From 74b4dd49c4eda38e306615f15ac171a7c4d7bdac Mon Sep 17 00:00:00 2001
From: Eric Eldredge <lettertwo@gmail.com>
Date: Wed, 3 Nov 2021 13:31:58 -0400
Subject: [PATCH] Fall back to FS for large blobs in LMDBCache

---
 packages/core/cache/src/LMDBCache.js | 55 ++++++++++++++++++----------
 1 file changed, 35 insertions(+), 20 deletions(-)

diff --git a/packages/core/cache/src/LMDBCache.js b/packages/core/cache/src/LMDBCache.js
index a9ac101a9de..2f8c988fa64 100644
--- a/packages/core/cache/src/LMDBCache.js
+++ b/packages/core/cache/src/LMDBCache.js
@@ -44,44 +44,46 @@ export class LMDBCache implements Cache {
   }
 
   has(key: string): Promise<boolean> {
-    return Promise.resolve(this.store.get(key) != null);
+    if (this.store.get(key) != null) return Promise.resolve(true);
+    return this.hasLargeBlob(key);
   }
 
-  get<T>(key: string): Promise<?T> {
-    let data = this.store.get(key);
-    if (data == null) {
-      return Promise.resolve(null);
-    }
-
-    return Promise.resolve(deserialize(data));
+  async get<T>(key: string): Promise<?T> {
+    let data = await this.getBuffer(key);
+    return data == null ? null : deserialize(data);
   }
 
   async set(key: string, value: mixed): Promise<void> {
-    await this.store.put(key, serialize(value));
+    await this.setBlob(key, serialize(value));
   }
 
   getStream(key: string): Readable {
-    return blobToStream(this.store.get(key));
+    return blobToStream(
+      this.store.get(key) ?? this.fs.readFileSync(path.join(this.dir, key)),
+    );
   }
 
   async setStream(key: string, stream: Readable): Promise<void> {
-    let buf = await bufferStream(stream);
-    await this.store.put(key, buf);
+    await this.setBlob(key, await bufferStream(stream));
   }
 
-  getBlob(key: string): Promise<Buffer> {
-    let buffer = this.store.get(key);
-    return buffer != null
-      ? Promise.resolve(buffer)
-      : Promise.reject(new Error(`Key ${key} not found in cache`));
+  async getBlob(key: string): Promise<Buffer> {
+    let buffer = await this.getBuffer(key);
+    if (buffer == null) throw new Error(`Key ${key} not found in cache`);
+    return buffer;
   }
 
   async setBlob(key: string, contents: Buffer | string): Promise<void> {
-    await this.store.put(key, contents);
+    if (isLargeBlob(contents)) await this.setLargeBlob(key, contents);
+    else await this.store.put(key, contents);
   }
 
-  getBuffer(key: string): Promise<?Buffer> {
-    return Promise.resolve(this.store.get(key));
+  async getBuffer(key: string): Promise<?Buffer> {
+    let buffer = this.store.get(key);
+    if (buffer == null && (await this.hasLargeBlob(key))) {
+      buffer = await this.getLargeBlob(key);
+    }
+    return buffer;
   }
 
   hasLargeBlob(key: string): Promise<boolean> {
@@ -97,4 +99,17 @@ export class LMDBCache implements Cache {
   }
 }
 
+// lmbd-store decodes cached binary data into a Node Buffer
+// via `Nan::NewBuffer`, which enforces a max size of ~1GB.
+// We subtract 9 bytes to account for any compression heaader
+// added by lmbd-store when encoding the data.
+// See: https://github.com/nodejs/nan/issues/883
+const MAX_BUFFER_SIZE = 0x3fffffff - 9;
+
+function isLargeBlob(contents: Buffer | string): boolean {
+  return typeof contents === 'string'
+    ? Buffer.byteLength(contents) > MAX_BUFFER_SIZE
+    : contents.length > MAX_BUFFER_SIZE;
+}
+
 registerSerializableClass(`${packageJson.version}:LMDBCache`, LMDBCache);