetch (via Undici) requires duplex: "half" for streaming request bodies; without it, new Request(...) throws TypeError: RequestInit: duplex option is required when sending a body.

Proposed fix

-    const req = new Request(url, {
-      method,
-      headers,
-      body,
-    });
+    const req = new Request(
+      url,
+      {
+        method,
+        headers,
+        body,
+        ...(body instanceof ReadableStream ? { duplex: "half" } : {}),
+      } as RequestInit & { duplex?: "half" },
+    );

src/loader.ts (1)

16-31: Handle file:// entries without breaking the notFound check

The docs say entry can be a URL, but resolve() + existsSync() will reject file://... inputs and prevent them from loading. The current code passes file URLs directly to resolve(), which treats them as path components, creating invalid paths like ./file:///path/to/server.ts. You should detect file URLs, convert them to paths using fileURLToPath for the existence check, and preserve the original URL for import.

🛠️ Proposed fix

-import { pathToFileURL } from "node:url";
+import { fileURLToPath, pathToFileURL } from "node:url";

-  let entry: string | undefined = opts.entry;
-  if (entry) {
-    entry = resolve(opts.dir || ".", entry);
-    if (!existsSync(entry)) {
-      return { notFound: true };
-    }
-  } else {
+  let entry: string | undefined = opts.entry;
+  let entryUrl: string | undefined;
+  if (entry) {
+    if (entry.startsWith("file://")) {
+      entryUrl = entry;
+      entry = fileURLToPath(entry);
+    } else {
+      entry = resolve(opts.dir || ".", entry);
+    }
+    if (!existsSync(entry)) {
+      return { notFound: true };
+    }
+  } else {
     for (const defEntry of defaultEntries) {
       for (const defExt of defaultExts) {
         const entryPath = resolve(opts.dir || ".", `${defEntry}${defExt}`);
         if (existsSync(entryPath)) {
           entry = entryPath;
           break;
         }
       }
       if (entry) break;
     }
     if (!entry) {
       return { notFound: true };
     }
   }
 
   // Convert to file:// URL for consistent imports
-  const url = entry.startsWith("file://") ? entry : pathToFileURL(resolve(entry)).href;
+  const url = entryUrl ?? pathToFileURL(entry!).href;

Also applies to: 90–115

[OskarLebuda]

LGTM ❤️

[codecov]

Codecov Report

❌ Patch coverage is 0% with 3 lines in your changes missing coverage. Please review.

Files with missing lines	Patch %	Lines
src/loader.ts	0.00%	1 Missing and 2 partials ⚠️

📢 Thoughts on this report? Let us know!