Merge based on file type

backend/sync_lib/src/lib.rs

@@ -19,11 +19,17 @@ pub mod errors;
 /// Encode binary data for easy transport over HTTP. Inverse of
 /// `base64_to_bytes`.
 #[wasm_bindgen(js_name = bytesToBase64)]
-pub fn bytes_to_base64(input: &[u8]) -> String { STANDARD.encode(input) }
+pub fn bytes_to_base64(input: &[u8]) -> String {
+    set_panic_hook();
+
+    STANDARD.encode(input)
+}
 
 /// Inverse of `bytes_to_base64`.
 #[wasm_bindgen(js_name = base64ToBytes)]
 pub fn base64_to_bytes(input: &str) -> Result<Vec<u8>, SyncLibError> {
+    set_panic_hook();
+
     STANDARD.decode(input).map_err(SyncLibError::from)
 }
 
@@ -32,6 +38,8 @@ pub fn base64_to_bytes(input: &str) -> Result<Vec<u8>, SyncLibError> {
 /// documents is binary.
 #[wasm_bindgen]
 pub fn merge(parent: &[u8], left: &[u8], right: &[u8]) -> Vec<u8> {
+    set_panic_hook();
+
     if is_binary(parent) || is_binary(left) || is_binary(right) {
         right.to_vec()
     } else {
@@ -47,6 +55,8 @@ pub fn merge(parent: &[u8], left: &[u8], right: &[u8]) -> Vec<u8> {
 /// WASM wrapper around `reconcile::reconcile` for text merging.
 #[wasm_bindgen(js_name = mergeText)]
 pub fn merge_text(parent: &str, left: &str, right: &str) -> String {
+    set_panic_hook();
+
     reconcile::reconcile(parent, left, right)
 }
 
@@ -54,6 +64,8 @@ pub fn merge_text(parent: &str, left: &str, right: &str) -> String {
 /// content.
 #[wasm_bindgen(js_name = isBinary)]
 pub fn is_binary(data: &[u8]) -> bool {
+    set_panic_hook();
+
     if data.iter().any(|&b| b == 0) {
         // Even though the NUL character is valid in UTF-8, it's highly suspicious in
         // human-readable text.
@@ -63,10 +75,18 @@ pub fn is_binary(data: &[u8]) -> bool {
     std::str::from_utf8(data).is_err()
 }
 
-/// Set up panic hook for better error messages in the browser console.
+/// We don't want to supporte merging structured data like JSON, YAML, etc.
-#[cfg(feature = "console_error_panic_hook")]
+#[wasm_bindgen(js_name = isFileTypeMergable)]
-#[wasm_bindgen(js_name = setPanicHook)]
+pub fn is_file_type_mergable(path_or_file_name: &str) -> bool {
-pub fn set_panic_hook() {
+    set_panic_hook();
+
+    let file_extension = path_or_file_name.split('.').last().unwrap_or_default();
+
+    matches!(file_extension.to_lowercase().as_str(), "md" | "txt")
+}
+
+fn set_panic_hook() {
     // https://github.com/rustwasm/console_error_panic_hook#readme
+    #[cfg(feature = "console_error_panic_hook")]
     console_error_panic_hook::set_once();
 }

backend/sync_lib/tests/web.rs

@@ -1,5 +1,3 @@
-//! Test suite for the Web and headless browsers.
-
 use insta::assert_debug_snapshot;
 use sync_lib::*;
 use wasm_bindgen_test::*;
@@ -44,3 +42,21 @@ fn test_is_binary() {
     assert!(is_binary(&[0, 12]));
     assert!(!is_binary(b"hello"));
 }
+
+#[wasm_bindgen_test(unsupported = test)]
+fn test_is_binary_empty() {
+    assert!(!is_binary(b""));
+}
+
+#[wasm_bindgen_test(unsupported = test)]
+fn test_is_file_type_mergable() {
+    assert!(is_file_type_mergable(".md"));
+    assert!(is_file_type_mergable("hi.md"));
+    assert!(is_file_type_mergable("my/path/to/my/document.md"));
+    assert!(is_file_type_mergable("hi.MD"));
+    assert!(is_file_type_mergable("my/path/to/my/DOCUMENT.MD"));
+
+    assert!(!is_file_type_mergable(".json"));
+    assert!(!is_file_type_mergable("HELLO.JSON"));
+    assert!(!is_file_type_mergable("my/config.yml"));
+}

backend/sync_server/src/server/create_document.rs

@@ -10,7 +10,7 @@ use chrono::{DateTime, Utc};
 use log::info;
 use schemars::JsonSchema;
 use serde::Deserialize;
-use sync_lib::{base64_to_bytes, merge};
+use sync_lib::{base64_to_bytes, is_file_type_mergable, merge};
 
 use super::{
     app_state::AppState,
@@ -127,11 +127,15 @@ async fn internal_create_document(
             )));
         }
 
-        let merged_content = merge(
+        let merged_content = if is_file_type_mergable(&sanitized_relative_path) {
-            &[], // the empty string is the first common parent of the two documents,
+            merge(
-            &existing_version.content,
+                &[], // the empty string is the first common parent of the two documents,
-            &content,
+                &existing_version.content,
-        );
+                &content,
+            )
+        } else {
+            content
+        };
 
         let new_version = StoredDocumentVersion {
             vault_id,

backend/sync_server/src/server/update_document.rs

@@ -10,7 +10,7 @@ use chrono::{DateTime, Utc};
 use log::info;
 use schemars::JsonSchema;
 use serde::Deserialize;
-use sync_lib::{base64_to_bytes, merge};
+use sync_lib::{base64_to_bytes, is_file_type_mergable, merge};
 
 use super::{
     app_state::AppState,
@@ -155,7 +155,12 @@ async fn internal_update_document(
         )));
     }
 
-    let merged_content = merge(&parent_document.content, &latest_version.content, &content);
+    let merged_content = if is_file_type_mergable(&sanitized_relative_path) {
+        merge(&parent_document.content, &latest_version.content, &content)
+    } else {
+        content.clone()
+    };
+
     let is_different_from_request_content = merged_content != content;
 
     // We can only update the relative path if we're the first one to do so