fix(core): batch calls to git hash-object to avoid E2BIG error

packages/workspace/src/core/hasher/git-hasher.ts

@@ -6,18 +6,42 @@ export async function getGitHashForFiles(
   potentialFilesToHash: string[],
   path: string
 ): Promise<{ hashes: Map<string, string>; deleted: string[] }> {
-  const filesToHash = [];
+  const { filesToHash, deleted } = getActualFilesToHash(
-  const deleted = [];
+    potentialFilesToHash,
-  for (let f of potentialFilesToHash) {
+    path
-    if (fileExists(joinPathFragments(path, f))) {
+  );
-      filesToHash.push(f);
-    } else {
-      deleted.push(f);
-    }
-  }
 
   const res: Map<string, string> = new Map<string, string>();
+  const promises: Promise<Map<string, string>>[] = [];
   if (filesToHash.length) {
+    // On windows the max length is limited by the length of
+    // the overall comand, rather than the number of individual
+    // arguments. Since file paths are large and rather variable,
+    // we use a smaller batchSize.
+    const batchSize = process.platform === 'win32' ? 500 : 4000;
+    for (
+      let startIndex = 0;
+      startIndex < filesToHash.length;
+      startIndex += batchSize
+    ) {
+      promises.push(
+        getGitHashForBatch(
+          filesToHash.slice(startIndex, startIndex + batchSize),
+          path
+        )
+      );
+    }
+  }
+  // Merge batch results into final result set
+  const batchResults = await Promise.all(promises);
+  for (const batch of batchResults) {
+    batch.forEach((v, k) => res.set(k, v));
+  }
+  return { hashes: res, deleted };
+}
+
+async function getGitHashForBatch(filesToHash: string[], path) {
+  const res: Map<string, string> = new Map<string, string>();
   const hashStdout = await spawnProcess(
     'git',
     ['hash-object', ...filesToHash],
@@ -34,8 +58,23 @@ export async function getGitHashForFiles(
     const filePath: string = filesToHash[i];
     res.set(filePath, hash);
   }
+  return res;
+}
+
+function getActualFilesToHash(
+  potentialFilesToHash: string[],
+  path: string
+): { filesToHash: string[]; deleted: string[] } {
+  const filesToHash = [];
+  const deleted = [];
+  for (let f of potentialFilesToHash) {
+    if (fileExists(joinPathFragments(path, f))) {
+      filesToHash.push(f);
+    } else {
+      deleted.push(f);
     }
-  return { hashes: res, deleted };
+  }
+  return { filesToHash, deleted };
 }
 
 async function spawnProcess(command: string, args: string[], cwd: string) {