downloadFileContent(): Follow URL redirects in Node [#585]

src/tests/webR/webr-worker.test.ts

@@ -60,8 +60,45 @@ describe('Download and install binary webR packages', () => {
     expect(await pkg.toBoolean()).toEqual(true);
     warnSpy.mockRestore();
   });
+
+  test('Install packages from R-universe using shim', async () => {
+    const warnSpy = jest.spyOn(console, 'warn').mockImplementation(() => null);
+    const pkg = (await webR.evalR(`
+      stopifnot(! "boot" %in% installed.packages()[, "Package"])
+      webr::shim_install()
+      install.packages("boot", repos = "https://cran.r-universe.dev")
+      "boot" %in% installed.packages()[, "Package"]
+    `)) as RLogical;
+    expect(await pkg.toBoolean()).toEqual(true);
+    warnSpy.mockRestore();
+  });
+});
+
+describe('Download files', () => {
+  test('download.file() without URL redirect', async () => {
+    const warnSpy = jest.spyOn(console, 'warn').mockImplementation(() => null);
+    const pkg = (await webR.evalR(`
+      tf <- tempfile()
+      utils::download.file("https://cran.r-project.org/web/packages/rlang/index.html", tf)
+      utils::file_test("-f", tf)
+    `)) as RLogical;
+    expect(await pkg.toBoolean()).toEqual(true);
+    warnSpy.mockRestore();
+  });
+
+  test('download.file() with URL redirect', async () => {
+    const warnSpy = jest.spyOn(console, 'warn').mockImplementation(() => null);
+    const pkg = (await webR.evalR(`
+      tf <- tempfile()
+      utils::download.file("https://cran.r-project.org/package=rlang", tf)
+      utils::file_test("-f", tf)
+    `)) as RLogical;
+    expect(await pkg.toBoolean()).toEqual(true);
+    warnSpy.mockRestore();
+  });
 });
 
+
 describe('Test webR virtual filesystem', () => {
   const testFileContents = new Uint8Array([1, 2, 4, 7, 11, 16, 22, 29, 37, 46]);
   test('Upload a file to the VFS', async () => {

src/webR/webr-worker.ts

@@ -592,9 +592,13 @@ function copyFSNode(obj: FSNode): FSNode {
   return retObj;
 }
 
-function downloadFileContent(URL: string, headers: Array<string> = []): XHRResponse {
+function downloadFileContent(url: string, headers: Array<string> = [], maxRedirects = 10): XHRResponse {
+  if (maxRedirects <= 0) {
+    return { status: 400, response: 'Too many redirects' };
+  }
+
   const request = new XMLHttpRequest();
-  request.open('GET', URL, false);
+  request.open('GET', url, false);
   request.responseType = 'arraybuffer';
 
   try {
@@ -610,15 +614,46 @@ function downloadFileContent(URL: string, headers: Array<string> = []): XHRRespo
 
   try {
     request.send(null);
-    const status = IN_NODE
-      ? (JSON.parse(String(request.status)) as { data: { statusCode: number } }).data.statusCode
-      : request.status;
+
+    let status: number;
+
+    if (IN_NODE) {
+      const parsed = JSON.parse(String(request.status)) as {
+        data: { statusCode: number; headers: Record<string, string> }
+      };
+      status = parsed.data.statusCode;
+
+      // Follow 3xx redirects
+      if (status >= 300 && status < 400) {
+        const location = parsed.data.headers?.location;
+
+        if (location) {
+          // Resolve relative URLs against the original URL
+          let redirectUrl: string;
+          try {
+            redirectUrl = new URL(location, url).href;
+          } catch (error) {
+            let responseText: string;
+            if (error instanceof TypeError) {
+              responseText = "Invalid redirect URL format";
+            } else {
+              responseText = "Unexpected redirect URL error";
+            }
+            console.error(responseText + ":", error);
+            return { status: 400, response: responseText };
+          }
+          return downloadFileContent(redirectUrl, headers, maxRedirects - 1);
+        }
+      }
+    } else {
+      status = request.status;
+    }
 
     if (status >= 200 && status < 300) {
       return { status: status, response: request.response as ArrayBuffer };
     } else {
       const responseText = new TextDecoder().decode(request.response as ArrayBuffer);
-      console.error(`Error fetching ${URL} - ${responseText}`);
+      console.error(`Error fetching ${url} - ${responseText}`);
       return { status: status, response: responseText };
     }
   } catch {