From 204c743bcc825e2f9aaa2b502d57de3020ed2969 Mon Sep 17 00:00:00 2001
From: mpostma <postma.marin@protonmail.com>
Date: Tue, 16 Mar 2021 14:09:01 +0100
Subject: [PATCH] add json payload check on document addition

---
 Cargo.lock                                    |  7 ++++
 meilisearch-http/Cargo.toml                   |  1 +
 .../src/index_controller/update_actor.rs      | 38 +++++++++++++++++--
 3 files changed, 42 insertions(+), 4 deletions(-)

diff --git a/Cargo.lock b/Cargo.lock
index bcdcec71c..85ebb9c02 100644
--- a/Cargo.lock
+++ b/Cargo.lock
@@ -1823,6 +1823,7 @@ dependencies = [
  "milli",
  "mime",
  "once_cell",
+ "oxidized-json-checker",
  "parking_lot",
  "rand 0.7.3",
  "rayon",
@@ -2122,6 +2123,12 @@ dependencies = [
  "num-traits",
 ]
 
+[[package]]
+name = "oxidized-json-checker"
+version = "0.3.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "938464aebf563f48ab86d1cfc0e2df952985c0b814d3108f41d1b85e7f5b0dac"
+
 [[package]]
 name = "page_size"
 version = "0.4.2"
diff --git a/meilisearch-http/Cargo.toml b/meilisearch-http/Cargo.toml
index 9ae7baae2..e61ecf647 100644
--- a/meilisearch-http/Cargo.toml
+++ b/meilisearch-http/Cargo.toml
@@ -61,6 +61,7 @@ tempfile = "3.1.0"
 thiserror = "1.0.24"
 tokio = { version = "1", features = ["full"] }
 uuid = "0.8.2"
+oxidized-json-checker = "0.3.2"
 
 [dependencies.sentry]
 default-features = false
diff --git a/meilisearch-http/src/index_controller/update_actor.rs b/meilisearch-http/src/index_controller/update_actor.rs
index 1e788ff34..abf2ab8bc 100644
--- a/meilisearch-http/src/index_controller/update_actor.rs
+++ b/meilisearch-http/src/index_controller/update_actor.rs
@@ -1,13 +1,15 @@
 use std::collections::{hash_map::Entry, HashMap};
+use std::io::SeekFrom;
 use std::fs::{create_dir_all, remove_dir_all};
 use std::path::{Path, PathBuf};
 use std::sync::Arc;
 
-use super::index_actor::IndexActorHandle;
 use log::info;
+use oxidized_json_checker::JsonChecker;
+use super::index_actor::IndexActorHandle;
 use thiserror::Error;
-use tokio::fs::File;
-use tokio::io::AsyncWriteExt;
+use tokio::fs::OpenOptions;
+use tokio::io::{AsyncWriteExt, AsyncSeekExt};
 use tokio::sync::{mpsc, oneshot, RwLock};
 use uuid::Uuid;
 
@@ -125,7 +127,11 @@ where
         let update_store = self.store.get_or_create(uuid).await?;
         let update_file_id = uuid::Uuid::new_v4();
         let path = self.path.join(format!("update_{}", update_file_id));
-        let mut file = File::create(&path)
+        let mut file = OpenOptions::new()
+            .read(true)
+            .write(true)
+            .create(true)
+            .open(&path)
             .await
             .map_err(|e| UpdateError::Error(Box::new(e)))?;
 
@@ -146,7 +152,31 @@ where
             .await
             .map_err(|e| UpdateError::Error(Box::new(e)))?;
 
+        file.seek(SeekFrom::Start(0))
+            .await
+            .map_err(|e| UpdateError::Error(Box::new(e)))?;
+
+        let mut file = file.into_std().await;
+
         tokio::task::spawn_blocking(move || {
+            use std::io::{BufReader, sink, copy, Seek};
+
+            // If the payload is empty, ignore the check.
+            if file.metadata().map_err(|e| UpdateError::Error(Box::new(e)))?.len() > 0 {
+                // Check that the json payload is valid:
+                let reader = BufReader::new(&mut file);
+                let mut checker = JsonChecker::new(reader);
+
+                if copy(&mut checker, &mut sink()).is_err() || checker.finish().is_err() {
+                    // The json file is invalid, we use Serde to get a nice error message:
+                    file.seek(SeekFrom::Start(0))
+                        .map_err(|e| UpdateError::Error(Box::new(e)))?;
+                    let _: serde_json::Value = serde_json::from_reader(file)
+                        .map_err(|e| UpdateError::Error(Box::new(e)))?;
+                }
+            }
+
+            // The payload is valid, we can register it to the update store.
             update_store
                 .register_update(meta, path, uuid)
                 .map(UpdateStatus::Pending)