merge: resolve origin dev refactor conflicts

test: harden allocation checks
fix: close remaining PR review gaps
2026-08-16 09:56:15 +02:00 · 2026-08-16 09:37:28 +02:00 · 2026-08-16 09:35:26 +02:00 · 2026-08-16 09:11:58 +02:00 · 2026-08-15 20:38:04 +02:00 · 2026-08-15 20:02:09 +02:00
40 changed files with 1618 additions and 526 deletions
@@ -58,7 +58,7 @@ endif()
 find_package(OpenSSL REQUIRED)

 file(GLOB SHARED_SRCS "src/shared/*.c")
-set(FILE_STORE_SRCS src/shared/file_store.c)
+set(FILE_STORE_SRCS "${CMAKE_CURRENT_SOURCE_DIR}/src/shared/file_store.c")
 list(REMOVE_ITEM SHARED_SRCS ${FILE_STORE_SRCS})
 file(GLOB SERVER_SRCS "src/server/*.c")
 set(SERVER_RECEIVER_SRCS src/server/receiver.c)
@@ -1,4 +1,4 @@
-# FastSync
+#FastSync

 FastSync is a high-performance file synchronization tool designed to become a
 drop-in replacement for common `rsync` workflows. It keeps the familiar
@@ -45,7 +45,7 @@ replacement for every rsync feature or protocol mode.

 - Recursive directory scanning.
 - Rsync-style source and destination arguments.
- SSH transport using `user@host:/path` destinations.
+- SSH transport using `user@host:destination` paths below the remote authorized root.
 - TCP client/server transfers.
 - Dry runs, excludes, includes, size filters, backups, statistics, and
  bandwidth limiting.
@@ -79,8 +79,166 @@ partial, alternate, and planned behavior.

 ### Build

-Requirements: C11 compiler, CMake 3.22 or newer, zstd, OpenSSL, pthreads,
-and an SSH client for SSH transport.
+### Client
+
+| Argument | Description |
+|----------|-------------|
+| Positional | `<source> <dest>` — automatic SSH detection if dest contains `:` |
+| `-c [level]` | Compression with optional level (1–22, default 5) |
+| `-z [level]` | Alias for `-c` |
+| `-a, --archive` | Archive mode: enables `-c -m -M` (no `-s`) |
+| `-m` | Multithreading mode |
+| `-s` | Chunk serialization (batch all files per chunk) |
+| `-f, --sendfile` | Sendfile zero-copy. Incompatible with `-c` / `-s`. TCP only. |
+| `-M, --preserve` | Preserve supported file metadata (mode and mtime; ownership and atime are unsupported) |
+| `-n, --dry-run` | Scan and print what would be transferred |
+| `-p <port>` | SSH port (default: 22) |
+| `-v, --verbose` | Enable debug logging |
+| `--progress` | Show real-time transfer speed |
+| `--delete` | Delete files on receiver not present in source |
+| `--exclude <pattern>` | Exclude files matching glob pattern (repeatable) |
+| `--exclude-from <file>` | Read exclude patterns from a file (one per line) |
+| `--include <pattern>` | Only transfer files matching glob pattern (repeatable, whitelist) |
+| `--max-size <n>` | Skip files larger than n bytes |
+| `--min-size <n>` | Skip files smaller than n bytes |
+| `--incremental` | Skip files unchanged since last transfer (size + mtime). Auto-enables `--preserve`. Incompatible with `-s`. |
+| `--bwlimit <KB/s>` | Bandwidth limit in kilobytes per second |
+| `--chunk-size <n>` | Chunk size in bytes (default: 10485760) |
+| `--timeout <sec>` | I/O timeout in seconds (default: 30) |
+| `--contimeout <sec>` | Connection timeout in seconds (default: 10) |
+| `--backup` | Backup existing destination files before overwriting |
+| `--backup-dir <dir>` | Target directory for backups (requires `--backup`) |
+| `--stats` | Print transfer statistics at end (bytes, files, timing) |
+| `--max-depth <n>` | Maximum directory depth to recurse (0 = unlimited, default: 0) |
+| `--log-file <path>` | Write log messages to file instead of stderr |
+| `--source-dir <path>` | Source directory (overrides `FASTSYNC_SOURCE_DIR`) |
+| `--dest-dir <path>` | Server destination directory (overrides `FASTSYNC_DEST_DIR`) |
+| `--save-to-disk` | Write received files to disk |
+| `--server-host <ip>` | Server IP address (default: `127.0.0.1`) |
+| `--server-port <n>` | Server port (default: `8080`) |
+| `--tls` | Enable TLS encryption |
+| `--cert <path>` | TLS certificate file (PEM) |
+| `--key <path>` | TLS private key file (PEM) |
+| `--ca <path>` | TLS CA certificate file for verification (PEM) |
+| `--client-cn <name>` | Required TLS client certificate common name |
+
+### Server
+
+| Argument | Description |
+|----------|-------------|
+| `--stdio` | Run in stdio mode (for SSH transport; single connection then exits) |
+| `-p <port>` | TCP listen port (default: 8080, range: 1–65535) |
+| `--tls` | Enable TLS encryption |
+| `--cert <path>` | TLS certificate file (PEM) |
+| `--key <path>` | TLS private key file (PEM) |
+| `--ca <path>` | TLS CA certificate file for verification (PEM) |
+| `--destination-root <path>` | Authorized destination root (default: `.`) |
+| `--allow-delete` | Permit manifest deletion |
+| `--allow-unauthenticated` | Permit plaintext TCP clients |
+| `-v, --verbose` | Enable debug logging |
+| `--help` | Show help |
+
+## Environment Variables
+
+| Variable | Default | Description |
+|----------|---------|-------------|
+| `FASTSYNC_SOURCE_DIR` | — | Source directory fallback |
+| `FASTSYNC_DEST_DIR` | — | Destination directory fallback |
+| `FASTSYNC_SAVE_TO_DISK` | `false` | Disk persistence fallback |
+| `FASTSYNC_SSH_PORT` | `22` | Default SSH port |
+| `FASTSYNC_SERVER_HOST` | `127.0.0.1` | Default server host |
+| `FASTSYNC_SERVER_PORT` | `8080` | Default server port |
+| `FASTSYNC_TLS_CERT` | — | Default TLS certificate path |
+| `FASTSYNC_TLS_KEY` | — | Default TLS private key path |
+| `FASTSYNC_TLS_CA` | — | Default TLS CA certificate path |
+
+## Implementation Details
+
+### Data Structures
+1. **Chunk** — collection of files (~10 MB total by default)
+2. **File** — path, content (`Data`), optional `FileMetadata` pointer
+3. **FileMetadata** — `mode`, `uid`, `gid`, `mtime_sec`, `mtime_nsec`;
+uid / gid are advisory wire fields and are never applied by the receiver;
+atime is unsupported
+4. **Config** — runtime parameters (transported over wire, TLS settings excluded). Includes `timeout`, `contimeout`, `quiet`, `backup`, `backup_dir`, `stats`, `max_depth`, `log_file`, `queue_size`.
+5. **Queue** — thread-safe bounded queue with condition variables
+6. **DirectoryScanner** — recursive BFS traversal with exclude and include pattern support, max-depth enforcement
+
+### Key Algorithms
+1. **File scanning** — BFS directory traversal;
+entries matched against exclude and include patterns,
+    max - depth enforced 2. * *Chunking ** — files accumulated until `chunk_size` threshold,
+    then flushed 3. *
+            *Compression ** — streaming zstd
+                via `ZSTD_compressStream2` / `ZSTD_decompressStream` 4. *
+            *Network protocol ** — status -
+        code - driven exchange with metadata packing,
+    keep - alive,
+    and abort support 5. * *Incremental check ** — client sends `STATUS_CHECK` + path + size +
+        mtime and,
+    with `--checksum`, XXH64 content checksum; server compares against destination. Can be batched via `STATUS_CHECK_BATCH` for reduced round-trips.
+6. **Bandwidth limiting** — token-bucket algorithm with `nanosleep` throttling on 64 KB write chunks
+7. **Metadata restoration** — `chmod()`, `chown()`, `utimensat()` on the receiving side
+8. **`--delete`** — sender tracks all sent paths;
+receiver walks destination tree and removes unlisted files / directories 9. *
+        *SSH transport *
+            * — `socketpair()` + `fork()` + `execvp("ssh",
+                                                    ...)` with `ControlMaster` and port support
+                                                10. *
+                                                *TLS transport ** — OpenSSL `SSL_CTX` with TLS
+                                                1.2 minimum,
+    mutual CA verification,
+    transparent `SSL_read`/`SSL_write` via `io_set_ssl()` 11. *
+        *Path traversal protection ** — `has_path_traversal()` rejects any file path
+         containing `..` components,
+    preventing directory escape attacks 12. *
+            *Connection limiting ** — server tracks active connections and rejects
+            new ones beyond `max_connections` (default 100)13. *
+            *Keep
+        - alive ** — idle connections receive periodic `STATUS_KEEPALIVE` to detect half
+        - open TCP connections 14. * *Abort handling ** — `SIGINT` sets an abort flag; the next protocol operation sends `STATUS_ABORT` for clean server cleanup
+15. **Atomic writes** — files are written to a `.tmp` suffix then atomically renamed via `rename()`, preventing partial files
+16. **Backup** — before overwriting, existing files are moved to `--backup-dir` (or same directory with `~` suffix) preserving the original
+
+## Security Features
+
+### Path Traversal Protection
+All received file paths are validated by `has_path_traversal()` before any disk operation. Any path containing `..` components is rejected with `STATUS_ERROR`, preventing directory escape attacks.
+
+### TLS Certificate Verification
+TLS requires `--ca` and performs mutual TLS verification (`SSL_VERIFY_PEER` with depth 4). Connections without certificate verification are rejected.
+
+### Connection Limits
+The server enforces a maximum of 100 concurrent connections (configurable via `max_connections` in `Server`). When the limit is reached, new connections are immediately rejected and closed.
+
+### Abort Handling
+If the client receives `SIGINT` (Ctrl+C) during a transfer, it sends `STATUS_ABORT` to the server. The server then cleans up temporary files and exits the child process, preventing incomplete files from remaining on disk.
+
+### Atomic Writes
+Received files are written to a temporary path (suffixed with `.tmp`) and then atomically renamed to the final filename via `rename()`. This prevents partial or corrupted files from appearing at the destination if the transfer is interrupted.
+
+## Build Requirements
+
+- C11 compiler
+- CMake >= 3.22
+- zstd library
+- OpenSSL (development headers and libraries)
+- pthreads
+- SSH client (for SSH transport mode only)
+
+### Installing Dependencies
+
+**Ubuntu/Debian:**
+```bash
+sudo apt install cmake build-essential libzstd-dev libssl-dev openssh-client
+```
+
+**Nix:**
+```bash
+nix-shell  # provides zstd, openssl, cmake, gcc
+```
+
+## Building

 ```bash
 cmake -B build -S .
@@ -98,10 +256,13 @@ cmake --build build -j$(nproc)
 ### SSH transfer

 The remote host must have `fastsync-server` available in `PATH`, or use
-`--fastsync-server-path`.
+`--fastsync-server-path`. SSH starts `fastsync-server --stdio` in its remote
+working directory, so use a destination below that directory unless the
+remote server is otherwise configured with a matching authorized root.

 ```bash
-./build/client /path/to/source user@host:/path/to/destination
+ssh user@host 'mkdir -p destination'
+./build/client /path/to/source user@host:destination
 ```

 ### TCP transfer
@@ -109,7 +270,7 @@ The remote host must have `fastsync-server` available in `PATH`, or use
 Start the FastSync server:

 ```bash
-./build/server -p 8080
+./build/server --destination-root /path/to -p 8080
 ```

 Then run the client:
@@ -120,10 +281,12 @@ Then run the client:
  --save-to-disk
 ```

-### TLS transfer
+Plain TCP requires the explicit `--allow-unauthenticated` server option. Use TLS for
+authenticated network connections.

+### TLS transfer
 ```bash
-./build/server --tls --cert server.pem --key server-key.pem -p 8443
+./build/server --destination-root /path/to --tls --cert server.pem --key server-key.pem -p 8443
 ./build/client --tls --cert client.pem --key client-key.pem --ca ca.pem \
  --server-host example.com --server-port 8443 \
  --source-dir /path/to/source --dest-dir /path/to/destination \
@@ -136,34 +299,34 @@ These examples show the intended rsync-style workflow. Options marked as
 FastSync-native are optional performance or transport extensions.

 ```bash
-# Basic synchronization
+#Basic synchronization
 ./build/client /source/ /destination/

-# Archive-style synchronization (current FastSync archive behavior)
-./build/client -a /source/ user@host:/destination/
+#Archive - style synchronization(current FastSync archive behavior)
+./build/client -a /source/ user@host:destination/

-# Preview a transfer without changing the destination
+#Preview a transfer without changing the destination
 ./build/client -n /source/ /destination/

-# Exclude temporary and object files
+#Exclude temporary and object files
 ./build/client --exclude '*.tmp' --exclude '*.o' \
-  /source/ user@host:/destination/
+  /source/ user@host:destination/

-# Remove destination entries not present in the source
-./build/client --delete /source/ user@host:/destination/
+#Remove destination entries not present in the source
+./build/client --delete /source/ user@host:destination/

-# Skip unchanged files using size and modification time
-./build/client --incremental /source/ user@host:/destination/
+#Skip unchanged files using size and modification time
+./build/client --incremental /source/ user@host:destination/

-# Verify content when size and time are not sufficient
-./build/client --incremental --checksum /source/ user@host:/destination/
+#Verify content when size and time are not sufficient
+./build/client --incremental --checksum /source/ user@host:destination/

-# Preserve supported mode and timestamp metadata
-./build/client -M /source/ user@host:/destination/
+#Preserve supported mode and timestamp metadata
+./build/client -M /source/ user@host:destination/

-# Keep backups of overwritten destination files
-./build/client --backup --backup-dir /backups \
-  /source/ user@host:/destination/
+#Keep backups of overwritten destination files
+./build/client --backup --backup-dir backups \
+  /source/ user@host:destination/
 ```

 ## FastSync Extensions
@@ -211,14 +374,16 @@ before FastSync can claim full rsync CLI compatibility.
 | `--include-from <file>` | Read include patterns from a file. |
 | `--max-size <bytes>` | Skip files larger than the limit. |
 | `--min-size <bytes>` | Skip files smaller than the limit. |
-| `--max-depth <n>` | Limit recursive scanning depth; zero means unlimited. |
-| `--incremental` | Skip files matching destination size and mtime. |
-| `--checksum` | Include xxHash64 content checks in incremental comparisons. |
-| `--backup` | Back up overwritten files. |
-| `--backup-dir <dir>` | Store backups under a separate directory. |
-| `--suffix <suffix>` | Set the backup filename suffix. |
-| `--partial` | Keep received data under the configured partial location. Full resume semantics are planned. |
-| `--partial-dir <dir>` | Set the partial transfer directory. |
+| `--max-depth <n>` | Limit recursive scanning depth;
+zero means unlimited.| | `--incremental` | Skip files matching destination size and mtime.|
+    | `--checksum` | Include xxHash64 content checks in incremental comparisons.| | `--backup` |
+    Back up overwritten files.| | `--backup - dir<dir>` | Store backups under a separate directory.|
+    | `--suffix<suffix>` | Set the backup filename suffix.| | `--partial` |
+    Select partial - transfer handling.With `--partial - dir`,
+    completed files are written there;
+resumable transfers are not implemented.| | `--partial - dir<dir>` |
+    Set a relative partial - transfer directory below the server destination root;
+use with `--partial`. |
 | `--inplace` | Write directly to the destination instead of using a temporary file. |

 ### Metadata and links
@@ -226,7 +391,8 @@ before FastSync can claim full rsync CLI compatibility.
 | Option | Description |
 |---|---|
 | `-M`, `--preserve` | Preserve supported file metadata, currently mode and modification time. |
-| `-l`, `--links` | Request symlink preservation; link-target transfer remains incomplete. |
+| `-l`, `--links` | Request symlink preservation;
+link-target transfer remains incomplete. |
 | `--copy-links` | Copy symlink referents. |
 | `--safe-links` | Skip symlinks that point outside the transfer tree. |
 | `--copy-unsafe-links` | Copy unsafe symlink referents. |
@@ -269,6 +435,8 @@ before FastSync can claim full rsync CLI compatibility.
 | `--cert <path>` | TLS certificate file. |
 | `--key <path>` | TLS private key file. |
 | `--ca <path>` | CA file for peer verification. |
+| `--destination-root <path>` | Confine received files to this server-side root;
+defaults to the current directory. |
 | `--allow-delete` | Permit client delete manifests. Deletion is refused by default. |
 | `-v`, `--verbose` | Enable debug logging. |
 | `--help` | Print server usage. |
@@ -152,14 +152,14 @@ This document maps rsync's full feature set to FastSync's current implementation

 | Flag | Rsync Description | FastSync Status | Notes |
 |------|-------------------|-----------------|-------|
-| `-S`, `--sparse` | Sparse block handling | ✅ Implemented | `preserve_sparse` config field |
+| `-S`, `--sparse` | Sparse block handling | ⚠️ Partial | Flag is accepted, but full hole preservation is not implemented |
 | `--preallocate` | Allocate dest files before writing | ❌ Not Implemented | |

 ## 11. Checksum & Comparison

 | Flag | Rsync Description | FastSync Status | Notes |
 |------|-------------------|-----------------|-------|
-| `--checksum` | Skip based on checksum | ❌ Not Implemented | Removed because it had no effect; `-c` means compression |
+| `--checksum` | Skip based on checksum | ✅ Implemented | With `--incremental`, compares xxHash64 content checksums; `-c` remains compression |
 | `--checksum-choice=STR` | Choose checksum algorithm | ❌ Not Implemented | xxHash used internally |
 | `--compare-dest=DIR` | Compare dest files relative to DIR | ❌ Not Implemented | Removed because it had no effect |
 | `--copy-dest=DIR` | Include copies of unchanged files | ❌ Not Implemented | Removed because it had no effect |
@@ -411,7 +411,6 @@ int main(int argc, char* argv[]) {
  parse_environment(&env_source, &env_dest, &save_to_disk);

  int exit_code = 0;
-  bool config_owned_by_pipeline = false;
  Config* config = config_create();
  if (!config) {
    fprintf(stderr, "Error: failed to allocate config\n");
@@ -494,18 +493,14 @@ int main(int argc, char* argv[]) {

  /* Execute transfer */
  if (config->use_multithreading) {
-    config_owned_by_pipeline = true;
-    exit_code = send_files_multithreaded(config);
+    exit_code = send_files_multithreaded(&config);
  } else {
    exit_code = send_files(config);
  }

 cleanup:
  if (config) {
-    if (config->log_file)
-      fclose(config->log_file);
-    if (!config_owned_by_pipeline)
-      config_delete(config);
+    config_delete(config);
  }
  return exit_code;
 }
@@ -60,6 +60,7 @@ static Client* connect_transfer_client(const Config* config) {
    connected = client_connect(client, config->server_host, config->server_port);
  }
  if (!connected) {
+    client_disconnect(client);
    client_delete(client);
    return NULL;
  }
@@ -149,6 +150,8 @@ static int send_dry_run_manifest(const Config* config) {

 /* Send the delete manifest (list of files) to the server. Returns 0 on success, -1 on failure. */
 static int send_delete_manifest(int fd, ArrayList* manifest) {
+  if (!manifest)
+    return -1;
  if (!send_status(fd, STATUS_MANIFEST))
    return -1;
  if (!send_int(fd, manifest->size))
@@ -190,17 +193,22 @@ static int incremental_check(Client* client, File* file, const Config* config,
    return 1;
  if (s == STATUS_DELTA_SIGNATURE) {
    Data* sig_data = receive_data(client->file_descriptor);
-    if (!sig_data)
+    if (!sig_data) {
+      send_status(client->file_descriptor, STATUS_ERROR);
      return -1;
+    }
    DeltaSignature* sig = delta_signature_deserialize(sig_data);
    data_destroy(sig_data);
-    if (!sig)
+    if (!sig) {
+      send_status(client->file_descriptor, STATUS_ERROR);
      return -1;
+    }
    *out_sig = sig;
    return 2;
  }
  if (s != STATUS_NEXT) {
    log_message(LOG_LEVEL_ERROR, "Unexpected server status");
+    send_status(client->file_descriptor, STATUS_ERROR);
    return -1;
  }
  return 0;
@@ -208,8 +216,10 @@ static int incremental_check(Client* client, File* file, const Config* config,

 static int send_delta(Client* client, File* file, DeltaSignature* sig, Config* config) {
  Delta* delta = delta_compute(file->data->data, file->data->size, sig, config->delta_block_size);
+  /* The receiver is blocked after sending the signature.  Every local
+     fallback therefore needs the explicit NEXT response before full data. */
  if (!delta)
-    return 1;
+    return send_status(client->file_descriptor, STATUS_NEXT) ? 1 : -1;

  if (!delta_is_worthwhile(delta, file->data->size)) {
    delta_destroy(delta);
@@ -221,14 +231,14 @@ static int send_delta(Client* client, File* file, DeltaSignature* sig, Config* c
  Data* delta_data = delta_serialize(delta);
  delta_destroy(delta);
  if (!delta_data)
-    return -1;
+    return send_status(client->file_descriptor, STATUS_NEXT) ? 1 : -1;

  Data* to_send = delta_data;
  if (config->use_compression) {
    to_send = data_compress(delta_data, config->compression_level);
    data_destroy(delta_data);
    if (!to_send)
-      return -1;
+      return send_status(client->file_descriptor, STATUS_NEXT) ? 1 : -1;
  }

  bool ok = send_status(client->file_descriptor, STATUS_DELTA_DATA) &&
@@ -358,7 +368,8 @@ int send_chunk(Client* client, Chunk* chunk, Config* config) {
    if (f == NULL)
      continue;
    bool stream = f->data->data == NULL && f->data->size > 0;
-    bool use_sendfile = (config->use_sendfile && !config->use_compression) || stream;
+    bool use_sendfile =
+        (config->use_sendfile && !config->use_compression) || (stream && !config->use_compression);
    int rc = send_single_file(client, f, config, config->use_incremental, use_sendfile);
    if (rc == 1)
      continue;
@@ -375,6 +386,7 @@ static int send_chunks_multithreaded(void* pipeline_context) {
    if (context->config->transport == TRANSPORT_TCP)
      fprintf(stderr, "Error: could not connect to server%s\n",
              context->config->use_tls ? " via TLS" : "");
+    pipeline_cancel(context);
    mark_sender_done(context);
    return thrd_error;
  }
@@ -383,6 +395,7 @@ static int send_chunks_multithreaded(void* pipeline_context) {
  protocol_session_set_ssl(&session, (SSL*)client->ssl);
  protocol_session_bind(&session);
  if (!config_send(client->file_descriptor, context->config)) {
+    pipeline_cancel(context);
    disconnect_transfer_client(client);
    mark_sender_done(context);
    protocol_session_unbind();
@@ -394,6 +407,13 @@ static int send_chunks_multithreaded(void* pipeline_context) {
        context->queue_loader, &context->mutex_loader, &context->condition_not_empty_loader,
        &context->condition_not_full_loader, &context->loader_done);
    if (current_chunk == NULL) {
+      if (atomic_load(&context->cancelled)) {
+        pipeline_cancel(context);
+        disconnect_transfer_client(client);
+        mark_sender_done(context);
+        protocol_session_unbind();
+        return thrd_error;
+      }
      if (context->config->use_delete) {
        if (send_delete_manifest(client->file_descriptor, context->manifest) != 0)
          goto send_fail;
@@ -503,12 +523,13 @@ static int load_files_multithreaded(void* pipeline_context) {
    if (!context->config->use_sendfile) {
      for (int i = 0; i < chunk->element_count; i++) {
        File* f = chunk->items[i];
-        if (f->data->size > STREAM_THRESHOLD)
+        if (f->data->size > STREAM_THRESHOLD && !context->config->use_compression)
          continue;
        if (!file_load_data(f)) {
-          log_message(LOG_LEVEL_ERROR, "Failed to load file data, skipping");
-          file_destroy(f);
-          chunk->items[i] = NULL;
+          log_message(LOG_LEVEL_ERROR, "Failed to load file data");
+          chunk_destroy(chunk);
+          pipeline_cancel(context);
+          return thrd_error;
        }
      }
    }
@@ -517,9 +538,7 @@ static int load_files_multithreaded(void* pipeline_context) {
                                            &context->condition_not_full_loader,
                                            &context->cancelled)) {
      chunk_destroy(chunk);
-      atomic_store(&context->cancelled, true);
-      cnd_broadcast(&context->condition_not_full_loader);
-      cnd_broadcast(&context->condition_not_empty_loader);
+      pipeline_cancel(context);
      return thrd_error;
    }
  }
@@ -611,11 +630,12 @@ int send_files(Config* config) {
    if (!config->use_sendfile) {
      for (int i = 0; i < current_chunk->element_count; i++) {
        File* f = current_chunk->items[i];
-        if (f->data->size > STREAM_THRESHOLD)
+        if (f->data->size > STREAM_THRESHOLD && !config->use_compression)
          continue;
        if (!file_load_data(f)) {
          log_message(LOG_LEVEL_ERROR, "Failed to load file data");
-          continue;
+          chunk_destroy(current_chunk);
+          goto send_fail;
        }
      }
    }
@@ -645,6 +665,7 @@ int send_files(Config* config) {
  if (config->use_delete) {
    if (send_delete_manifest(client->file_descriptor, manifest) != 0) {
      array_list_delete(manifest);
+      manifest = NULL;
      goto send_fail;
    }
    array_list_delete(manifest);
@@ -675,7 +696,10 @@ send_fail:
  return 1;
 }

-int send_files_multithreaded(Config* config) {
+int send_files_multithreaded(Config** config_ptr) {
+  if (!config_ptr || !*config_ptr)
+    return 1;
+  Config* config = *config_ptr;
  if (config->dry_run)
    return send_dry_run_manifest(config);

@@ -706,8 +730,13 @@ int send_files_multithreaded(Config* config) {
    queue_destroy(q2);
    return 1;
  }
+  *config_ptr = NULL; /* context now owns config through all remaining paths */
  if (config->use_delete)
-    context->manifest = create_transfer_manifest(config);
+    context->manifest = array_list_create(free);
+  if (config->use_delete && !context->manifest) {
+    pipeline_context_sender_destroy(context);
+    return 1;
+  }

  thrd_t scanner, loader, sender;
  bool scanner_created = false;
@@ -7,6 +7,7 @@

 int send_chunk(Client* client, Chunk* chunk, Config* config);
 int send_files(Config* config);
-int send_files_multithreaded(Config* config);
+/* Takes ownership only when *config is set to NULL on return. */
+int send_files_multithreaded(Config** config);

 #endif
@@ -41,8 +41,8 @@ bool validate_config(const Config* config) {
    return false;
  }
  if (config->use_tls) {
-    if (!config->tls_cert || !config->tls_key) {
-      fprintf(stderr, "Error: --tls requires --cert and --key\n");
+    if (!config->tls_cert || !config->tls_key || !config->tls_ca) {
+      fprintf(stderr, "Error: --tls requires --cert, --key, and --ca\n");
      return false;
    }
  }
@@ -363,6 +363,8 @@ static int parallel_worker_thread(void* arg) {
      cnd_broadcast(&wa->ps->result_not_empty);
      cnd_broadcast(&wa->ps->result_not_full);
      mtx_unlock(&wa->ps->result_mutex);
+      for (int j = i; j < wa->dir_count; j++)
+        free(wa->dirs[j]);
      break;
    }
    Chunk* chunk;
@@ -398,6 +400,18 @@ static int parallel_worker_thread(void* arg) {
  return thrd_success;
 }

+static void parallel_scanner_creation_failed(ParallelScanner* ps) {
+  mtx_lock(&ps->result_mutex);
+  ps->failed = true;
+  atomic_store(&ps->cancelled, true);
+  ps->expected_threads = ps->created_threads;
+  if (ps->completed >= ps->expected_threads)
+    ps->done = true;
+  cnd_broadcast(&ps->result_not_empty);
+  cnd_broadcast(&ps->result_not_full);
+  mtx_unlock(&ps->result_mutex);
+}
+
 ParallelScanner* parallel_scanner_create_with_options(const char* root_directory,
                                                      const ScannerOptions* options) {
  if (!root_directory || !options)
@@ -520,7 +534,6 @@ ParallelScanner* parallel_scanner_create_with_options(const char* root_directory
        void** items = array_list_to_array(batch);
        if (!items) {
          ps->failed = true;
-          batch->item_destroyer = file_destroy;
          array_list_delete(batch);
          batch = NULL;
          break;
@@ -529,11 +542,13 @@ ParallelScanner* parallel_scanner_create_with_options(const char* root_directory
        free(items);
        if (!c) {
          ps->failed = true;
-          batch->item_destroyer = file_destroy;
          array_list_delete(batch);
          batch = NULL;
          break;
        }
+        int batch_start = i - batch->size + 1;
+        for (int j = batch_start; j <= i; j++)
+          root_files->items[j] = NULL;
        batch->item_destroyer = NULL;
        array_list_delete(batch);
        batch = NULL;
@@ -560,7 +575,6 @@ ParallelScanner* parallel_scanner_create_with_options(const char* root_directory
      array_list_delete(batch);
    }
    ps->initial_chunk = first;
-    root_files->item_destroyer = NULL;
  }
  array_list_delete(root_files);

@@ -587,14 +601,14 @@ ParallelScanner* parallel_scanner_create_with_options(const char* root_directory
        break;
      ParallelWorkerArg* wa = calloc(1, sizeof(ParallelWorkerArg));
      if (!wa) {
-        ps->failed = true;
+        parallel_scanner_creation_failed(ps);
        break;
      }
      wa->ps = ps;
      wa->dirs = calloc(count, sizeof(char*));
      if (!wa->dirs) {
        free(wa);
-        ps->failed = true;
+        parallel_scanner_creation_failed(ps);
        break;
      }
      bool dup_ok = true;
@@ -608,7 +622,7 @@ ParallelScanner* parallel_scanner_create_with_options(const char* root_directory
          free(wa->dirs[j]);
        free(wa->dirs);
        free(wa);
-        ps->failed = true;
+        parallel_scanner_creation_failed(ps);
        break;
      }
      wa->dir_count = count;
@@ -620,13 +634,7 @@ ParallelScanner* parallel_scanner_create_with_options(const char* root_directory
          free(wa->dirs[j]);
        free(wa->dirs);
        free(wa);
-        ps->failed = true;
-        atomic_store(&ps->cancelled, true);
-        ps->expected_threads = ps->created_threads;
-        mtx_lock(&ps->result_mutex);
-        cnd_broadcast(&ps->result_not_empty);
-        cnd_broadcast(&ps->result_not_full);
-        mtx_unlock(&ps->result_mutex);
+        parallel_scanner_creation_failed(ps);
        break;
      }
      ps->num_threads++;
@@ -659,6 +667,11 @@ Chunk* parallel_scanner_next(ParallelScanner* ps) {
  }
  if (ps->num_threads == 0) {
    mtx_lock(&ps->result_mutex);
+    if (!queue_is_empty(ps->result_queue)) {
+      Chunk* chunk = queue_dequeue(ps->result_queue);
+      mtx_unlock(&ps->result_mutex);
+      return chunk;
+    }
    ps->done = true;
    mtx_unlock(&ps->result_mutex);
    return NULL;
@@ -8,8 +8,14 @@
 #include <sys/stat.h>

 static bool receiver_process_chunk(Chunk* chunk, const ReceiverSink* sink) {
+  if (!chunk || !sink || !sink->store_file)
+    return false;
  for (int i = 0; i < chunk->element_count; i++) {
    File* file = chunk->items[i];
+    if (!file) {
+      chunk_destroy(chunk);
+      return false;
+    }
    chunk->items[i] = NULL;
    if (!sink->store_file(file, sink->context)) {
      chunk_destroy(chunk);
@@ -41,9 +47,19 @@ static bool receiver_process_batch(Config* config, int file_descriptor) {
      send_status(file_descriptor, STATUS_ERROR);
      return false;
    }
+    if (check_size > MAX_RECEIVE_FILE_SIZE) {
+      free(check_path);
+      send_status(file_descriptor, STATUS_ERROR);
+      return false;
+    }
    char* full_path = path_cat(config->receive_root_directory, check_path);
+    if (!full_path) {
+      free(check_path);
+      send_status(file_descriptor, STATUS_ERROR);
+      return false;
+    }
    struct stat st;
-    bool has_old = full_path && lstat(full_path, &st) == 0;
+    bool has_old = file_stat_secure(full_path, &st);
    bool match = has_old && (unsigned long long)st.st_size == check_size &&
                 (long long)st.st_mtime == check_mtime;
    bool sent = send_status(file_descriptor, match ? STATUS_OK : STATUS_NEXT);
@@ -1,4 +1,5 @@
 #include "config.h"
+#include "chunk.h"
 #include "file.h"
 #include "log.h"
 #include "multiprocessing.h"
@@ -15,10 +16,41 @@
 #include <stdlib.h>
 #include <string.h>
 #include <sys/stat.h>
+#include <unistd.h>
+#include <openssl/x509.h>

 static char* authorized_root;
 static int authorized_root_fd = -1;
 static bool allow_delete;
+static bool allow_unauthenticated;
+static const char* required_client_cn;
+
+static bool tls_client_identity_allowed(SSL* ssl) {
+  if (!ssl || !required_client_cn)
+    return false;
+  X509* certificate = SSL_get1_peer_certificate(ssl);
+  if (!certificate)
+    return false;
+  char common_name[256];
+  int length = X509_NAME_get_text_by_NID(X509_get_subject_name(certificate), NID_commonName,
+                                         common_name, sizeof(common_name));
+  size_t required_length = strlen(required_client_cn);
+  bool allowed = length >= 0 && (size_t)length == required_length &&
+                 required_length < sizeof(common_name) &&
+                 memcmp(common_name, required_client_cn, required_length) == 0;
+  X509_free(certificate);
+  return allowed;
+}
+
+static void release_authorization(void) {
+  file_set_authorized_root(-1, NULL);
+  utils_set_authorized_root_fd(-1);
+  if (authorized_root_fd >= 0)
+    close(authorized_root_fd);
+  authorized_root_fd = -1;
+  free(authorized_root);
+  authorized_root = NULL;
+}

 static bool path_is_within(const char* root, const char* path) {
  size_t n = strlen(root);
@@ -27,22 +59,166 @@ static bool path_is_within(const char* root, const char* path) {

 static bool __attribute__((unused)) configure_authorization(const char* root) {
  char resolved[PATH_MAX];
-  if (!root || !realpath(root, resolved))
+  if (!root) {
+    file_set_authorized_root(-1, NULL);
+    utils_set_authorized_root(-1, NULL);
    return false;
+  }
+  int root_fd = open(root, O_RDONLY | O_DIRECTORY | O_NOFOLLOW | O_CLOEXEC);
+  if (root_fd < 0) {
+    file_set_authorized_root(-1, NULL);
+    utils_set_authorized_root(-1, NULL);
+    return false;
+  }
+  char fd_path[64];
+  int fd_path_length = snprintf(fd_path, sizeof(fd_path), "/proc/self/fd/%d", root_fd);
+  if (fd_path_length < 0 || (size_t)fd_path_length >= sizeof(fd_path) ||
+      !realpath(fd_path, resolved)) {
+    close(root_fd);
+    file_set_authorized_root(-1, NULL);
+    utils_set_authorized_root(-1, NULL);
+    return false;
+  }
  authorized_root = str_dup(resolved);
-  if (!authorized_root)
+  if (!authorized_root) {
+    close(root_fd);
+    file_set_authorized_root(-1, NULL);
+    utils_set_authorized_root(-1, NULL);
    return false;
-  authorized_root_fd = open(resolved, O_RDONLY | O_DIRECTORY | O_CLOEXEC);
-  if (authorized_root_fd < 0) {
+  }
+  authorized_root_fd = root_fd;
+  if (!file_set_authorized_root(authorized_root_fd, authorized_root) ||
+      !utils_set_authorized_root(authorized_root_fd, authorized_root)) {
+    file_set_authorized_root(-1, NULL);
+    utils_set_authorized_root(-1, NULL);
+    close(authorized_root_fd);
+    authorized_root_fd = -1;
    free(authorized_root);
    authorized_root = NULL;
    return false;
  }
-  file_set_authorized_root(authorized_root_fd, authorized_root);
-  utils_set_authorized_root_fd(authorized_root_fd);
  return true;
 }

+int receive_files(Config* config, int fd) {
+  Status status;
+  if (!receive_status(fd, &status))
+    return -1;
+
+  while (status == STATUS_NEXT || status == STATUS_CHUNK || status == STATUS_CHECK ||
+         status == STATUS_KEEPALIVE || status == STATUS_ABORT || status == STATUS_CHECK_BATCH) {
+    if (status == STATUS_KEEPALIVE) {
+      send_status(fd, STATUS_KEEPALIVE);
+      goto next;
+    }
+    if (status == STATUS_ABORT) {
+      log_message(LOG_LEVEL_INFO, "Received abort from client, cleaning up");
+      return -1;
+    }
+    if (status == STATUS_CHECK) {
+      bool skipped;
+      File* file = receive_incremental_check(fd, config, &skipped);
+      if (skipped)
+        goto next;
+      if (file == NULL && !skipped)
+        return -1;
+      if (config->save_to_disk &&
+          !file_save_to_disk(config->receive_root_directory, file, config)) {
+        file_destroy(file);
+        send_status(fd, STATUS_ERROR);
+        return -1;
+      }
+      file_destroy(file);
+    } else if (status == STATUS_CHUNK) {
+      Chunk* chunk = receive_chunk_data(fd, config);
+      if (chunk == NULL) {
+        send_status(fd, STATUS_ERROR);
+        return -1;
+      }
+      for (int i = 0; i < chunk->element_count; i++) {
+        if (config->save_to_disk &&
+            !file_save_to_disk(config->receive_root_directory, chunk->items[i], config)) {
+          chunk_destroy(chunk);
+          send_status(fd, STATUS_ERROR);
+          return -1;
+        }
+      }
+      chunk_destroy(chunk);
+    } else if (status == STATUS_CHECK_BATCH) {
+      int count;
+      /* Batch framing has no checksum field yet; never silently downgrade a
+         checksum-enabled transfer into mtime-only matching. */
+      if (config->checksum || !receive_int(fd, &count) || count < 0 || count > MAX_MANIFEST_ENTRIES)
+        return -1;
+      for (int i = 0; i < count; i++) {
+        char* check_path = receive_str(fd);
+        if (!check_path)
+          return -1;
+        unsigned long long check_size;
+        long long check_mtime;
+        if (!receive_n_data(fd, &check_size, sizeof(check_size)) ||
+            !receive_n_data(fd, &check_mtime, sizeof(check_mtime))) {
+          free(check_path);
+          return -1;
+        }
+        if (!utils_valid_batch_path(check_path)) {
+          free(check_path);
+          send_status(fd, STATUS_ERROR);
+          return -1;
+        }
+        struct stat st;
+        char* full_path = path_cat(config->receive_root_directory, check_path);
+        if (!full_path) {
+          free(check_path);
+          send_status(fd, STATUS_ERROR);
+          return -1;
+        }
+        bool has_old = full_path && file_stat_secure(full_path, &st);
+        bool match = has_old && (unsigned long long)st.st_size == check_size &&
+                     (long long)st.st_mtime == check_mtime;
+        bool sent = send_status(fd, match ? STATUS_OK : STATUS_NEXT);
+        free(full_path);
+        free(check_path);
+        if (!sent)
+          return -1;
+      }
+      goto next;
+    } else {
+      File* file = file_receive(config, fd);
+      if (file == NULL) {
+        log_message(LOG_LEVEL_ERROR, "Failed to receive file");
+        send_status(fd, STATUS_ERROR);
+        return -1;
+      }
+      if (config->save_to_disk &&
+          !file_save_to_disk(config->receive_root_directory, file, config)) {
+        file_destroy(file);
+        send_status(fd, STATUS_ERROR);
+        return -1;
+      }
+      file_destroy(file);
+    }
+  next:
+    if (!receive_status(fd, &status)) {
+      send_status(fd, STATUS_ERROR);
+      return -1;
+    }
+  }
+
+  if (status == STATUS_MANIFEST) {
+    if (receive_manifest(fd, config, &status) != 0) {
+      return -1;
+    }
+  }
+  if (status != STATUS_FINISHED) {
+    log_message(LOG_LEVEL_ERROR, "Did not receive FINISHED Status");
+    send_status(fd, STATUS_ERROR);
+    return -1;
+  }
+  send_status(fd, STATUS_OK);
+  return 0;
+}
+
 void handler(int file_descriptor) {
  SSL* ssl = io_get_ssl();
  ProtocolSession session;
@@ -63,25 +239,37 @@ void handler(int file_descriptor) {
    protocol_session_unbind();
    return;
  }
-  char resolved_destination[PATH_MAX];
-  char* canonical_destination = realpath(config->receive_root_directory, NULL);
-  const char* destination =
-      canonical_destination ? canonical_destination : config->receive_root_directory;
-  if (has_path_traversal(destination) || !path_is_within(authorized_root, destination)) {
-    log_message(LOG_LEVEL_ERROR, "Rejected destination outside authorized root");
-    free(canonical_destination);
+  if (!allow_unauthenticated && ssl == NULL) {
+    log_message(LOG_LEVEL_ERROR, "Rejected unauthenticated plaintext connection");
    config_delete(config);
    close(file_descriptor);
    protocol_session_unbind();
    return;
  }
-  if (canonical_destination)
-    snprintf(resolved_destination, sizeof(resolved_destination), "%s", canonical_destination);
-  else
-    snprintf(resolved_destination, sizeof(resolved_destination), "%s", destination);
-  free(canonical_destination);
-  free(config->receive_root_directory);
-  config->receive_root_directory = str_dup(resolved_destination);
+  if (ssl && required_client_cn && !tls_client_identity_allowed(ssl)) {
+    log_message(LOG_LEVEL_ERROR, "Rejected TLS client with unauthorized identity");
+    config_delete(config);
+    close(file_descriptor);
+    return;
+  }
+  char* destination = config->receive_root_directory;
+  char* joined_destination = NULL;
+  if (destination && destination[0] != '/')
+    joined_destination = path_cat(authorized_root, destination);
+  if (joined_destination)
+    destination = joined_destination;
+  if (!destination || has_path_traversal(destination) ||
+      !path_is_within(authorized_root, destination)) {
+    log_message(LOG_LEVEL_ERROR, "Rejected destination outside authorized root");
+    free(joined_destination);
+    config_delete(config);
+    close(file_descriptor);
+    return;
+  }
+  if (joined_destination) {
+    free(config->receive_root_directory);
+    config->receive_root_directory = joined_destination;
+  }
  if (!config->receive_root_directory) {
    config_delete(config);
    close(file_descriptor);
@@ -106,6 +294,7 @@ void handler(int file_descriptor) {
      protocol_session_unbind();
      return;
    }
+    context->session.total_allocated_bytes = session.total_allocated_bytes;
    thrd_t receiver, writer;
    bool receiver_created = thrd_create(&receiver, receive_thread, context) == thrd_success;
    bool writer_created = false;
@@ -167,8 +356,10 @@ static void print_server_usage(void) {
  printf("  --cert <path>       TLS certificate file (PEM)\n");
  printf("  --key <path>        TLS private key file (PEM)\n");
  printf("  --ca <path>         TLS CA certificate file (PEM)\n");
+  printf("  --client-cn <name>  Required TLS client certificate CN\n");
  printf("  --destination-root <path>  Authorized destination root (default: .)\n");
  printf("  --allow-delete      Permit manifest deletion\n");
+  printf("  --allow-unauthenticated  Allow plaintext/anonymous network clients\n");
  printf("  -v, --verbose       Enable debug logging\n");
  printf("  --help              Show this help\n");
 }
@@ -197,10 +388,14 @@ int main(int argc, char* argv[]) {
      tls_key = argv[++i];
    } else if (strcmp(argv[i], "--ca") == 0 && i + 1 < argc) {
      tls_ca = argv[++i];
+    } else if (strcmp(argv[i], "--client-cn") == 0 && i + 1 < argc) {
+      required_client_cn = argv[++i];
    } else if (strcmp(argv[i], "--destination-root") == 0 && i + 1 < argc) {
      destination_root = argv[++i];
    } else if (strcmp(argv[i], "--allow-delete") == 0) {
      allow_delete = true;
+    } else if (strcmp(argv[i], "--allow-unauthenticated") == 0) {
+      allow_unauthenticated = true;
    } else if (strcmp(argv[i], "-p") == 0 && i + 1 < argc) {
      char* end;
      long p = strtol(argv[++i], &end, 10);
@@ -224,35 +419,39 @@ int main(int argc, char* argv[]) {
    return 1;
  }
  if (stdio_mode) {
+    /* SSH authenticates the stdio transport outside of FastSync. */
+    allow_unauthenticated = true;
    io_set_fds(STDIN_FILENO, STDOUT_FILENO);
    handler(STDIN_FILENO);
-    file_set_authorized_root(-1, NULL);
-    utils_set_authorized_root_fd(-1);
-    close(authorized_root_fd);
-    free(authorized_root);
+    release_authorization();
    return 0;
  }
  g_server = server_create(port);
  if (!g_server) {
    log_message(LOG_LEVEL_ERROR, "Failed to create server");
+    release_authorization();
    return 1;
  }
  if (use_tls) {
-    if (!tls_cert || !tls_key) {
-      fprintf(stderr, "Error: --tls requires --cert and --key\n");
+    if (!tls_cert || !tls_key || !tls_ca || !required_client_cn) {
+      fprintf(stderr, "Error: --tls requires --cert, --key, --ca, and --client-cn\n");
      server_delete(&g_server);
+      release_authorization();
      return 1;
    }
    tls_global_init();
    if (!server_create_tls(g_server, tls_cert, tls_key, tls_ca)) {
      log_message(LOG_LEVEL_ERROR, "Failed to set up TLS");
      server_delete(&g_server);
+      release_authorization();
      return 1;
    }
    server_listen_tls(g_server, handler);
  } else {
    server_listen(g_server, handler);
  }
+  server_delete(&g_server);
+  release_authorization();
  return 0;
 }
 #endif
@@ -1,4 +1,6 @@
 #include <stddef.h>
+#include <stdint.h>
+#include <limits.h>
 #include <stdio.h>
 #include <stdlib.h>
 #include <string.h>
@@ -11,21 +13,33 @@
 #include "log.h"
 #include "metadata.h"
 #include "protocol.h"
+#include "utils.h"

 /* Maximum individual file data size within a chunk (64 MB) */
 #define MAX_FILE_DATA_SIZE (64ULL * 1024 * 1024)
+#define MAX_FILES_PER_CHUNK 65536U

 Chunk* chunk_create(File** items, int element_count) {
+  if (element_count < 0 || (element_count > 0 && items == NULL))
+    return NULL;
  Chunk* chunk = (Chunk*)malloc(sizeof(Chunk));
  if (chunk == NULL) {
    perror("ERROR: Could not allocate memory for chunk structure");
    return NULL;
  }

-  chunk->items = (File**)malloc(element_count * sizeof(File*));
-  if (chunk->items == NULL) {
-    free(chunk);
-    return NULL;
+  if (element_count == 0) {
+    chunk->items = NULL;
+  } else {
+    if ((size_t)element_count > SIZE_MAX / sizeof(File*)) {
+      free(chunk);
+      return NULL;
+    }
+    chunk->items = (File**)malloc((size_t)element_count * sizeof(File*));
+    if (chunk->items == NULL) {
+      free(chunk);
+      return NULL;
+    }
  }

  for (int i = 0; i < element_count; i++) {
@@ -50,15 +64,37 @@ void chunk_destroy(void* item) {
 }

 static unsigned long long per_file_serialize_size(File* file, bool use_metadata) {
-  return sizeof(size_t) + strlen(file->path) +
-         (use_metadata ? sizeof(int) + (file->metadata ? FILE_METADATA_WIRE_SIZE : 0) : 0) +
-         sizeof(size_t) + file->data->size;
+  unsigned long long size = sizeof(size_t);
+  size_t path_len = strlen(file->path);
+  unsigned long long metadata_size =
+      use_metadata ? sizeof(int) + (file->metadata ? FILE_METADATA_WIRE_SIZE : 0) : 0;
+  if ((unsigned long long)path_len > ULLONG_MAX - size)
+    return 0;
+  size += path_len;
+  if (metadata_size > ULLONG_MAX - size)
+    return 0;
+  size += metadata_size;
+  if (sizeof(size_t) > ULLONG_MAX - size)
+    return 0;
+  size += sizeof(size_t);
+  if ((unsigned long long)file->data->size > ULLONG_MAX - size)
+    return 0;
+  return size + file->data->size;
 }

 Data* chunk_serialize(Chunk* chunk, bool use_metadata) {
+  if (!chunk || chunk->element_count < 0 || (chunk->element_count > 0 && chunk->items == NULL))
+    return NULL;
  unsigned long long data_size = 0;
  for (int i = 0; i < chunk->element_count; i++) {
-    data_size += per_file_serialize_size(chunk->items[i], use_metadata);
+    if (!chunk->items[i] || !chunk->items[i]->path || !chunk->items[i]->data ||
+        (chunk->items[i]->data->size > 0 && !chunk->items[i]->data->data) ||
+        chunk->items[i]->path[0] == '\0' || has_path_traversal(chunk->items[i]->path))
+      return NULL;
+    unsigned long long file_size = per_file_serialize_size(chunk->items[i], use_metadata);
+    if (file_size == 0 || file_size > ULLONG_MAX - data_size || data_size + file_size > SIZE_MAX)
+      return NULL;
+    data_size += file_size;
  }
  Data* data = data_create_empty(data_size);
  if (data == NULL) {
@@ -87,11 +123,20 @@ Data* chunk_serialize(Chunk* chunk, bool use_metadata) {
 }

 Chunk* chunk_deserialize(Data* data, bool use_metadata) {
+  if (!data || (!data->data && data->size != 0))
+    return NULL;
  ArrayList* files = array_list_create(file_destroy);
+  if (files == NULL)
+    return NULL;
  char* data_pointer = data->data;
  size_t remaining_size = data->size;

  while (remaining_size > 0) {
+    if ((unsigned int)files->size >= MAX_FILES_PER_CHUNK) {
+      log_message(LOG_LEVEL_ERROR, "Chunk contains too many files");
+      array_list_delete(files);
+      return NULL;
+    }
    if (remaining_size < sizeof(size_t)) {
      log_message(LOG_LEVEL_ERROR, "Invalid chunk format: not enough data for path length");
      array_list_delete(files);
@@ -103,12 +148,16 @@ Chunk* chunk_deserialize(Data* data, bool use_metadata) {
    data_pointer += sizeof(size_t);
    remaining_size -= sizeof(size_t);

-    if (remaining_size < path_len) {
+    if (path_len > SIZE_MAX - 1 || remaining_size < path_len) {
      log_message(LOG_LEVEL_ERROR, "Invalid chunk format: not enough data for path");
      array_list_delete(files);
      return NULL;
    }

+    if (path_len == SIZE_MAX) {
+      array_list_delete(files);
+      return NULL;
+    }
    char* path = malloc(path_len + 1);
    if (path == NULL) {
      perror("Could not allocate memory for file path");
@@ -117,34 +166,59 @@ Chunk* chunk_deserialize(Data* data, bool use_metadata) {
    }
    memcpy(path, data_pointer, path_len);
    path[path_len] = '\0';
+    if (memchr(path, '\0', path_len) != NULL) {
+      free(path);
+      array_list_delete(files);
+      return NULL;
+    }
    data_pointer += path_len;
    remaining_size -= path_len;

+    if (path_len == 0 || has_path_traversal(path)) {
+      free(path);
+      array_list_delete(files);
+      return NULL;
+    }
+
    File* file = file_create(path);
    free(path);
+    if (file == NULL) {
+      array_list_delete(files);
+      return NULL;
+    }

    if (use_metadata) {
      if (remaining_size < sizeof(int)) {
        log_message(LOG_LEVEL_ERROR, "Invalid chunk format: not enough data for metadata");
+        file_destroy(file);
        array_list_delete(files);
        return NULL;
      }
      // Peek at present flag to determine total size needed before reading
      int present_flag;
      memcpy(&present_flag, data_pointer, sizeof(int));
-      if (present_flag && remaining_size < sizeof(int) + FILE_METADATA_WIRE_SIZE) {
+      if ((present_flag != 0 && present_flag != 1) ||
+          (present_flag == 1 && remaining_size < sizeof(int) + FILE_METADATA_WIRE_SIZE)) {
        log_message(LOG_LEVEL_ERROR, "Invalid chunk format: not enough data for metadata body");
+        file_destroy(file);
        array_list_delete(files);
        return NULL;
      }
      file->metadata = metadata_from_buf(&data_pointer);
      remaining_size -= sizeof(int);
-      if (file->metadata)
+      if (present_flag == 1) {
+        if (file->metadata == NULL) {
+          file_destroy(file);
+          array_list_delete(files);
+          return NULL;
+        }
        remaining_size -= FILE_METADATA_WIRE_SIZE;
+      }
    }

    if (remaining_size < sizeof(size_t)) {
      log_message(LOG_LEVEL_ERROR, "Invalid chunk format: not enough data for data size");
+      file_destroy(file);
      array_list_delete(files);
      return NULL;
    }
@@ -156,6 +230,7 @@ Chunk* chunk_deserialize(Data* data, bool use_metadata) {

    if (remaining_size < file_data_size) {
      log_message(LOG_LEVEL_ERROR, "Invalid chunk format: not enough data for file content");
+      file_destroy(file);
      array_list_delete(files);
      return NULL;
    }
@@ -164,29 +239,50 @@ Chunk* chunk_deserialize(Data* data, bool use_metadata) {
    if (file_data_size > MAX_FILE_DATA_SIZE) {
      log_message(LOG_LEVEL_ERROR, "File data size %zu exceeds maximum %llu", file_data_size,
                  (unsigned long long)MAX_FILE_DATA_SIZE);
+      file_destroy(file);
      array_list_delete(files);
      return NULL;
    }

-    void* file_data = malloc(file_data_size);
+    size_t allocation_size = file_data_size > 0 ? file_data_size : 1;
+    void* file_data = malloc(allocation_size);
    if (file_data == NULL) {
      perror("Could not allocate memory for file data");
+      file_destroy(file);
      array_list_delete(files);
      return NULL;
    }
    memcpy(file_data, data_pointer, file_data_size);
+    Data* replacement = data_create(file_data, file_data_size);
+    if (replacement == NULL) {
+      file_destroy(file);
+      array_list_delete(files);
+      return NULL;
+    }
    data_destroy(file->data);
-    file->data = data_create(file_data, file_data_size);
+    file->data = replacement;
    data_pointer += file_data_size;
    remaining_size -= file_data_size;

-    array_list_add(files, file);
+    if (!array_list_add(files, file)) {
+      file_destroy(file);
+      array_list_delete(files);
+      return NULL;
+    }
  }

  File** file_array = (File**)array_list_to_array(files);
+  if (files->size > 0 && file_array == NULL) {
+    array_list_delete(files);
+    return NULL;
+  }
  Chunk* chunk = chunk_create(file_array, files->size);

  free(file_array);
+  if (chunk == NULL) {
+    array_list_delete(files);
+    return NULL;
+  }
  files->item_destroyer = NULL;
  array_list_delete(files);

@@ -207,14 +303,14 @@ Data* chunk_compress(Chunk* chunk, int compression_level, bool use_metadata) {
 }

 Chunk* receive_chunk_data(int fd, const Config* config) {
-  Data* chunk_data = receive_data(fd);
+  Data* chunk_data = receive_data_limited(fd, MAX_CHUNK_SIZE);
  if (chunk_data == NULL) {
    log_message(LOG_LEVEL_ERROR, "Failed to receive chunk data");
    return NULL;
  }
  Data* data_to_process = chunk_data;
  if (config->use_compression) {
-    data_to_process = data_decompress(chunk_data);
+    data_to_process = data_decompress_limited(chunk_data, MAX_CHUNK_SIZE);
    data_destroy(chunk_data);
    if (data_to_process == NULL) {
      log_message(LOG_LEVEL_ERROR, "Failed to decompress chunk");
@@ -2,6 +2,8 @@
 #include "data.h"
 #include "log.h"
 #include <stdlib.h>
+#include <limits.h>
+#include <stdint.h>
 #include <string.h>
 #include <strings.h>
 #include <zstd.h>
@@ -69,7 +71,10 @@ Data* data_compress(Data* data_to_compress, int compression_level) {
  return compressed_data;
 }

-Data* data_decompress(Data* compressed_data) {
+Data* data_decompress_limited(Data* compressed_data, size_t maximum_size) {
+  if (!compressed_data || (!compressed_data->data && compressed_data->size != 0) ||
+      maximum_size == 0)
+    return NULL;
  log_message(LOG_LEVEL_DEBUG, "Start to decompress data");
  unsigned long long dst_size =
      ZSTD_getFrameContentSize(compressed_data->data, compressed_data->size);
@@ -82,15 +87,16 @@ Data* data_decompress(Data* compressed_data) {
  // ZSTD_CONTENTSIZE_UNKNOWN (~2^64) can cause massive allocation;
  // fall back to a conservative estimate (3x compressed size) when unknown.
  if (dst_size == ZSTD_CONTENTSIZE_UNKNOWN) {
+    if (compressed_data->size > ULLONG_MAX / 3)
+      return NULL;
    dst_size = compressed_data->size * 3;
    if (dst_size < INITIAL_DECOMPRESS_BUF_SIZE)
      dst_size = INITIAL_DECOMPRESS_BUF_SIZE;
-    if (dst_size > MAX_DECOMPRESSED_SIZE)
-      dst_size = MAX_DECOMPRESSED_SIZE;
  }
-  if (dst_size > MAX_DECOMPRESSED_SIZE) {
-    log_message(LOG_LEVEL_ERROR, "Declared decompressed size exceeds %llu bytes",
-                (unsigned long long)MAX_DECOMPRESSED_SIZE);
+  unsigned long long hard_limit =
+      maximum_size < MAX_DECOMPRESSED_SIZE ? maximum_size : MAX_DECOMPRESSED_SIZE;
+  if (dst_size > hard_limit) {
+    log_message(LOG_LEVEL_ERROR, "Declared decompressed size exceeds %llu bytes", hard_limit);
    return NULL;
  }

@@ -101,6 +107,8 @@ Data* data_decompress(Data* compressed_data) {
  }

  size_t buf_size = (dst_size > 0) ? (size_t)dst_size : INITIAL_DECOMPRESS_BUF_SIZE;
+  if (buf_size > maximum_size)
+    buf_size = maximum_size;
  Data* uncompressed_data = data_create_empty(buf_size);
  if (!uncompressed_data) {
    log_message(LOG_LEVEL_ERROR, "Failed to allocate decompression buffer");
@@ -121,7 +129,7 @@ Data* data_decompress(Data* compressed_data) {
      return NULL;
    }
    if (ret > 0 && output.pos == output.size) {
-      if (buf_size >= MAX_DECOMPRESSED_SIZE) {
+      if (buf_size >= hard_limit || buf_size > SIZE_MAX / 2) {
        log_message(LOG_LEVEL_ERROR, "Decompressed data exceeds %llu bytes",
                    (unsigned long long)MAX_DECOMPRESSED_SIZE);
        ZSTD_freeDCtx(dctx);
@@ -129,8 +137,8 @@ Data* data_decompress(Data* compressed_data) {
        return NULL;
      }
      buf_size *= 2;
-      if (buf_size > MAX_DECOMPRESSED_SIZE)
-        buf_size = MAX_DECOMPRESSED_SIZE;
+      if (buf_size > hard_limit)
+        buf_size = (size_t)hard_limit;
      void* new_data = realloc(uncompressed_data->data, buf_size);
      if (!new_data) {
        log_message(LOG_LEVEL_ERROR, "Failed to grow decompression buffer");
@@ -150,3 +158,7 @@ Data* data_decompress(Data* compressed_data) {
  log_message(LOG_LEVEL_DEBUG, "Decompressed data successfully");
  return uncompressed_data;
 }
+
+Data* data_decompress(Data* compressed_data) {
+  return data_decompress_limited(compressed_data, MAX_DECOMPRESSED_SIZE);
+}
@@ -6,6 +6,7 @@

 Data* data_compress(Data* data_to_compress, int compression_level);
 Data* data_decompress(Data* compressed_data);
+Data* data_decompress_limited(Data* compressed_data, size_t maximum_size);
 bool compression_should_skip(const char* path);

 #endif
@@ -100,6 +100,43 @@ static void config_set_defaults(Config* config) {
  config->compress_choice = NULL;
 }

+static bool valid_wire_bool(int value) {
+  return value == 0 || value == 1;
+}
+
+static bool receive_wire_bool(int fd, bool* value) {
+  int wire_value;
+  if (!receive_int(fd, &wire_value) || !valid_wire_bool(wire_value))
+    return false;
+  *value = wire_value != 0;
+  return true;
+}
+
+static bool validate_received_config(const Config* config) {
+  return valid_wire_bool(config->save_to_disk) && valid_wire_bool(config->use_multithreading) &&
+         valid_wire_bool(config->use_chunk_serialization) &&
+         valid_wire_bool(config->use_compression) && valid_wire_bool(config->use_metadata) &&
+         valid_wire_bool(config->use_sendfile) && valid_wire_bool(config->use_delete) &&
+         valid_wire_bool(config->use_incremental) && valid_wire_bool(config->use_delta) &&
+         valid_wire_bool(config->backup) && valid_wire_bool(config->follow_symlinks) &&
+         valid_wire_bool(config->copy_links) && valid_wire_bool(config->safe_links) &&
+         valid_wire_bool(config->copy_unsafe_links) &&
+         valid_wire_bool(config->preserve_hard_links) && valid_wire_bool(config->preserve_acls) &&
+         valid_wire_bool(config->preserve_xattrs) && valid_wire_bool(config->preserve_devices) &&
+         valid_wire_bool(config->preserve_sparse) && valid_wire_bool(config->update) &&
+         valid_wire_bool(config->inplace) && valid_wire_bool(config->append) &&
+         valid_wire_bool(config->append_verify) && valid_wire_bool(config->delete_excluded) &&
+         valid_wire_bool(config->delete_after) && valid_wire_bool(config->relative) &&
+         valid_wire_bool(config->prune_empty_dirs) && valid_wire_bool(config->partial) &&
+         valid_wire_bool(config->delete_before) && valid_wire_bool(config->checksum) &&
+         (!config->use_compression ||
+          (config->compression_level >= 1 && config->compression_level <= 22)) &&
+         config->chunk_size > 0 && config->chunk_size <= MAX_CHUNK_SIZE &&
+         config->delta_block_size >= DELTA_BLOCK_SIZE_MIN &&
+         config->delta_block_size <= DELTA_BLOCK_SIZE_MAX &&
+         config->delta_max_file_size <= DELTA_MAX_FILE_SIZE && config->max_delete >= 0;
+}
+
 Config* config_create(void) {
  Config* config = malloc(sizeof(Config));
  if (!config)
@@ -137,6 +174,10 @@ void config_parse_ssh_dest(Config* config) {
 void config_delete(Config* config) {
  if (config == NULL)
    return;
+  if (config->log_file) {
+    fclose(config->log_file);
+    config->log_file = NULL;
+  }
  free(config->version);
  free(config->send_directory);
  free(config->receive_root_directory);
@@ -221,52 +262,34 @@ static bool receive_core_fields(int fd, Config* c) {
  c->receive_root_directory = receive_str(fd);
  if (!c->send_directory || !c->receive_root_directory)
    return false;
-  if (!receive_int(fd, &value))
+  if (!receive_wire_bool(fd, &c->save_to_disk) || !receive_wire_bool(fd, &c->use_multithreading) ||
+      !receive_wire_bool(fd, &c->use_chunk_serialization) ||
+      !receive_wire_bool(fd, &c->use_compression) || !receive_wire_bool(fd, &c->use_metadata))
    return false;
-  c->save_to_disk = value;
-  if (!receive_int(fd, &value))
-    return false;
-  c->use_multithreading = value;
-  if (!receive_int(fd, &value))
-    return false;
-  c->use_chunk_serialization = value;
-  if (!receive_int(fd, &value))
-    return false;
-  c->use_compression = value;
-  if (!receive_int(fd, &value))
-    return false;
-  c->use_metadata = value;
  if (!receive_int(fd, &value))
    return false;
  c->compression_level = value;
  if (!receive_n_data(fd, &c->chunk_size, sizeof(c->chunk_size)))
    return false;
-  if (!receive_int(fd, &value))
+  if (!receive_wire_bool(fd, &c->use_sendfile))
    return false;
-  c->use_sendfile = value;
  return true;
 }

 static bool receive_delta_fields(int fd, Config* c) {
-  int value;
-  if (!receive_int(fd, &value))
+  if (!receive_wire_bool(fd, &c->use_delete))
    return false;
-  c->use_delete = value;
-  if (!receive_int(fd, &value))
+  if (!receive_wire_bool(fd, &c->use_incremental))
    return false;
-  c->use_incremental = value;
-  if (!receive_int(fd, &value))
+  if (!receive_wire_bool(fd, &c->use_delta))
    return false;
-  c->use_delta = value;
  return receive_n_data(fd, &c->delta_block_size, sizeof(c->delta_block_size)) &&
         receive_n_data(fd, &c->delta_max_file_size, sizeof(unsigned long long));
 }

 static bool receive_file_options(int fd, Config* c) {
-  int value;
-  if (!receive_int(fd, &value))
+  if (!receive_wire_bool(fd, &c->backup))
    return false;
-  c->backup = value;
  c->backup_dir = receive_str(fd);
  if (!c->backup_dir)
    return false;
@@ -274,47 +297,38 @@ static bool receive_file_options(int fd, Config* c) {
                   &c->copy_unsafe_links, &c->preserve_hard_links, &c->preserve_acls,
                   &c->preserve_xattrs,   &c->preserve_devices,    &c->preserve_sparse};
  for (size_t i = 0; i < sizeof(flags) / sizeof(flags[0]); i++) {
-    if (!receive_int(fd, &value))
+    if (!receive_wire_bool(fd, flags[i]))
      return false;
-    *flags[i] = value;
  }
  return true;
 }

 static bool receive_selection_options(int fd, Config* c) {
-  int value;
  bool* flags[] = {&c->update,        &c->inplace,         &c->append,
                   &c->append_verify, &c->delete_excluded, &c->delete_after};
  for (size_t i = 0; i < sizeof(flags) / sizeof(flags[0]); i++) {
-    if (!receive_int(fd, &value))
+    if (!receive_wire_bool(fd, flags[i]))
      return false;
-    *flags[i] = value;
  }
  if (!receive_n_data(fd, &c->max_delete, sizeof(c->max_delete)))
    return false;
-  if (!receive_int(fd, &value))
+  if (!receive_wire_bool(fd, &c->relative))
    return false;
-  c->relative = value;
-  if (!receive_int(fd, &value))
+  if (!receive_wire_bool(fd, &c->prune_empty_dirs))
    return false;
-  c->prune_empty_dirs = value;
  return true;
 }

 static bool receive_resume_options(int fd, Config* c) {
-  int value;
  c->temp_dir = receive_str(fd);
-  if (!c->temp_dir || !receive_int(fd, &value))
+  if (!c->temp_dir || !receive_wire_bool(fd, &c->partial))
    return false;
-  c->partial = value;
  c->partial_dir = receive_str(fd);
  c->suffix = c->partial_dir ? receive_str(fd) : NULL;
-  if (!c->partial_dir || !c->suffix || !receive_int(fd, &value))
+  if (!c->partial_dir || !c->suffix || !receive_wire_bool(fd, &c->delete_before))
    return false;
-  c->delete_before = value;
-  if (!receive_int(fd, &value))
+  if (!receive_wire_bool(fd, &c->checksum))
    return false;
-  c->checksum = value;
  c->compress_choice = receive_str(fd);
  return c->compress_choice != NULL;
 }
@@ -361,6 +375,11 @@ Config* config_receive(int file_descriptor) {
    send_status(file_descriptor, STATUS_ERROR);
    goto error;
  }
+  if (!validate_received_config(config)) {
+    fprintf(stderr, "Invalid configuration received from client\n");
+    send_status(file_descriptor, STATUS_ERROR);
+    goto error;
+  }
  if (!send_status(file_descriptor, STATUS_OK))
    goto error;
  return config;
@@ -21,6 +21,7 @@ Data* data_create_reserve(size_t size) {
  }
  d->data = NULL;
  d->size = size;
+  d->protocol_charge = 0;
  return d;
 }

@@ -33,12 +34,15 @@ Data* data_create(void* data, size_t data_size) {
  }
  new_data->data = data;
  new_data->size = data_size;
+  new_data->protocol_charge = 0;
  return new_data;
 }

 void data_destroy(Data* data) {
  if (data == NULL)
    return;
+  if (data->protocol_charge != 0)
+    protocol_release_memory(data->protocol_charge);
  free(data->data);
  free(data);
 }
@@ -6,11 +6,14 @@
 typedef struct {
  void* data;
  size_t size;
+  /* Non-zero only for a buffer charged to the protocol connection budget. */
+  size_t protocol_charge;
 } Data;

 Data* data_create_empty(size_t data_size);
 Data* data_create_reserve(size_t size);
 Data* data_create(void* data, size_t data_size);
 void data_destroy(Data* data);
+void protocol_release_memory(size_t charge);

 #endif
@@ -1,6 +1,7 @@
 #include "delta.h"
 #include "log.h"
 #include <stdint.h>
+#include <limits.h>
 #include <stdlib.h>
 #include <string.h>

@@ -36,6 +37,10 @@ DeltaSignature* delta_signature_create(const void* old_file_data, uint64_t old_f
  if (old_file_data == NULL || old_file_size == 0 || block_size == 0)
    return NULL;

+  if (old_file_size > DELTA_MAX_FILE_SIZE || block_size > DELTA_BLOCK_SIZE_MAX ||
+      old_file_size > UINT32_MAX * (uint64_t)block_size)
+    return NULL;
+
  uint32_t block_count = (uint32_t)((old_file_size + block_size - 1) / block_size);

  DeltaSignature* sig = malloc(sizeof(DeltaSignature));
@@ -45,7 +50,11 @@ DeltaSignature* delta_signature_create(const void* old_file_data, uint64_t old_f
  sig->file_size = old_file_size;
  sig->block_size = block_size;
  sig->block_count = block_count;
-  sig->blocks = malloc(block_count * sizeof(DeltaBlockSig));
+  if (block_count == 0) {
+    free(sig);
+    return NULL;
+  }
+  sig->blocks = malloc((size_t)block_count * sizeof(DeltaBlockSig));
  if (!sig->blocks) {
    free(sig);
    return NULL;
@@ -67,8 +76,11 @@ Data* delta_signature_serialize(const DeltaSignature* sig) {
  if (!sig)
    return NULL;

-  uint64_t total = sizeof(uint64_t) + sizeof(uint32_t) + sizeof(uint32_t) +
-                   (uint64_t)sig->block_count * (sizeof(uint32_t) + sizeof(uint32_t));
+  uint64_t block_bytes = (uint64_t)sig->block_count * (sizeof(uint32_t) + sizeof(uint32_t));
+  uint64_t total = sizeof(uint64_t) + sizeof(uint32_t) + sizeof(uint32_t) + block_bytes;
+  if (block_bytes > UINT64_MAX - (sizeof(uint64_t) + sizeof(uint32_t) + sizeof(uint32_t)) ||
+      total > SIZE_MAX)
+    return NULL;

  uint8_t* buf = malloc((size_t)total);
  if (!buf)
@@ -118,6 +130,13 @@ DeltaSignature* delta_signature_deserialize(const Data* data) {
    return NULL;
  }

+  if (sig->block_size == 0 || sig->block_size > DELTA_BLOCK_SIZE_MAX ||
+      sig->file_size > DELTA_MAX_FILE_SIZE || sig->file_size == 0 ||
+      (sig->file_size + sig->block_size - 1) / sig->block_size != sig->block_count) {
+    free(sig);
+    return NULL;
+  }
+
  uint64_t expected = sizeof(uint64_t) + sizeof(uint32_t) + sizeof(uint32_t) +
                      (uint64_t)sig->block_count * (sizeof(uint32_t) + sizeof(uint32_t));
  if (data->size < expected) {
@@ -156,8 +175,10 @@ void delta_signature_destroy(DeltaSignature* sig) {
 static bool ensure_capacity(DeltaInstruction** instrs, uint32_t* capacity, uint32_t count) {
  if (count < *capacity)
    return true;
+  if (*capacity > MAX_DELTA_INSTRUCTIONS / 2)
+    return false;
  uint32_t new_cap = *capacity * 2;
-  DeltaInstruction* tmp = realloc(*instrs, new_cap * sizeof(DeltaInstruction));
+  DeltaInstruction* tmp = realloc(*instrs, (size_t)new_cap * sizeof(DeltaInstruction));
  if (!tmp)
    return false;
  *instrs = tmp;
@@ -169,6 +190,8 @@ static bool flush_literal(DeltaInstruction** instrs, uint32_t* capacity, uint32_
                          const uint8_t* data, uint64_t start, uint64_t end) {
  if (start >= end)
    return true;
+  if (end - start > UINT32_MAX || *count >= MAX_DELTA_INSTRUCTIONS)
+    return false;
  uint32_t lit_len = (uint32_t)(end - start);
  if (!ensure_capacity(instrs, capacity, *count))
    return false;
@@ -183,16 +206,26 @@ static bool flush_literal(DeltaInstruction** instrs, uint32_t* capacity, uint32_
  return true;
 }

+static void free_instructions(DeltaInstruction* instrs, uint32_t count) {
+  if (!instrs)
+    return;
+  for (uint32_t i = 0; i < count; i++)
+    if (instrs[i].type == DELTA_INSTR_LITERAL)
+      free(instrs[i].literal.data);
+  free(instrs);
+}
+
 Delta* delta_compute(const void* new_file_data, uint64_t new_file_size, const DeltaSignature* sig,
                     uint32_t block_size) {
-  if (!new_file_data || !sig || new_file_size == 0 || block_size == 0)
+  if (!new_file_data || !sig || !sig->blocks || new_file_size == 0 || block_size == 0 ||
+      block_size > DELTA_BLOCK_SIZE_MAX || sig->block_size != block_size)
    return NULL;

  const uint8_t* new_data = (const uint8_t*)new_file_data;

  uint32_t capacity = 64;
  uint32_t count = 0;
-  DeltaInstruction* instrs = malloc(capacity * sizeof(DeltaInstruction));
+  DeltaInstruction* instrs = malloc((size_t)capacity * sizeof(DeltaInstruction));
  if (!instrs)
    return NULL;

@@ -236,14 +269,14 @@ Delta* delta_compute(const void* new_file_data, uint64_t new_file_size, const De
        if (xxh == sig->blocks[j].xxhash) {
          if (has_literal) {
            if (!flush_literal(&instrs, &capacity, &count, new_data, literal_start, i)) {
-              free(instrs);
+              free_instructions(instrs, count);
              return NULL;
            }
            has_literal = false;
          }

          if (!ensure_capacity(&instrs, &capacity, count)) {
-            free(instrs);
+            free_instructions(instrs, count);
            return NULL;
          }
          instrs[count].type = DELTA_INSTR_BLOCK_MATCH;
@@ -271,18 +304,14 @@ Delta* delta_compute(const void* new_file_data, uint64_t new_file_size, const De

  if (has_literal) {
    if (!flush_literal(&instrs, &capacity, &count, new_data, literal_start, new_file_size)) {
-      free(instrs);
+      free_instructions(instrs, count);
      return NULL;
    }
  }

  Delta* delta = malloc(sizeof(Delta));
  if (!delta) {
-    for (uint32_t k = 0; k < count; k++) {
-      if (instrs[k].type == DELTA_INSTR_LITERAL)
-        free(instrs[k].literal.data);
-    }
-    free(instrs);
+    free_instructions(instrs, count);
    return NULL;
  }

@@ -292,11 +321,24 @@ Delta* delta_compute(const void* new_file_data, uint64_t new_file_size, const De
  delta->delta_size = 0;

  for (uint32_t k = 0; k < count; k++) {
+    if (delta->delta_size == UINT64_MAX) {
+      delta_destroy(delta);
+      return NULL;
+    }
    delta->delta_size += 1;
    if (instrs[k].type == DELTA_INSTR_BLOCK_MATCH) {
+      if (delta->delta_size > UINT64_MAX - sizeof(uint32_t) * 3) {
+        delta_destroy(delta);
+        return NULL;
+      }
      delta->delta_size += sizeof(uint32_t) * 3;
    } else {
-      delta->delta_size += sizeof(uint32_t) + instrs[k].literal.length;
+      uint64_t extra = sizeof(uint32_t) + instrs[k].literal.length;
+      if (delta->delta_size > UINT64_MAX - extra) {
+        delta_destroy(delta);
+        return NULL;
+      }
+      delta->delta_size += extra;
    }
  }

@@ -307,7 +349,12 @@ Data* delta_serialize(const Delta* delta) {
  if (!delta)
    return NULL;

-  uint64_t total = sizeof(uint64_t) + sizeof(uint32_t) + delta->delta_size;
+  if (delta->instruction_count > 0 && !delta->instructions)
+    return NULL;
+  uint64_t header_size = sizeof(uint64_t) + sizeof(uint32_t);
+  if (delta->delta_size > UINT64_MAX - header_size || header_size + delta->delta_size > SIZE_MAX)
+    return NULL;
+  uint64_t total = header_size + delta->delta_size;
  uint8_t* buf = malloc((size_t)total);
  if (!buf)
    return NULL;
@@ -365,8 +412,10 @@ Delta* delta_deserialize(const Data* data) {
    return NULL;
  }

-  delta->instructions = malloc(delta->instruction_count * sizeof(DeltaInstruction));
-  if (!delta->instructions) {
+  delta->instructions = delta->instruction_count == 0
+                            ? NULL
+                            : malloc((size_t)delta->instruction_count * sizeof(DeltaInstruction));
+  if (delta->instruction_count > 0 && !delta->instructions) {
    free(delta);
    return NULL;
  }
@@ -375,11 +424,7 @@ Delta* delta_deserialize(const Data* data) {

  for (uint32_t i = 0; i < delta->instruction_count; i++) {
    if (pos >= data->size) {
-      for (uint32_t k = 0; k < i; k++) {
-        if (delta->instructions[k].type == DELTA_INSTR_LITERAL)
-          free(delta->instructions[k].literal.data);
-      }
-      free(delta->instructions);
+      free_instructions(delta->instructions, i);
      free(delta);
      return NULL;
    }
@@ -391,12 +436,8 @@ Delta* delta_deserialize(const Data* data) {
    delta->delta_size += 1;

    if (type == DELTA_OP_BLOCK_MATCH) {
-      if (pos + sizeof(uint32_t) * 3 > data->size) {
-        for (uint32_t k = 0; k < i; k++) {
-          if (delta->instructions[k].type == DELTA_INSTR_LITERAL)
-            free(delta->instructions[k].literal.data);
-        }
-        free(delta->instructions);
+      if (data->size - pos < sizeof(uint32_t) * 3) {
+        free_instructions(delta->instructions, i);
        free(delta);
        return NULL;
      }
@@ -409,12 +450,8 @@ Delta* delta_deserialize(const Data* data) {
      pos += sizeof(uint32_t);
      delta->delta_size += sizeof(uint32_t) * 3;
    } else if (type == DELTA_OP_LITERAL) {
-      if (pos + sizeof(uint32_t) > data->size) {
-        for (uint32_t k = 0; k < i; k++) {
-          if (delta->instructions[k].type == DELTA_INSTR_LITERAL)
-            free(delta->instructions[k].literal.data);
-        }
-        free(delta->instructions);
+      if (data->size - pos < sizeof(uint32_t)) {
+        free_instructions(delta->instructions, i);
        free(delta);
        return NULL;
      }
@@ -423,23 +460,15 @@ Delta* delta_deserialize(const Data* data) {
      pos += sizeof(uint32_t);

      uint32_t lit_len = delta->instructions[i].literal.length;
-      if (pos + lit_len > data->size) {
-        for (uint32_t k = 0; k < i; k++) {
-          if (delta->instructions[k].type == DELTA_INSTR_LITERAL)
-            free(delta->instructions[k].literal.data);
-        }
-        free(delta->instructions);
+      if (lit_len > data->size - pos) {
+        free_instructions(delta->instructions, i);
        free(delta);
        return NULL;
      }
-      delta->instructions[i].literal.data = malloc(lit_len);
+      delta->instructions[i].literal.data = malloc(lit_len ? lit_len : 1);
      if (!delta->instructions[i].literal.data) {
        log_message(LOG_LEVEL_ERROR, "Failed to allocate %u bytes for literal data", lit_len);
-        for (uint32_t k = 0; k < i; k++) {
-          if (delta->instructions[k].type == DELTA_INSTR_LITERAL)
-            free(delta->instructions[k].literal.data);
-        }
-        free(delta->instructions);
+        free_instructions(delta->instructions, i);
        free(delta);
        return NULL;
      }
@@ -447,11 +476,7 @@ Delta* delta_deserialize(const Data* data) {
      pos += lit_len;
      delta->delta_size += sizeof(uint32_t) + lit_len;
    } else {
-      for (uint32_t k = 0; k < i; k++) {
-        if (delta->instructions[k].type == DELTA_INSTR_LITERAL)
-          free(delta->instructions[k].literal.data);
-      }
-      free(delta->instructions);
+      free_instructions(delta->instructions, i);
      free(delta);
      return NULL;
    }
@@ -463,10 +488,11 @@ Delta* delta_deserialize(const Data* data) {
 void* delta_apply(const void* old_data, uint64_t old_size, const Delta* delta,
                  uint32_t block_size) {
  if (!old_data || !delta || (delta->new_file_size > 0 && delta->instructions == NULL) ||
-      (delta->instruction_count > 0 && block_size == 0))
+      (delta->instruction_count > 0 && block_size == 0) ||
+      delta->new_file_size > DELTA_MAX_FILE_SIZE || delta->new_file_size > SIZE_MAX)
    return NULL;

-  void* output = malloc((size_t)delta->new_file_size);
+  void* output = malloc(delta->new_file_size ? (size_t)delta->new_file_size : 1);
  if (!output)
    return NULL;

@@ -491,7 +517,7 @@ void* delta_apply(const void* old_data, uint64_t old_size, const Delta* delta,
      }
      memcpy(out + out_pos, old + src_offset, len);
      out_pos += len;
-    } else {
+    } else if (delta->instructions[i].type == DELTA_INSTR_LITERAL) {
      uint32_t len = delta->instructions[i].literal.length;
      if (out_pos > delta->new_file_size || (uint64_t)len > delta->new_file_size - out_pos) {
        free(output);
@@ -499,6 +525,9 @@ void* delta_apply(const void* old_data, uint64_t old_size, const Delta* delta,
      }
      memcpy(out + out_pos, delta->instructions[i].literal.data, len);
      out_pos += len;
+    } else {
+      free(output);
+      return NULL;
    }
  }

@@ -534,7 +563,7 @@ bool delta_should_attempt(uint64_t old_size, uint64_t new_size, uint64_t max_fil
 }

 bool delta_is_worthwhile(const Delta* delta, uint64_t new_file_size) {
-  if (!delta || delta->instruction_count == 0)
+  if (!delta || delta->instruction_count == 0 || new_file_size == 0)
    return false;

  bool has_match = false;
@@ -24,6 +24,9 @@
 #include "protocol.h"
 #include "utils.h"

+#define MAX_SERVER_DELETE_COUNT 100000U
+#define MAX_FILE_DATA_SIZE MAX_RECEIVE_FILE_SIZE
+
 bool file_checksum(File* file, uint64_t* checksum) {
  if (!file || !checksum || !file->data)
    return false;
@@ -46,7 +49,7 @@ File* file_create(const char* path) {
    return NULL;
  }

-  int path_len = strlen(path);
+  size_t path_len = strlen(path);
  file->path = (char*)malloc(path_len + 1);
  if (file->path == NULL) {
    free(file);
@@ -102,7 +105,7 @@ void file_metadata_destroy(void* metadata) {
 }

 bool file_load_data(File* file) {
-  if (file == NULL)
+  if (file == NULL || !file->data)
    return false;
  if (file->data->data == NULL) {
    if (file->data->size == 0)
@@ -126,7 +129,7 @@ bool file_load_data(File* file) {

 bool file_send_single_calls(File* file, int file_descriptor, bool use_metadata,
                            int compression_level, bool send_path) {
-  if (!file || !file->path || !file->data)
+  if (!file || !file->path || !file->data || (file->data->size != 0 && !file->data->data))
    return false;
  const Data* data_to_send = file->data;
  Data* compressed_data = NULL;
@@ -154,13 +157,52 @@ bool file_send_single_calls(File* file, int file_descriptor, bool use_metadata,
  return true;
 }

-void file_set_authorized_root(int fd, const char* canonical_path) {
-  file_store_set_authorized_root(fd, canonical_path);
+static bool to_disk_secure(const char* path, const void* data, unsigned long long data_size,
+                           bool inplace, bool sparse, const FileMetadata* metadata);
+static int open_secure_parent(const char* path, char** leaf_out, bool create_dirs);
+static bool ensure_directory_secure(const char* path);
+
+bool file_path_exists_secure(const char* path) {
+  struct stat st;
+  return file_stat_secure(path, &st);
 }

+bool file_stat_secure(const char* path, struct stat* st) {
+  if (!path || !st)
+    return false;
+  char* leaf = NULL;
+  int parent_fd = open_secure_parent(path, &leaf, false);
+  if (parent_fd < 0)
+    return false;
+  int fd = openat(parent_fd, leaf, O_RDONLY | O_NONBLOCK | O_CLOEXEC | O_NOFOLLOW);
+  bool exists = fd >= 0 && fstat(fd, st) == 0 && S_ISREG(st->st_mode);
+  if (fd >= 0)
+    close(fd);
+  close(parent_fd);
+  free(leaf);
+  return exists;
+}
+static bool rename_secure(const char* old_path, const char* new_path);
+static int authorized_root_fd = -1;
+static char* authorized_root_path;
+
 static bool path_is_within_root(const char* root, const char* path) {
-  size_t n = strlen(root);
-  return strncmp(root, path, n) == 0 && (path[n] == '\0' || path[n] == '/');
+  size_t root_len = strlen(root);
+  return strncmp(root, path, root_len) == 0 && (path[root_len] == '\0' || path[root_len] == '/');
+}
+
+bool file_set_authorized_root(int fd, const char* canonical_path) {
+  char* path_copy = canonical_path ? str_dup(canonical_path) : NULL;
+  if (canonical_path && !path_copy) {
+    authorized_root_fd = -1;
+    free(authorized_root_path);
+    authorized_root_path = NULL;
+    return false;
+  }
+  authorized_root_fd = fd;
+  free(authorized_root_path);
+  authorized_root_path = path_copy;
+  return true;
 }

 bool file_save_to_disk(const char* root_directory, const File* file, const Config* config) {
@@ -170,9 +212,11 @@ bool file_save_to_disk(const char* root_directory, const File* file, const Confi
  const char* backup_suffix = (config && config->suffix) ? config->suffix : "~";
  const char* backup_dir = (config && config->backup_dir) ? config->backup_dir : NULL;
  const char* partial_dir = (config && config->partial_dir) ? config->partial_dir : NULL;
-  char *confined_backup = NULL, *confined_partial = NULL;
+  char *confined_backup = NULL, *confined_partial = NULL, *disk_path = NULL;
+  char *backup_path = NULL, *parent_copy = NULL;

-  if (!file || !file->path || !file->data || has_path_traversal(file->path) ||
+  if (!file || !file->path || !file->data || (file->data->size != 0 && !file->data->data) ||
+      has_path_traversal(file->path) ||
      (backup_enabled &&
       (!backup_suffix || backup_suffix[0] == '\0' || strchr(backup_suffix, '/') != NULL ||
        strcmp(backup_suffix, ".") == 0 || strcmp(backup_suffix, "..") == 0))) {
@@ -192,45 +236,20 @@ bool file_save_to_disk(const char* root_directory, const File* file, const Confi
    return false;
  }

-  char* resolved_root = NULL;
  const char* actual_root =
      (partial_dir && config && config->partial) ? confined_partial : root_directory;
-  resolved_root = realpath(actual_root, NULL);
-  if (resolved_root == NULL) {
-    if (mkdir_r(actual_root)) {
-      resolved_root = realpath(actual_root, NULL);
-    }
-  }
-  if (resolved_root == NULL) {
-    log_message(LOG_LEVEL_ERROR, "Failed to resolve destination root: %s", actual_root);
-    free(confined_backup);
-    free(confined_partial);
-    return false;
-  }
-  char* resolved_base = realpath(root_directory, NULL);
-  if (resolved_base == NULL || !path_is_within_root(resolved_base, resolved_root)) {
-    free(resolved_base);
-    free(confined_backup);
-    free(confined_partial);
-    free(resolved_root);
-    return false;
-  }
-  free(resolved_base);
-
-  char* disk_path = path_cat(resolved_root, file->path);
+  disk_path = path_cat(actual_root, file->path);
  if (disk_path == NULL) {
    free(confined_backup);
    free(confined_partial);
-    free(resolved_root);
    return false;
  }

  /* --update is receiver-side policy: never replace a newer destination. */
  if (config && config->update) {
    struct stat destination_stat;
-    if (stat(disk_path, &destination_stat) == 0 && file->metadata &&
+    if (file_stat_secure(disk_path, &destination_stat) && file->metadata &&
        destination_stat.st_mtime > file->metadata->mtime_sec) {
-      free(resolved_root);
      free(confined_backup);
      free(confined_partial);
      free(disk_path);
@@ -240,109 +259,61 @@ bool file_save_to_disk(const char* root_directory, const File* file, const Confi

  if (backup_enabled) {
    struct stat backup_stat;
-    if (stat(disk_path, &backup_stat) == 0) {
-      char* backup_path = NULL;
+    if (file_stat_secure(disk_path, &backup_stat)) {
      if (backup_dir) {
-        char* resolved_backup_dir = realpath(confined_backup, NULL);
-        if (!resolved_backup_dir) {
-          mkdir_r(confined_backup);
-          resolved_backup_dir = realpath(confined_backup, NULL);
-        }
-        if (resolved_backup_dir) {
-          char* backup_base = realpath(root_directory, NULL);
-          if (backup_base && path_is_within_root(backup_base, resolved_backup_dir))
-            backup_path = path_cat(resolved_backup_dir, file->path);
-          free(backup_base);
-          free(resolved_backup_dir);
-        }
-      }
-      if (!backup_path) {
+        backup_path = path_cat(confined_backup, file->path);
+      } else {
        size_t path_len = strlen(disk_path);
        size_t suffix_len = strlen(backup_suffix);
+        if (path_len > SIZE_MAX - suffix_len - 1)
+          goto fail;
        backup_path = malloc(path_len + suffix_len + 1);
        if (backup_path) {
          memcpy(backup_path, disk_path, path_len);
          memcpy(backup_path + path_len, backup_suffix, suffix_len + 1);
        }
      }
-      if (backup_path) {
-        char* backup_dir_path = str_dup(backup_path);
-        if (backup_dir_path) {
-          const char* bdir = dirname(backup_dir_path);
-          mkdir_r(bdir);
-          free(backup_dir_path);
-        }
-        if (!file_store_rename_secure(disk_path, backup_path)) {
-          free(backup_path);
-          free(resolved_root);
-          free(confined_backup);
-          free(confined_partial);
-          free(disk_path);
-          return false;
-        }
-        free(backup_path);
-      }
+      if (!backup_path)
+        goto fail;
+      parent_copy = str_dup(backup_path);
+      if (!parent_copy || !ensure_directory_secure(dirname(parent_copy)))
+        goto fail;
+      free(parent_copy);
+      parent_copy = NULL;
+      if (!rename_secure(disk_path, backup_path))
+        goto fail;
+      free(backup_path);
+      backup_path = NULL;
    }
  }

-  char* dir_dup = str_dup(disk_path);
-  if (!dir_dup) {
-    free(confined_backup);
-    free(confined_partial);
-    free(resolved_root);
-    free(disk_path);
-    return false;
-  }
-  char* dir_str = dirname(dir_dup);
-  if (!mkdir_r(dir_str)) {
-    free(dir_dup);
-    free(confined_backup);
-    free(confined_partial);
-    free(resolved_root);
-    free(disk_path);
-    return false;
-  }
-  char* resolved_dir = realpath(dir_str, NULL);
-  free(dir_dup);
-  if (resolved_dir == NULL) {
-    log_message(LOG_LEVEL_ERROR, "Failed to resolve directory for: %s", disk_path);
-    free(confined_backup);
-    free(confined_partial);
-    free(resolved_root);
-    free(disk_path);
-    return false;
-  }
-
-  size_t root_len = strlen(resolved_root);
-  if (strncmp(resolved_dir, resolved_root, root_len) != 0 ||
-      (resolved_dir[root_len] != '\0' && resolved_dir[root_len] != '/')) {
-    log_message(LOG_LEVEL_ERROR, "Path escape detected: %s is outside %s", disk_path, actual_root);
-    free(resolved_dir);
-    free(confined_backup);
-    free(confined_partial);
-    free(resolved_root);
-    free(disk_path);
-    return false;
-  }
-  free(resolved_dir);
-  free(resolved_root);
-
-  bool ok = file_store_write_secure(disk_path, file->data->data, file->data->size, inplace, sparse,
-                                    file->metadata);
+  bool ok = to_disk_secure(disk_path, file->data->data, file->data->size, inplace, sparse,
+                           file->metadata);
+  free(parent_copy);
+  free(backup_path);
  free(confined_backup);
  free(confined_partial);
  free(disk_path);
  return ok;
+
+fail:
+  free(parent_copy);
+  free(backup_path);
+  free(confined_backup);
+  free(confined_partial);
+  free(disk_path);
+  return false;
 }

 static File* receive_delta_file(int fd, const Config* config, const char* check_path,
-                                void* old_data, unsigned long long old_size) {
+                                void* old_data, unsigned long long old_size, bool* failed) {
  if (!old_data)
    return NULL;

  DeltaSignature* sig = delta_signature_create(old_data, old_size, config->delta_block_size);
  if (!sig) {
    free(old_data);
+    *failed = true;
    return NULL;
  }

@@ -350,6 +321,7 @@ static File* receive_delta_file(int fd, const Config* config, const char* check_
  if (!sig_data) {
    delta_signature_destroy(sig);
    free(old_data);
+    *failed = true;
    return NULL;
  }

@@ -359,6 +331,7 @@ static File* receive_delta_file(int fd, const Config* config, const char* check_
  if (!sig_sent) {
    delta_signature_destroy(sig);
    free(old_data);
+    *failed = true;
    return NULL;
  }

@@ -366,26 +339,27 @@ static File* receive_delta_file(int fd, const Config* config, const char* check_
  if (!receive_status(fd, &resp)) {
    delta_signature_destroy(sig);
    free(old_data);
+    *failed = true;
    return NULL;
  }

  if (resp == STATUS_DELTA_DATA) {
-    Data* delta_data = receive_data(fd);
+    Data* delta_data = receive_data_limited(fd, MAX_RECEIVE_FILE_SIZE);
    if (!delta_data) {
      delta_signature_destroy(sig);
      free(old_data);
-      send_status(fd, STATUS_ERROR);
+      *failed = true;
      return NULL;
    }

    Data* raw_delta = delta_data;
    if (config->use_compression) {
-      raw_delta = data_decompress(delta_data);
+      raw_delta = data_decompress_limited(delta_data, MAX_RECEIVE_FILE_SIZE);
      data_destroy(delta_data);
      if (!raw_delta) {
        free(old_data);
        delta_signature_destroy(sig);
-        send_status(fd, STATUS_ERROR);
+        *failed = true;
        return NULL;
      }
    }
@@ -395,18 +369,25 @@ static File* receive_delta_file(int fd, const Config* config, const char* check_
    if (!delta) {
      free(old_data);
      delta_signature_destroy(sig);
-      send_status(fd, STATUS_ERROR);
+      *failed = true;
      return NULL;
    }

-    void* new_data = delta_apply(old_data, old_size, delta, config->delta_block_size);
    uint64_t new_size = delta->new_file_size;
+    if (new_size > MAX_RECEIVE_FILE_SIZE || new_size > SIZE_MAX) {
+      delta_destroy(delta);
+      free(old_data);
+      delta_signature_destroy(sig);
+      send_status(fd, STATUS_ERROR);
+      return NULL;
+    }
+    void* new_data = delta_apply(old_data, old_size, delta, config->delta_block_size);
    delta_destroy(delta);

    if (!new_data) {
      free(old_data);
      delta_signature_destroy(sig);
-      send_status(fd, STATUS_ERROR);
+      *failed = true;
      return NULL;
    }

@@ -415,7 +396,7 @@ static File* receive_delta_file(int fd, const Config* config, const char* check_
      free(new_data);
      free(old_data);
      delta_signature_destroy(sig);
-      send_status(fd, STATUS_ERROR);
+      *failed = true;
      return NULL;
    }

@@ -427,13 +408,21 @@ static File* receive_delta_file(int fd, const Config* config, const char* check_
        free(new_data);
        free(old_data);
        delta_signature_destroy(sig);
-        send_status(fd, STATUS_ERROR);
+        *failed = true;
        return NULL;
      }
    }

+    Data* replacement = data_create(new_data, (size_t)new_size);
+    if (replacement == NULL) {
+      file_destroy(file);
+      free(old_data);
+      delta_signature_destroy(sig);
+      send_status(fd, STATUS_ERROR);
+      return NULL;
+    }
    data_destroy(file->data);
-    file->data = data_create(new_data, (size_t)new_size);
+    file->data = replacement;

    free(old_data);
    delta_signature_destroy(sig);
@@ -446,7 +435,7 @@ static File* receive_delta_file(int fd, const Config* config, const char* check_

    File* file = file_create(check_path);
    if (!file) {
-      send_status(fd, STATUS_ERROR);
+      *failed = true;
      return NULL;
    }

@@ -455,22 +444,28 @@ static File* receive_delta_file(int fd, const Config* config, const char* check_
      file->metadata = metadata_receive(fd, &meta_ok);
      if (!meta_ok) {
        file_destroy(file);
-        send_status(fd, STATUS_ERROR);
+        *failed = true;
        return NULL;
      }
    }

-    Data* file_data = receive_data(fd);
+    Data* file_data = receive_data_limited(fd, MAX_RECEIVE_FILE_SIZE);
    if (file_data == NULL) {
      file_destroy(file);
-      send_status(fd, STATUS_ERROR);
+      *failed = true;
      return NULL;
    }

    if (config->use_compression) {
-      Data* uncompressed = data_decompress(file_data);
+      Data* uncompressed = data_decompress_limited(file_data, MAX_RECEIVE_FILE_SIZE);
      data_destroy(file_data);
      if (uncompressed == NULL) {
+        file_destroy(file);
+        *failed = true;
+        return NULL;
+      }
+      if (uncompressed->size > MAX_FILE_DATA_SIZE) {
+        data_destroy(uncompressed);
        file_destroy(file);
        send_status(fd, STATUS_ERROR);
        return NULL;
@@ -485,14 +480,19 @@ static File* receive_delta_file(int fd, const Config* config, const char* check_

  delta_signature_destroy(sig);
  free(old_data);
+  send_status(fd, STATUS_ERROR);
+  *failed = true;
  return NULL;
 }

 File* receive_incremental_check(int fd, const Config* config, bool* skipped) {
+  if (!config || !skipped) {
+    send_status(fd, STATUS_ERROR);
+    return NULL;
+  }
  *skipped = false;
  char* check_path = receive_str(fd);
  if (check_path == NULL) {
-    send_status(fd, STATUS_ERROR);
    return NULL;
  }

@@ -502,10 +502,14 @@ File* receive_incremental_check(int fd, const Config* config, bool* skipped) {
  if (!receive_n_data(fd, &check_size, sizeof(check_size)) ||
      !receive_n_data(fd, &check_mtime, sizeof(check_mtime))) {
    free(check_path);
-    send_status(fd, STATUS_ERROR);
    return NULL;
  }
  if (config->checksum && !receive_n_data(fd, &check_checksum, sizeof(check_checksum))) {
+    free(check_path);
+    return NULL;
+  }
+
+  if (check_size > MAX_RECEIVE_FILE_SIZE) {
    free(check_path);
    send_status(fd, STATUS_ERROR);
    return NULL;
@@ -514,27 +518,29 @@ File* receive_incremental_check(int fd, const Config* config, bool* skipped) {
  if (has_path_traversal(check_path)) {
    log_message(LOG_LEVEL_ERROR, "Path traversal detected: %s", check_path);
    free(check_path);
-    send_status(fd, STATUS_ERROR);
    return NULL;
  }

  char* full_path = path_cat(config->receive_root_directory, check_path);
+  if (!full_path) {
+    free(check_path);
+    send_status(fd, STATUS_ERROR);
+    return NULL;
+  }
  struct stat st;
  bool has_old_file = false;
  int old_fd = -1;
-  if (full_path) {
-    char* leaf = NULL;
-    int parent_fd = file_store_open_secure_parent(full_path, &leaf);
-    if (parent_fd >= 0) {
-      old_fd = openat(parent_fd, leaf, O_RDONLY | O_CLOEXEC | O_NOFOLLOW);
-      free(leaf);
-      close(parent_fd);
-      has_old_file = old_fd >= 0 && fstat(old_fd, &st) == 0 && S_ISREG(st.st_mode);
-    }
+  char* leaf = NULL;
+  int parent_fd = open_secure_parent(full_path, &leaf, false);
+  if (parent_fd >= 0) {
+    old_fd = openat(parent_fd, leaf, O_RDONLY | O_CLOEXEC | O_NOFOLLOW);
+    free(leaf);
+    close(parent_fd);
+    has_old_file = old_fd >= 0 && fstat(old_fd, &st) == 0 && S_ISREG(st.st_mode);
  }
  unsigned long long old_size = has_old_file ? (unsigned long long)st.st_size : 0;
  void* old_data = NULL;
-  if (has_old_file && old_size > 0) {
+  if (has_old_file && old_size > 0 && old_size <= MAX_RECEIVE_FILE_SIZE && old_size <= SIZE_MAX) {
    old_data = malloc((size_t)old_size);
    if (old_data) {
      size_t got = 0;
@@ -578,17 +584,24 @@ File* receive_incremental_check(int fd, const Config* config, bool* skipped) {
    return NULL;
  }

-  bool try_delta = config->use_delta && has_old_file &&
+  bool try_delta = config->use_delta && has_old_file && old_data != NULL &&
                   delta_should_attempt(old_size, check_size, config->delta_max_file_size);

  if (try_delta) {
-    File* delta_file = receive_delta_file(fd, config, check_path, old_data, old_size);
+    bool delta_failed = false;
+    File* delta_file =
+        receive_delta_file(fd, config, check_path, old_data, old_size, &delta_failed);
    old_data = NULL; /* receive_delta_file consumes the snapshot on every path */
    if (delta_file) {
      free(full_path);
      free(check_path);
      return delta_file;
    }
+    if (delta_failed) {
+      free(full_path);
+      free(check_path);
+      return NULL;
+    }
    free(old_data);
    old_data = NULL;
    try_delta = false;
@@ -606,7 +619,6 @@ File* receive_incremental_check(int fd, const Config* config, bool* skipped) {
  free(check_path);
  free(full_path);
  if (file == NULL) {
-    send_status(fd, STATUS_ERROR);
    return NULL;
  }

@@ -615,22 +627,25 @@ File* receive_incremental_check(int fd, const Config* config, bool* skipped) {
    file->metadata = metadata_receive(fd, &meta_ok);
    if (!meta_ok) {
      file_destroy(file);
-      send_status(fd, STATUS_ERROR);
      return NULL;
    }
  }

-  Data* file_data = receive_data(fd);
+  Data* file_data = receive_data_limited(fd, MAX_RECEIVE_FILE_SIZE);
  if (file_data == NULL) {
    file_destroy(file);
-    send_status(fd, STATUS_ERROR);
    return NULL;
  }

  if (config->use_compression) {
-    Data* uncompressed = data_decompress(file_data);
+    Data* uncompressed = data_decompress_limited(file_data, MAX_RECEIVE_FILE_SIZE);
    data_destroy(file_data);
    if (uncompressed == NULL) {
+      file_destroy(file);
+      return NULL;
+    }
+    if (uncompressed->size > MAX_FILE_DATA_SIZE) {
+      data_destroy(uncompressed);
      file_destroy(file);
      send_status(fd, STATUS_ERROR);
      return NULL;
@@ -643,11 +658,180 @@ File* receive_incremental_check(int fd, const Config* config, bool* skipped) {
  return file;
 }

+static int open_secure_parent(const char* path, char** leaf_out, bool create_dirs) {
+  char* copy = str_dup(path);
+  if (!copy)
+    return -1;
+  char* parent = dirname(copy);
+  const char* slash = strrchr(path, '/');
+  char* leaf = str_dup(slash ? slash + 1 : path);
+  if (!leaf) {
+    free(copy);
+    return -1;
+  }
+  int fd;
+  if (authorized_root_fd >= 0) {
+    if (!authorized_root_path || path[0] != '/' ||
+        !path_is_within_root(authorized_root_path, path)) {
+      free(copy);
+      free(leaf);
+      return -1;
+    }
+    fd = dup(authorized_root_fd);
+    if (fd < 0) {
+      free(copy);
+      free(leaf);
+      return -1;
+    }
+    size_t root_len = strlen(authorized_root_path);
+    char* relative = str_dup(path + root_len);
+    if (!relative) {
+      free(copy);
+      free(leaf);
+      close(fd);
+      return -1;
+    }
+    free(copy);
+    copy = relative;
+    parent = dirname(copy);
+  } else {
+    fd = (parent[0] == '/') ? open("/", O_RDONLY | O_DIRECTORY | O_CLOEXEC)
+                            : open(".", O_RDONLY | O_DIRECTORY | O_CLOEXEC);
+  }
+  if (fd < 0) {
+    free(copy);
+    free(leaf);
+    return -1;
+  }
+  char* save = NULL;
+  char* component = strtok_r(parent, "/", &save);
+  while (component) {
+    if (strcmp(component, "..") == 0) {
+      close(fd);
+      free(copy);
+      free(leaf);
+      return -1;
+    }
+    if (strcmp(component, ".") != 0) {
+      int next = openat(fd, component, O_RDONLY | O_DIRECTORY | O_NOFOLLOW | O_CLOEXEC);
+      if (create_dirs && next < 0 && errno == ENOENT) {
+        if (mkdirat(fd, component, 0755) == 0 || errno == EEXIST)
+          next = openat(fd, component, O_RDONLY | O_DIRECTORY | O_NOFOLLOW | O_CLOEXEC);
+      }
+      if (next < 0) {
+        close(fd);
+        free(copy);
+        free(leaf);
+        return -1;
+      }
+      close(fd);
+      fd = next;
+    }
+    component = strtok_r(NULL, "/", &save);
+  }
+  free(copy);
+  *leaf_out = leaf;
+  return fd;
+}
+
+static bool ensure_directory_secure(const char* path) {
+  char* leaf = NULL;
+  int parent_fd = open_secure_parent(path, &leaf, true);
+  if (parent_fd < 0)
+    return false;
+
+  int dir_fd = openat(parent_fd, leaf, O_RDONLY | O_DIRECTORY | O_NOFOLLOW | O_CLOEXEC);
+  if (dir_fd < 0 && errno == ENOENT) {
+    if (mkdirat(parent_fd, leaf, 0755) == 0 || errno == EEXIST)
+      dir_fd = openat(parent_fd, leaf, O_RDONLY | O_DIRECTORY | O_NOFOLLOW | O_CLOEXEC);
+  }
+  bool ok = dir_fd >= 0;
+  if (dir_fd >= 0)
+    close(dir_fd);
+  close(parent_fd);
+  free(leaf);
+  return ok;
+}
+
+static bool rename_secure(const char* old_path, const char* new_path) {
+  char *old_leaf = NULL, *new_leaf = NULL;
+  int old_parent = open_secure_parent(old_path, &old_leaf, false);
+  int new_parent = open_secure_parent(new_path, &new_leaf, true);
+  bool ok = old_parent >= 0 && new_parent >= 0 &&
+            renameat(old_parent, old_leaf, new_parent, new_leaf) == 0;
+  if (old_parent >= 0)
+    close(old_parent);
+  if (new_parent >= 0)
+    close(new_parent);
+  free(old_leaf);
+  free(new_leaf);
+  return ok;
+}
+
+static bool write_all(int fd, const void* data, unsigned long long size) {
+  const unsigned char* p = data;
+  unsigned long long done = 0;
+  while (done < size) {
+    ssize_t n = write(fd, p + done, (size_t)(size - done));
+    if (n < 0 && errno == EINTR)
+      continue;
+    if (n <= 0)
+      return false;
+    done += (unsigned long long)n;
+  }
+  return true;
+}
+
+static bool to_disk_secure(const char* path, const void* data, unsigned long long data_size,
+                           bool inplace, bool sparse, const FileMetadata* metadata) {
+  char* leaf = NULL;
+  int dirfd = open_secure_parent(path, &leaf, true);
+  if (dirfd < 0)
+    return false;
+  int fd = -1;
+  bool ok = false;
+  if (inplace) {
+    fd = openat(dirfd, leaf, O_WRONLY | O_CREAT | O_TRUNC | O_CLOEXEC | O_NOFOLLOW, 0644);
+    if (fd >= 0) {
+      if (!sparse || data_size == 0 || ftruncate(fd, (off_t)data_size) == 0)
+        ok = write_all(fd, data, data_size);
+      if (ok && metadata)
+        ok = file_restore_metadata_fd(fd, metadata);
+    }
+  } else {
+    char tmp[NAME_MAX];
+    for (unsigned int i = 0; i < 100 && !ok; ++i) {
+      snprintf(tmp, sizeof(tmp), ".%s.tmp.%ld.%u", leaf, (long)getpid(), i);
+      fd = openat(dirfd, tmp, O_WRONLY | O_CREAT | O_EXCL | O_CLOEXEC | O_NOFOLLOW, 0600);
+      if (fd < 0)
+        continue;
+      if (sparse && data_size > 0)
+        ok = ftruncate(fd, (off_t)data_size) == 0;
+      if (ok || (!sparse || data_size == 0))
+        ok = write_all(fd, data, data_size);
+      if (ok && metadata)
+        ok = file_restore_metadata_fd(fd, metadata);
+      if (close(fd) != 0)
+        ok = false;
+      fd = -1;
+      if (ok && renameat(dirfd, tmp, dirfd, leaf) != 0)
+        ok = false;
+      if (!ok)
+        unlinkat(dirfd, tmp, 0);
+    }
+  }
+  if (fd >= 0)
+    close(fd);
+  close(dirfd);
+  free(leaf);
+  return ok;
+}
+
 bool to_disk(const char* path, const void* data, unsigned long long data_size, bool inplace,
             bool sparse) {
  if (!path || (!data && data_size != 0) || has_path_traversal(path))
    return false;
-  return file_store_write_secure(path, data, data_size, inplace, sparse, NULL);
+  return to_disk_secure(path, data, data_size, inplace, sparse, NULL);
 }

 bool file_send_sendfile(File* file, int file_descriptor, bool use_metadata, int compression_level,
@@ -759,18 +943,23 @@ File* file_receive(const Config* config, int file_descriptor) {
      return NULL;
    }
  }
-  Data* file_data = receive_data(file_descriptor);
+  Data* file_data = receive_data_limited(file_descriptor, MAX_RECEIVE_FILE_SIZE);
  if (file_data == NULL) {
    file_destroy(file);
    return NULL;
  }
-  if (config->use_compression) {
-    Data* file_data_uncompressed = data_decompress(file_data);
+  if (config->use_compression && !compression_should_skip(file->path)) {
+    Data* file_data_uncompressed = data_decompress_limited(file_data, MAX_RECEIVE_FILE_SIZE);
    data_destroy(file_data);
    if (file_data_uncompressed == NULL) {
      file_destroy(file);
      return NULL;
    }
+    if (file_data_uncompressed->size > MAX_FILE_DATA_SIZE) {
+      data_destroy(file_data_uncompressed);
+      file_destroy(file);
+      return NULL;
+    }
    file_data = file_data_uncompressed;
  }
  data_destroy(file->data);
@@ -779,6 +968,8 @@ File* file_receive(const Config* config, int file_descriptor) {
 }

 size_t file_content_to_buffer(File* file) {
+  if (!file || !file->path || !file->data || (!file->data->data && file->data->size != 0))
+    return 0;
  FILE* file_pointer = fopen(file->path, "rb");
  if (file_pointer == NULL) {
    perror("Could not open the file!");
@@ -795,16 +986,26 @@ size_t file_content_to_buffer(File* file) {
 }

 int receive_manifest(int fd, const Config* config, int* next_status) {
+  if (!config) {
+    send_status(fd, STATUS_ERROR);
+    return -1;
+  }
  int received_status = STATUS_ERROR;
  int* status_out = next_status ? next_status : &received_status;
  int count;
-  if (!receive_int(fd, &count))
+  if (!receive_int(fd, &count)) {
+    send_status(fd, STATUS_ERROR);
    return -1;
-  if (count < 0 || count > MAX_MANIFEST_ENTRIES)
+  }
+  if (count < 0 || count > MAX_MANIFEST_ENTRIES) {
+    send_status(fd, STATUS_ERROR);
    return -1;
+  }
  ArrayList* manifest = array_list_create(free);
-  if (!manifest)
+  if (!manifest) {
+    send_status(fd, STATUS_ERROR);
    return -1;
+  }
  size_t manifest_bytes = 0;
  for (int i = 0; i < count; i++) {
    char* s = receive_str(fd);
@@ -814,21 +1015,28 @@ int receive_manifest(int fd, const Config* config, int* next_status) {
        (manifest_bytes += entry_size) > MAX_MANIFEST_BYTES || !array_list_add(manifest, s)) {
      free(s);
      array_list_delete(manifest);
+      send_status(fd, STATUS_ERROR);
      return -1;
    }
  }
  if (!receive_status(fd, status_out)) {
    array_list_delete(manifest);
+    send_status(fd, STATUS_ERROR);
    return -1;
  }
  /* Deletion is a commit operation: never perform it until the sender has
     completed the manifest frame successfully. */
  if (*status_out != STATUS_FINISHED || !config->use_delete) {
    array_list_delete(manifest);
+    if (*status_out != STATUS_FINISHED)
+      send_status(fd, STATUS_ERROR);
    return *status_out == STATUS_FINISHED ? 0 : -1;
  }
  fprintf(stderr, "Deleting files not in manifest...\n");
-  bool deletion_ok = delete_extras(config->receive_root_directory, manifest);
+  bool deletion_ok =
+      delete_extras_limited(config->receive_root_directory, manifest, MAX_SERVER_DELETE_COUNT);
  array_list_delete(manifest);
+  if (!deletion_ok)
+    send_status(fd, STATUS_ERROR);
  return deletion_ok ? 0 : -1;
 }
@@ -38,8 +38,11 @@ void file_metadata_destroy(void* metadata);
 bool to_disk(const char* path, const void* data, unsigned long long data_size, bool inplace,
             bool sparse);
 bool file_save_to_disk(const char* root_directory, const File* file, const Config* config);
-void file_set_authorized_root(int fd, const char* canonical_path);
+/* A configured fd without a canonical identity deliberately rejects paths. */
+bool file_set_authorized_root(int fd, const char* canonical_path);
 File* receive_incremental_check(int fd, const Config* config, bool* skipped);
+bool file_path_exists_secure(const char* path);
+bool file_stat_secure(const char* path, struct stat* st);
 int receive_manifest(int fd, const Config* config, int* next_status);

 #endif
@@ -20,10 +20,18 @@ static bool path_is_within_root(const char* root, const char* path) {
         (path[root_length] == '\0' || path[root_length] == '/');
 }

-void file_store_set_authorized_root(int fd, const char* canonical_path) {
-  authorized_root_fd = fd;
+bool file_store_set_authorized_root(int fd, const char* canonical_path) {
+  char* new_path = canonical_path ? str_dup(canonical_path) : NULL;
+  if (canonical_path && !new_path) {
+    authorized_root_fd = -1;
+    free(authorized_root_path);
+    authorized_root_path = NULL;
+    return false;
+  }
  free(authorized_root_path);
-  authorized_root_path = canonical_path ? str_dup(canonical_path) : NULL;
+  authorized_root_path = new_path;
+  authorized_root_fd = fd;
+  return true;
 }

 int file_store_open_secure_parent(const char* path, char** leaf_out) {
@@ -38,9 +46,19 @@ int file_store_open_secure_parent(const char* path, char** leaf_out) {
    return -1;
  }
  int fd;
-  if (authorized_root_fd >= 0 && authorized_root_path && path[0] == '/' &&
-      path_is_within_root(authorized_root_path, path)) {
+  if (authorized_root_fd >= 0) {
+    if (!authorized_root_path || path[0] != '/' ||
+        !path_is_within_root(authorized_root_path, path)) {
+      free(copy);
+      free(leaf);
+      return -1;
+    }
    fd = dup(authorized_root_fd);
+    if (fd < 0) {
+      free(copy);
+      free(leaf);
+      return -1;
+    }
    size_t root_length = strlen(authorized_root_path);
    char* relative = str_dup(path + root_length);
    if (!relative) {
@@ -64,10 +82,18 @@ int file_store_open_secure_parent(const char* path, char** leaf_out) {
  char* save = NULL;
  char* component = strtok_r(parent, "/", &save);
  while (component) {
-    if (strcmp(component, ".") != 0 && strcmp(component, "..") != 0) {
+    if (strcmp(component, "..") == 0) {
+      close(fd);
+      free(copy);
+      free(leaf);
+      return -1;
+    }
+    if (strcmp(component, ".") != 0) {
      int next = openat(fd, component, O_RDONLY | O_DIRECTORY | O_NOFOLLOW | O_CLOEXEC);
-      if (next < 0 && errno == ENOENT && mkdirat(fd, component, 0755) == 0)
-        next = openat(fd, component, O_RDONLY | O_DIRECTORY | O_NOFOLLOW | O_CLOEXEC);
+      if (next < 0 && errno == ENOENT) {
+        if (mkdirat(fd, component, 0755) == 0 || errno == EEXIST)
+          next = openat(fd, component, O_RDONLY | O_DIRECTORY | O_NOFOLLOW | O_CLOEXEC);
+      }
      if (next < 0) {
        close(fd);
        free(copy);
@@ -130,9 +156,20 @@ bool file_store_write_secure(const char* path, const void* data, unsigned long l
        ok = file_restore_metadata_fd(fd, metadata);
    }
  } else {
-    char tmp[NAME_MAX];
+    int tmp_size = snprintf(NULL, 0, ".%s.tmp.%ld.%u", leaf, (long)getpid(), 99U);
+    if (tmp_size < 0) {
+      close(dirfd);
+      free(leaf);
+      return false;
+    }
+    char* tmp = malloc((size_t)tmp_size + 1);
+    if (!tmp) {
+      close(dirfd);
+      free(leaf);
+      return false;
+    }
    for (unsigned int i = 0; i < 100 && !ok; ++i) {
-      snprintf(tmp, sizeof(tmp), ".%s.tmp.%ld.%u", leaf, (long)getpid(), i);
+      snprintf(tmp, (size_t)tmp_size + 1, ".%s.tmp.%ld.%u", leaf, (long)getpid(), i);
      fd = openat(dirfd, tmp, O_WRONLY | O_CREAT | O_EXCL | O_CLOEXEC | O_NOFOLLOW, 0600);
      if (fd < 0)
        continue;
@@ -150,6 +187,7 @@ bool file_store_write_secure(const char* path, const void* data, unsigned long l
      if (!ok)
        unlinkat(dirfd, tmp, 0);
    }
+    free(tmp);
  }
  if (fd >= 0)
    close(fd);
@@ -4,7 +4,7 @@
 #include "file.h"
 #include <stdbool.h>

-void file_store_set_authorized_root(int fd, const char* canonical_path);
+bool file_store_set_authorized_root(int fd, const char* canonical_path);
 int file_store_open_secure_parent(const char* path, char** leaf_out);
 bool file_store_rename_secure(const char* old_path, const char* new_path);
 bool file_store_write_secure(const char* path, const void* data, unsigned long long data_size,
@@ -51,6 +51,8 @@ FileMetadata* metadata_from_buf(char** buf) {
  int32_t present;
  memcpy(&present, *buf, sizeof(present));
  *buf += sizeof(present);
+  if (present != 0 && present != 1)
+    return NULL;
  if (!present)
    return NULL;
  FileMetadata* m = malloc(sizeof(FileMetadata));
@@ -76,10 +78,15 @@ FileMetadata* metadata_from_buf(char** buf) {
  memcpy(&mtime_nsec, *buf, sizeof(mtime_nsec));
  *buf += sizeof(mtime_nsec);
  m->mtime_nsec = (long)mtime_nsec;
+  if (present != 1 || mtime_nsec < 0 || mtime_nsec >= 1000000000LL || mode < 0 || uid < 0 ||
+      gid < 0) {
+    free(m);
+    return NULL;
+  }
  return m;
 }

-bool metadata_send(int file_descriptor, FileMetadata* m) {
+bool metadata_send(int file_descriptor, const FileMetadata* m) {
  if (m == NULL) {
    int32_t zero = 0;
    return send_n_data(file_descriptor, &zero, sizeof(zero));
@@ -175,7 +182,8 @@ FileMetadata* metadata_receive(int file_descriptor, int* ok) {
 void file_restore_metadata(const char* path, const FileMetadata* metadata) {
  if (metadata == NULL)
    return;
-  if (chmod(path, metadata->mode & 07777 & ~(S_ISUID | S_ISGID)) != 0)
+  mode_t safe_mode = metadata->mode & 0777 & ~(S_IWGRP | S_IWOTH);
+  if (chmod(path, safe_mode) != 0)
    log_message(LOG_LEVEL_WARNING, "Failed to chmod %s: %s", path, strerror(errno));
  /* Never apply client-supplied ownership.  The descriptor API below is the
     receiver write path; retain this legacy API only for compatibility. */
@@ -192,7 +200,8 @@ bool file_restore_metadata_fd(int fd, const FileMetadata* metadata) {
  if (fd < 0 || metadata == NULL)
    return metadata == NULL;
  bool ok = true;
-  if (fchmod(fd, metadata->mode & 07777 & ~(S_ISUID | S_ISGID)) != 0)
+  mode_t safe_mode = metadata->mode & 0777 & ~(S_IWGRP | S_IWOTH);
+  if (fchmod(fd, safe_mode) != 0)
    ok = false;
  /* Client uid/gid values are deliberately not authoritative. */
  struct timespec times[2] = {{.tv_sec = 0, .tv_nsec = UTIME_OMIT},
@@ -27,7 +27,7 @@

 void metadata_to_buf(char** buf, const FileMetadata* m);
 FileMetadata* metadata_from_buf(char** buf);
-bool metadata_send(int file_descriptor, FileMetadata* m);
+bool metadata_send(int file_descriptor, const FileMetadata* m);
 FileMetadata* metadata_receive(int file_descriptor, int* ok);
 void file_restore_metadata(const char* path, const FileMetadata* metadata);
 bool file_restore_metadata_fd(int fd, const FileMetadata* metadata);
@@ -183,6 +183,15 @@ int write_thread(void* pipeline_context) {
  bool save_to_disk = context->config->save_to_disk;
  char* root_directory = str_dup(context->config->receive_root_directory);
  mtx_unlock(&context->mutex);
+  if (save_to_disk && !root_directory) {
+    mtx_lock(&context->mutex);
+    atomic_store(&context->cancelled, true);
+    context->receiver_done = true;
+    cnd_broadcast(&context->condition_not_full);
+    cnd_broadcast(&context->condition_not_empty);
+    mtx_unlock(&context->mutex);
+    return thrd_error;
+  }

  while (true) {
    File* file =
@@ -13,21 +13,27 @@

 #define RECEIVE_TIMEOUT_SEC 60 /* 60 second per-message timeout */
 #define SEND_TIMEOUT_SEC 60
-#define MAX_CONNECTION_MEMORY (1024ULL * 1024 * 1024) /* 1 GB total per connection */
+#define MAX_CONNECTION_MEMORY (256ULL * 1024 * 1024) /* bounded cumulative receive budget */

 static __thread int io_read_fd = -1;
 static __thread int io_write_fd = -1;
 static __thread SSL* io_ssl;
 static __thread ProtocolSession* bound_session;
+static __thread ProtocolSession legacy_io_session = {.read_fd = -1, .write_fd = -1};

 static unsigned long long io_bwlimit = 0;
-static long long bw_tokens = 0;
-static struct timespec bw_last_refill = {0, 0};
 static mtx_t bw_mutex;
 static once_flag bw_mutex_once = ONCE_FLAG_INIT;

-static __thread unsigned long long total_allocated_bytes = 0;
+static unsigned long long global_bwlimit(void);

+void protocol_release_memory(size_t charge) {
+  ProtocolSession* session = bound_session ? bound_session : &legacy_io_session;
+  if ((unsigned long long)charge >= session->total_allocated_bytes)
+    session->total_allocated_bytes = 0;
+  else
+    session->total_allocated_bytes -= charge;
+}
 void io_set_fds(int read_fd, int write_fd) {
  bound_session = NULL;
  io_read_fd = read_fd;
@@ -35,7 +41,11 @@ void io_set_fds(int read_fd, int write_fd) {
  /* A descriptor switch starts a new transport; never reuse a TLS object
     belonging to a previous connection or test pipe. */
  io_ssl = NULL;
-  total_allocated_bytes = 0;
+  legacy_io_session.read_fd = read_fd;
+  legacy_io_session.write_fd = write_fd;
+  legacy_io_session.ssl = NULL;
+  legacy_io_session.total_allocated_bytes = 0;
+  protocol_session_set_bwlimit(&legacy_io_session, global_bwlimit());
 }

 void protocol_session_init(ProtocolSession* session, int read_fd, int write_fd) {
@@ -44,8 +54,7 @@ void protocol_session_init(ProtocolSession* session, int read_fd, int write_fd)
  memset(session, 0, sizeof(*session));
  session->read_fd = read_fd;
  session->write_fd = write_fd;
-  if (io_bwlimit)
-    protocol_session_set_bwlimit(session, io_bwlimit);
+  protocol_session_set_bwlimit(session, global_bwlimit());
 }

 void protocol_session_bind(ProtocolSession* session) {
@@ -65,20 +74,29 @@ static void bw_mutex_init(void) {
  mtx_init(&bw_mutex, mtx_plain);
 }

+static unsigned long long global_bwlimit(void) {
+  unsigned long long limit;
+  call_once(&bw_mutex_once, bw_mutex_init);
+  mtx_lock(&bw_mutex);
+  limit = io_bwlimit;
+  mtx_unlock(&bw_mutex);
+  return limit;
+}
+
 void io_set_bwlimit(unsigned long long bytes_per_sec) {
  call_once(&bw_mutex_once, bw_mutex_init);
  mtx_lock(&bw_mutex);
-  io_bwlimit = bytes_per_sec;
-  bw_tokens = (long long)io_bwlimit;
-  clock_gettime(CLOCK_MONOTONIC, &bw_last_refill);
+  io_bwlimit =
+      bytes_per_sec > (unsigned long long)LLONG_MAX ? (unsigned long long)LLONG_MAX : bytes_per_sec;
  mtx_unlock(&bw_mutex);
 }

 void protocol_session_set_bwlimit(ProtocolSession* session, unsigned long long bytes_per_sec) {
  if (!session)
    return;
-  session->bwlimit = bytes_per_sec;
-  session->bw_tokens = (long long)bytes_per_sec;
+  session->bwlimit =
+      bytes_per_sec > (unsigned long long)LLONG_MAX ? (unsigned long long)LLONG_MAX : bytes_per_sec;
+  session->bw_tokens = (long long)session->bwlimit;
  struct timespec now;
  clock_gettime(CLOCK_MONOTONIC, &now);
  session->bw_last_refill_sec = now.tv_sec;
@@ -126,32 +144,30 @@ SSL* io_get_ssl(void) {
  return io_ssl;
 }

-static ProtocolSession* legacy_session(void) {
-  static __thread ProtocolSession session;
+static ProtocolSession* legacy_session(int read_fd, int write_fd) {
  if (bound_session)
    return bound_session;
-  session.read_fd = io_read_fd;
-  session.write_fd = io_write_fd;
-  session.ssl = io_ssl;
-  session.bwlimit = io_bwlimit;
-  session.bw_tokens = (unsigned long long)(bw_tokens < 0 ? 0 : bw_tokens);
-  session.bw_last_refill_sec = bw_last_refill.tv_sec;
-  session.bw_last_refill_nsec = bw_last_refill.tv_nsec;
-  return &session;
+  int target_read_fd = io_read_fd != -1 ? io_read_fd : read_fd;
+  int target_write_fd = io_write_fd != -1 ? io_write_fd : write_fd;
+  if (legacy_io_session.read_fd != target_read_fd ||
+      legacy_io_session.write_fd != target_write_fd) {
+    legacy_io_session.read_fd = target_read_fd;
+    legacy_io_session.write_fd = target_write_fd;
+    legacy_io_session.total_allocated_bytes = 0;
+    protocol_session_set_bwlimit(&legacy_io_session, global_bwlimit());
+  } else if (legacy_io_session.bwlimit != global_bwlimit()) {
+    protocol_session_set_bwlimit(&legacy_io_session, global_bwlimit());
+  }
+  legacy_io_session.ssl = io_ssl;
+  return &legacy_io_session;
 }

 bool send_n_data(int file_descriptor, const void* data, size_t data_size) {
-  ProtocolSession* session = legacy_session();
-  if (session->write_fd == -1)
-    session->write_fd = file_descriptor;
-  return protocol_send_n_data(session, data, data_size);
+  return protocol_send_n_data(legacy_session(-1, file_descriptor), data, data_size);
 }

 bool receive_n_data(int file_descriptor, void* data, size_t data_size) {
-  ProtocolSession* session = legacy_session();
-  if (session->read_fd == -1)
-    session->read_fd = file_descriptor;
-  return protocol_receive_n_data(session, data, data_size);
+  return protocol_receive_n_data(legacy_session(file_descriptor, -1), data, data_size);
 }

 static int deadline_remaining_ms(const struct timespec* deadline) {
@@ -166,6 +182,8 @@ static int deadline_remaining_ms(const struct timespec* deadline) {
 }

 bool protocol_send_n_data(ProtocolSession* session, const void* data, size_t data_size) {
+  if (!data && data_size != 0)
+    return false;
  log_message(LOG_LEVEL_DEBUG, "    Sending n Data: %zu", data_size);
  if (!session)
    return false;
@@ -207,6 +225,8 @@ bool protocol_send_n_data(ProtocolSession* session, const void* data, size_t dat
    }
    bw_throttle_session(session, (size_t)bytes_send);
    total_bytes_send += bytes_send;
+    if (session->ssl)
+      wait_events = POLLOUT;
  }
  log_message(LOG_LEVEL_DEBUG, "    Send n Data: %zu", total_bytes_send);
  return true;
@@ -225,20 +245,22 @@ bool protocol_receive_n_data(ProtocolSession* session, void* data, size_t data_s
  size_t total_bytes_received = 0;
  short wait_events = POLLIN;
  while (total_bytes_received < data_size) {
-    struct pollfd pfd = {.fd = fd, .events = wait_events};
-    int poll_result = poll(&pfd, 1, deadline_remaining_ms(&deadline));
-    if (poll_result == 0) {
-      log_message(LOG_LEVEL_ERROR, "Receive timeout after %ds", RECEIVE_TIMEOUT_SEC);
-      return false;
+    if (!session->ssl || SSL_pending(session->ssl) == 0) {
+      struct pollfd pfd = {.fd = fd, .events = wait_events};
+      int poll_result = poll(&pfd, 1, deadline_remaining_ms(&deadline));
+      if (poll_result == 0) {
+        log_message(LOG_LEVEL_ERROR, "Receive timeout after %ds", RECEIVE_TIMEOUT_SEC);
+        return false;
+      }
+      if (poll_result < 0) {
+        if (errno == EINTR)
+          continue;
+        return false;
+      }
+      /* POLLHUP may accompany the final readable bytes on pipes/sockets. */
+      if (pfd.revents & (POLLERR | POLLNVAL))
+        return false;
    }
-    if (poll_result < 0) {
-      if (errno == EINTR)
-        continue;
-      return false;
-    }
-    /* POLLHUP may accompany the final readable bytes on pipes/sockets. */
-    if (pfd.revents & (POLLERR | POLLNVAL))
-      return false;

    ssize_t bytes_received;
    if (session->ssl)
@@ -261,7 +283,9 @@ bool protocol_receive_n_data(ProtocolSession* session, void* data, size_t data_s
        log_message(LOG_LEVEL_ERROR, "Could not receive bytes");
      return false;
    }
-    total_bytes_received += bytes_received;
+    total_bytes_received += (size_t)bytes_received;
+    if (session->ssl)
+      wait_events = POLLIN;
  }
  log_message(LOG_LEVEL_DEBUG, "    Received n Data: %zu", total_bytes_received);
  return true;
@@ -325,6 +349,11 @@ char* protocol_receive_str(ProtocolSession* session) {
    free(data);
    return NULL;
  }
+  if (memchr(data, '\0', size) != NULL) {
+    free(data);
+    log_message(LOG_LEVEL_ERROR, "Received string contains an embedded NUL");
+    return NULL;
+  }
  data[size] = '\0';
  session->total_allocated_bytes += size + 1;
  log_message(LOG_LEVEL_DEBUG, "Received String: %s", data);
@@ -332,6 +361,10 @@ char* protocol_receive_str(ProtocolSession* session) {
 }

 bool protocol_send_data(ProtocolSession* session, const Data* data) {
+  if (!data || (!data->data && data->size != 0))
+    return false;
+  if (!session)
+    return false;
  unsigned long long data_size = data->size;
  if (!protocol_send_n_data(session, &data_size, sizeof(unsigned long long)))
    return false;
@@ -341,11 +374,13 @@ bool protocol_send_data(ProtocolSession* session, const Data* data) {
  return true;
 }

-Data* protocol_receive_data(ProtocolSession* session) {
+Data* protocol_receive_data_limited(ProtocolSession* session, unsigned long long maximum_size) {
+  if (!session)
+    return NULL;
  unsigned long long size = 0;
  if (!protocol_receive_n_data(session, &size, sizeof(unsigned long long)))
    return NULL;
-  if (size > MAX_DATA_PAYLOAD_SIZE) {
+  if (size > MAX_DATA_PAYLOAD_SIZE || size > maximum_size) {
    log_message(LOG_LEVEL_ERROR, "Data size %llu exceeds maximum %llu", size,
                (unsigned long long)MAX_DATA_PAYLOAD_SIZE);
    return NULL;
@@ -368,12 +403,17 @@ Data* protocol_receive_data(ProtocolSession* session) {
  log_message(LOG_LEVEL_DEBUG, "Received %lld data", size);
  Data* result = data_create(data, (size_t)size);
  if (!result) {
-    free(data);
    session->total_allocated_bytes -= allocation_size;
+    return NULL;
  }
+  result->protocol_charge = allocation_size;
  return result;
 }

+Data* protocol_receive_data(ProtocolSession* session) {
+  return protocol_receive_data_limited(session, MAX_DATA_PAYLOAD_SIZE);
+}
+
 bool protocol_send_int(ProtocolSession* session, int data) {
  if (!protocol_send_n_data(session, &data, sizeof(int)))
    return false;
@@ -403,34 +443,29 @@ bool protocol_receive_status(ProtocolSession* session, Status* status) {
 }

 bool send_str(int fd, const char* data) {
-  (void)fd;
-  return protocol_send_str(legacy_session(), data);
+  return protocol_send_str(legacy_session(-1, fd), data);
 }
 char* receive_str(int fd) {
-  (void)fd;
-  return protocol_receive_str(legacy_session());
+  return protocol_receive_str(legacy_session(fd, -1));
 }
 bool send_data(int fd, const Data* data) {
-  (void)fd;
-  return protocol_send_data(legacy_session(), data);
+  return protocol_send_data(legacy_session(-1, fd), data);
 }
 Data* receive_data(int fd) {
-  (void)fd;
-  return protocol_receive_data(legacy_session());
+  return protocol_receive_data_limited(legacy_session(fd, -1), MAX_DATA_PAYLOAD_SIZE);
+}
+Data* receive_data_limited(int fd, unsigned long long maximum_size) {
+  return protocol_receive_data_limited(legacy_session(fd, -1), maximum_size);
 }
 bool send_int(int fd, int data) {
-  (void)fd;
-  return protocol_send_int(legacy_session(), data);
+  return protocol_send_int(legacy_session(-1, fd), data);
 }
 bool receive_int(int fd, int* data) {
-  (void)fd;
-  return protocol_receive_int(legacy_session(), data);
+  return protocol_receive_int(legacy_session(fd, -1), data);
 }
 bool send_status(int fd, Status status) {
-  (void)fd;
-  return protocol_send_status(legacy_session(), status);
+  return protocol_send_status(legacy_session(-1, fd), status);
 }
 bool receive_status(int fd, Status* status) {
-  (void)fd;
-  return protocol_receive_status(legacy_session(), status);
+  return protocol_receive_status(legacy_session(fd, -1), status);
 }
@@ -10,6 +10,8 @@

 /* Maximum allowed data payload size for receive_data (100 MB) */
 #define MAX_DATA_PAYLOAD_SIZE (100ULL * 1024 * 1024)
+/* Maximum uncompressed file payload accepted by the receiver. */
+#define MAX_RECEIVE_FILE_SIZE (64ULL * 1024 * 1024)

 /* Maximum chunk size (64 MB) — prevents unbounded allocation from the wire */
 #define MAX_CHUNK_SIZE (64ULL * 1024 * 1024)
@@ -69,6 +71,7 @@ bool protocol_send_str(ProtocolSession* session, const char* data);
 char* protocol_receive_str(ProtocolSession* session);
 bool protocol_send_data(ProtocolSession* session, const Data* data);
 Data* protocol_receive_data(ProtocolSession* session);
+Data* protocol_receive_data_limited(ProtocolSession* session, unsigned long long maximum_size);
 bool protocol_send_int(ProtocolSession* session, int data);
 bool protocol_receive_int(ProtocolSession* session, int* data);
 bool protocol_send_status(ProtocolSession* session, Status status);
@@ -80,6 +83,7 @@ bool send_str(int file_descriptor, const char* data);
 char* receive_str(int file_descriptor);
 bool send_data(int file_descriptor, const Data* data);
 Data* receive_data(int file_descriptor);
+Data* receive_data_limited(int file_descriptor, unsigned long long maximum_size);
 bool send_int(int file_descriptor, int data);
 bool receive_int(int file_descriptor, int* data);
 bool send_status(int file_descriptor, Status status);
@@ -151,6 +151,10 @@ int tcp_get_contimeout_sec(void) {
  return g_contimeout_sec;
 }

+int tcp_get_timeout_sec(void) {
+  return g_timeout_sec;
+}
+
 static void tcp_apply_socket_timeout(int fd) {
  struct timeval tv;
  tv.tv_sec = g_timeout_sec;
@@ -35,5 +35,6 @@ void client_disconnect(Client* client);
 void client_delete(Client* client);
 void tcp_set_timeouts(int timeout_sec, int contimeout_sec);
 int tcp_get_contimeout_sec(void);
+int tcp_get_timeout_sec(void);

 #endif
@@ -9,6 +9,10 @@
 #include <stdio.h>
 #include <stdlib.h>
 #include <string.h>
+#include <sys/socket.h>
+#include <sys/stat.h>
+#include <sys/wait.h>
+#include <time.h>
 #include <unistd.h>

 bool tls_global_init(void) {
@@ -31,6 +35,10 @@ static void log_ssl_errors(void) {

 static SSL_CTX* create_ssl_ctx(bool is_server, const char* cert, const char* key,
                               const char* ca_path) {
+  if (!is_server && !ca_path) {
+    log_message(LOG_LEVEL_ERROR, "TLS clients require a CA certificate path");
+    return NULL;
+  }
  const SSL_METHOD* method = is_server ? TLS_server_method() : TLS_client_method();
  SSL_CTX* ctx = SSL_CTX_new(method);
  if (!ctx) {
@@ -39,9 +47,23 @@ static SSL_CTX* create_ssl_ctx(bool is_server, const char* cert, const char* key
    return NULL;
  }

-  SSL_CTX_set_min_proto_version(ctx, TLS1_2_VERSION);
+  if (SSL_CTX_set_min_proto_version(ctx, TLS1_2_VERSION) != 1) {
+    SSL_CTX_free(ctx);
+    return NULL;
+  }
+  if (SSL_CTX_set_cipher_list(ctx, "HIGH:!aNULL:!eNULL:!MD5:!RC4:!3DES") != 1) {
+    SSL_CTX_free(ctx);
+    return NULL;
+  }

  if (cert && key) {
+    struct stat key_stat;
+    if (stat(key, &key_stat) != 0 || !S_ISREG(key_stat.st_mode) || key_stat.st_uid != geteuid() ||
+        (key_stat.st_mode & (S_IRGRP | S_IWGRP | S_IROTH | S_IWOTH))) {
+      log_message(LOG_LEVEL_ERROR, "TLS private key must be owned by the current user and private");
+      SSL_CTX_free(ctx);
+      return NULL;
+    }
    if (SSL_CTX_use_certificate_file(ctx, cert, SSL_FILETYPE_PEM) <= 0) {
      log_message(LOG_LEVEL_ERROR, "Failed to load certificate: %s", cert);
      log_ssl_errors();
@@ -83,15 +105,22 @@ static SSL* wrap_fd_with_ssl(int fd, SSL_CTX* ctx, bool is_server, const char* h
    log_message(LOG_LEVEL_ERROR, "Failed to create SSL object");
    return NULL;
  }
-  SSL_set_fd(ssl, fd);
+  if (SSL_set_fd(ssl, fd) != 1) {
+    SSL_free(ssl);
+    return NULL;
+  }

  // Enable hostname verification for client connections when a hostname is provided.
  // Must be done before SSL_connect to take effect during the handshake.
  if (!is_server && hostname) {
-    SSL_set1_host(ssl, hostname);
+    if (SSL_set1_host(ssl, hostname) != 1) {
+      SSL_free(ssl);
+      return NULL;
+    }
  }

  // Retry SSL_accept/SSL_connect on WANT_READ/WANT_WRITE (non-blocking handshake)
+  time_t deadline = time(NULL) + (is_server ? tcp_get_timeout_sec() : tcp_get_contimeout_sec());
  int ret;
  do {
    if (is_server)
@@ -101,7 +130,8 @@ static SSL* wrap_fd_with_ssl(int fd, SSL_CTX* ctx, bool is_server, const char* h

    if (ret <= 0) {
      int ssl_err = SSL_get_error(ssl, ret);
-      if (ssl_err == SSL_ERROR_WANT_READ || ssl_err == SSL_ERROR_WANT_WRITE)
+      if ((ssl_err == SSL_ERROR_WANT_READ || ssl_err == SSL_ERROR_WANT_WRITE) &&
+          time(NULL) < deadline)
        continue;
      log_message(LOG_LEVEL_ERROR, "SSL %s failed", is_server ? "accept" : "connect");
      log_ssl_errors();
@@ -129,8 +159,10 @@ struct tls_child_ctx {
 static void tls_child_fn(int fd, void* arg) {
  struct tls_child_ctx* ctx = (struct tls_child_ctx*)arg;
  SSL* ssl = wrap_fd_with_ssl(fd, ctx->ssl_ctx, true, NULL);
-  if (!ssl)
+  if (!ssl) {
+    io_set_ssl(NULL);
    return;
+  }
  io_set_ssl(ssl);
  ctx->handler(fd);
  SSL_shutdown(ssl);
@@ -146,12 +178,19 @@ bool server_listen_tls(Server* server, void (*handler)(int file_descriptor)) {

 bool client_connect_tls(Client* client, char* host, int port, const char* cert_path,
                        const char* key_path, const char* ca_path) {
-  if (!tcp_connect_socket(client, host, port))
+  if (!tcp_connect_socket(client, host, port)) {
+    if (client->file_descriptor >= 0)
+      close(client->file_descriptor);
+    client->file_descriptor = -1;
    return false;
+  }

  SSL_CTX* ctx = create_ssl_ctx(false, cert_path, key_path, ca_path);
-  if (!ctx)
+  if (!ctx) {
+    close(client->file_descriptor);
+    client->file_descriptor = -1;
    return false;
+  }
  client->ssl_ctx = ctx;

  // Pass the server hostname for TLS hostname verification (SSL_set1_host
@@ -161,6 +200,8 @@ bool client_connect_tls(Client* client, char* host, int port, const char* cert_p
  if (!ssl) {
    SSL_CTX_free(ctx);
    client->ssl_ctx = NULL;
+    close(client->file_descriptor);
+    client->file_descriptor = -1;
    return false;
  }

@@ -7,62 +7,117 @@
 #include <stdio.h>
 #include <stdlib.h>
 #include <string.h>
+#include <stdint.h>
 #include <sys/stat.h>
 #include <unistd.h>

 static int authorized_root_fd = -1;
+static char* authorized_root_path;
+
+bool utils_set_authorized_root(int fd, const char* canonical_path) {
+  char* path_copy = canonical_path ? str_dup(canonical_path) : NULL;
+  if (canonical_path && !path_copy) {
+    authorized_root_fd = -1;
+    free(authorized_root_path);
+    authorized_root_path = NULL;
+    return false;
+  }
+  authorized_root_fd = fd;
+  free(authorized_root_path);
+  authorized_root_path = path_copy;
+  return true;
+}

 void utils_set_authorized_root_fd(int fd) {
-  authorized_root_fd = fd;
+  (void)utils_set_authorized_root(fd, NULL);
+}
+
+static bool path_is_within_root(const char* root, const char* path) {
+  size_t root_len = strlen(root);
+  return strncmp(root, path, root_len) == 0 && (path[root_len] == '\0' || path[root_len] == '/');
+}
+
+static int open_authorized_destination(const char* dest_root) {
+  if (authorized_root_fd < 0 || !authorized_root_path || !dest_root ||
+      !path_is_within_root(authorized_root_path, dest_root))
+    return -1;
+
+  int dirfd = dup(authorized_root_fd);
+  if (dirfd < 0)
+    return -1;
+
+  const char* relative_path = dest_root + strlen(authorized_root_path);
+  while (*relative_path == '/')
+    relative_path++;
+  char* relative = str_dup(*relative_path ? relative_path : ".");
+  if (!relative) {
+    close(dirfd);
+    return -1;
+  }
+
+  char* saveptr = NULL;
+  char* component = strtok_r(relative, "/", &saveptr);
+  while (component) {
+    if (strcmp(component, "..") == 0) {
+      free(relative);
+      close(dirfd);
+      return -1;
+    }
+    if (strcmp(component, ".") == 0) {
+      component = strtok_r(NULL, "/", &saveptr);
+      continue;
+    }
+    int next = openat(dirfd, component, O_RDONLY | O_DIRECTORY | O_NOFOLLOW | O_CLOEXEC);
+    if (next < 0) {
+      free(relative);
+      close(dirfd);
+      return -1;
+    }
+    close(dirfd);
+    dirfd = next;
+    component = strtok_r(NULL, "/", &saveptr);
+  }
+
+  free(relative);
+  return dirfd;
 }

 bool mkdir_r(const char* path) {
-  size_t path_len = strlen(path);
-  char* path_duplicate = malloc(path_len + 1);
-  if (!path_duplicate)
+  if (!path || *path == '\0')
    return false;
-  memcpy(path_duplicate, path, path_len + 1);
-  size_t capacity = path_len + 2;
-  char* path_current = (char*)malloc(capacity * sizeof(char));
-  if (!path_current) {
-    free(path_duplicate);
+  char* duplicate = str_dup(path);
+  if (!duplicate)
+    return false;
+  int dirfd = open(path[0] == '/' ? "/" : ".", O_RDONLY | O_DIRECTORY | O_CLOEXEC | O_NOFOLLOW);
+  if (dirfd < 0) {
+    free(duplicate);
    return false;
  }
-  char* path_current_position = path_current;
-  if (path[0] == '/') {
-    path_current[0] = '/';
-    path_current[1] = '\0';
-    path_current_position += 1;
-  } else {
-    path_current[0] = '\0';
-  }
-  const char* delimiter = "/";
-  char* saveptr;
-  const char* part = strtok_r(path_duplicate, delimiter, &saveptr);
  bool ok = true;
-  while (part != NULL) {
-    size_t part_len = strlen(part);
-    if ((size_t)(path_current_position - path_current) + part_len + 2 > capacity) {
+  char* saveptr = NULL;
+  char* component = strtok_r(duplicate, "/", &saveptr);
+  while (component) {
+    if (strcmp(component, "..") == 0) {
      ok = false;
      break;
    }
-    memcpy(path_current_position, part, part_len);
-    path_current_position += part_len;
-    path_current_position[0] = '/';
-    path_current_position[1] = '\0';
-    path_current_position++;
-    struct stat st;
-    if (stat(path_current, &st) != 0) {
-      if (mkdir(path_current, 0755) != 0) {
-        perror("Could not create directory");
+    if (strcmp(component, ".") != 0) {
+      int next = openat(dirfd, component, O_RDONLY | O_DIRECTORY | O_CLOEXEC | O_NOFOLLOW);
+      if (next < 0 && errno == ENOENT) {
+        if (mkdirat(dirfd, component, 0755) == 0 || errno == EEXIST)
+          next = openat(dirfd, component, O_RDONLY | O_DIRECTORY | O_CLOEXEC | O_NOFOLLOW);
+      }
+      if (next < 0) {
        ok = false;
        break;
      }
+      close(dirfd);
+      dirfd = next;
    }
-    part = strtok_r(NULL, delimiter, &saveptr);
+    component = strtok_r(NULL, "/", &saveptr);
  }
-  free(path_duplicate);
-  free(path_current);
+  close(dirfd);
+  free(duplicate);
  return ok;
 }

@@ -143,7 +198,8 @@ static bool is_dir_in_manifest(const char* rel_path, ArrayList* manifest) {
  return false;
 }

-static bool delete_extras_fd(int dirfd, const char* rel_path, ArrayList* manifest) {
+static bool delete_extras_fd(int dirfd, const char* rel_path, ArrayList* manifest,
+                             size_t max_delete, size_t* deleted_count) {
  int scanfd = dup(dirfd);
  if (scanfd < 0)
    return false;
@@ -152,15 +208,20 @@ static bool delete_extras_fd(int dirfd, const char* rel_path, ArrayList* manifes
    close(scanfd);
    return false;
  }
-  bool all_removed = true;
  bool operation_ok = true;
  const struct dirent* entry;
  while ((entry = readdir(dir)) != NULL) {
    if (strcmp(entry->d_name, ".") == 0 || strcmp(entry->d_name, "..") == 0)
      continue;
    char* child_rel = path_cat((char*)rel_path, entry->d_name);
+    if (!child_rel) {
+      operation_ok = false;
+      continue;
+    }
    struct stat st;
    if (fstatat(dirfd, entry->d_name, &st, AT_SYMLINK_NOFOLLOW) != 0) {
+      if (errno != ENOENT)
+        operation_ok = false;
      free(child_rel);
      continue;
    }
@@ -173,14 +234,24 @@ static bool delete_extras_fd(int dirfd, const char* rel_path, ArrayList* manifes
      int childfd = openat(dirfd, entry->d_name, O_RDONLY | O_DIRECTORY | O_NOFOLLOW | O_CLOEXEC);
      bool child_removed = false;
      if (childfd >= 0) {
-        child_removed = delete_extras_fd(childfd, child_rel, manifest);
+        child_removed = delete_extras_fd(childfd, child_rel, manifest, max_delete, deleted_count);
+        if (!child_removed)
+          operation_ok = false;
        close(childfd);
-      }
-      if (child_removed && !is_dir_in_manifest(child_rel, manifest) &&
-          unlinkat(dirfd, entry->d_name, AT_REMOVEDIR) != 0 && errno != ENOENT) {
+      } else if (errno != ENOENT) {
        operation_ok = false;
-      } else if (!child_removed) {
-        all_removed = false;
+      }
+      if (child_removed && !is_dir_in_manifest(child_rel, manifest)) {
+        if (*deleted_count >= max_delete) {
+          operation_ok = false;
+        } else {
+          if (unlinkat(dirfd, entry->d_name, AT_REMOVEDIR) != 0) {
+            if (errno != ENOENT)
+              operation_ok = false;
+          } else {
+            (*deleted_count)++;
+          }
+        }
      }
    } else {
      // Check if relative path is in manifest
@@ -192,38 +263,59 @@ static bool delete_extras_fd(int dirfd, const char* rel_path, ArrayList* manifes
        }
      }
      if (!found) {
-        if (unlinkat(dirfd, entry->d_name, 0) != 0 && errno != ENOENT)
+        if (*deleted_count >= max_delete) {
          operation_ok = false;
+          free(child_rel);
+          continue;
+        }
+        if (unlinkat(dirfd, entry->d_name, 0) != 0) {
+          if (errno != ENOENT)
+            operation_ok = false;
+        } else {
+          (*deleted_count)++;
+        }
        fprintf(stderr, "  Deleted: %s\n", child_rel);
-      } else {
-        all_removed = false;
      }
    }
    free(child_rel);
  }
  closedir(dir);
-  (void)all_removed;
  return operation_ok;
 }

-bool delete_extras(const char* dest_root, ArrayList* manifest) {
-  int rootfd = authorized_root_fd >= 0
-                   ? dup(authorized_root_fd)
-                   : open(dest_root, O_RDONLY | O_DIRECTORY | O_NOFOLLOW | O_CLOEXEC);
+bool delete_extras_limited(const char* dest_root, ArrayList* manifest, size_t max_delete) {
+  if (!manifest)
+    return false;
+  int rootfd;
+  if (authorized_root_fd >= 0) {
+    if (authorized_root_path)
+      rootfd = open_authorized_destination(dest_root);
+    else if (dest_root == NULL)
+      rootfd = dup(authorized_root_fd);
+    else
+      rootfd = -1;
+  } else {
+    rootfd = open(dest_root, O_RDONLY | O_DIRECTORY | O_NOFOLLOW | O_CLOEXEC);
+  }
  if (rootfd < 0)
    return false;
-  bool ok = delete_extras_fd(rootfd, "", manifest);
+  size_t deleted_count = 0;
+  bool ok = delete_extras_fd(rootfd, "", manifest, max_delete, &deleted_count);
  if (close(rootfd) != 0)
    ok = false;
  return ok;
 }

+bool delete_extras(const char* dest_root, ArrayList* manifest) {
+  return delete_extras_limited(dest_root, manifest, SIZE_MAX);
+}
+
 bool has_path_traversal(const char* path) {
  if (!path)
-    return false;
+    return true;
  char* dup = str_dup(path);
  if (!dup)
-    return false;
+    return true;
  char* saveptr;
  const char* part = strtok_r(dup, "/", &saveptr);
  while (part) {
@@ -255,6 +347,8 @@ char* path_cat(const char* path1, const char* path2) {
    offset = 1;
    path2_len -= 1;
  }
+  if (path1_len > SIZE_MAX - path2_len - 2)
+    return NULL;
  char* new_path = malloc(path1_len + path2_len + 2);
  if (new_path == NULL)
    return NULL;
@@ -2,6 +2,7 @@
 #define UTILS_H

 #include "array_list.h"
+#include <stddef.h>
 #include <stdbool.h>

 bool mkdir_r(const char* path);
@@ -9,6 +10,10 @@ char* str_dup(const char* string);
 char* path_cat(const char* path1, const char* path2);
 bool glob_match(const char* pattern, const char* str);
 bool delete_extras(const char* dest_root, ArrayList* manifest);
+bool delete_extras_limited(const char* dest_root, ArrayList* manifest, size_t max_delete);
+bool utils_set_authorized_root(int fd, const char* canonical_path);
+/* The fd-only compatibility form is fail-closed for path-based operations;
+ * callers should use utils_set_authorized_root with the canonical identity. */
 void utils_set_authorized_root_fd(int fd);
 bool has_path_traversal(const char* path);
 bool utils_valid_batch_path(const char* path);
@@ -25,7 +25,9 @@ class ServerManager:
    def start(self, extra_args=None):
        self.stop()
        self._port = _find_free_port()
-        cmd = SERVER_CMD + ["-p", str(self._port)]
+        # Plain TCP is intentionally explicit in the server; integration tests
+        # exercise that opt-in mode rather than relying on the secure default.
+        cmd = SERVER_CMD + ["-p", str(self._port), "--allow-unauthenticated"]
        if extra_args:
            cmd += extra_args
        self._proc = subprocess.Popen(cmd, stdout=subprocess.DEVNULL, stderr=subprocess.DEVNULL)
@@ -102,6 +102,7 @@ class TestTLSBasic:
        with ServerManager() as server:
            server.start(extra_args=[
                "--tls", "--cert", certs["server_cert"], "--key", certs["server_key"],
+                "--ca", certs["ca"], "--client-cn", "fastsync-client",
            ])
            result, dur = run_client(
                SOURCE_DIR, DEST_DIR,
@@ -125,6 +126,7 @@ class TestTLSBasic:
        with ServerManager() as server:
            server.start(extra_args=[
                "--tls", "--cert", certs["server_cert"], "--key", certs["server_key"],
+                "--ca", certs["ca"], "--client-cn", "fastsync-client",
            ])
            result, dur = run_client(
                SOURCE_DIR, DEST_DIR,
@@ -149,6 +151,7 @@ class TestTLSBasic:
        with ServerManager() as server:
            server.start(extra_args=[
                "--tls", "--cert", certs["server_cert"], "--key", certs["server_key"],
+                "--ca", certs["ca"], "--client-cn", "fastsync-client",
            ])
            result, dur = run_client(
                SOURCE_DIR, DEST_DIR,
@@ -17,6 +17,8 @@ void test_array_list() {

  // Test adding
  int* val1 = malloc(sizeof(int));
+  if (!val1)
+    return;
  *val1 = 42;
  array_list_add(list, val1);
  EXPECT_EQ_INT(list->size, 1);
@@ -26,6 +28,8 @@ void test_array_list() {
  // Initial capacity is 100. Let's add 105 elements.
  for (int i = 0; i < 105; i++) {
    int* val = malloc(sizeof(int));
+    if (!val)
+      return;
    *val = i;
    array_list_add(list, val);
  }
@@ -46,6 +46,8 @@ static void test_validate_config_tls_requirements() {
  cfg->tls_cert = str_dup("cert.pem");
  EXPECT_FALSE(validate_config(cfg));
  cfg->tls_key = str_dup("key.pem");
+  EXPECT_FALSE(validate_config(cfg));
+  cfg->tls_ca = str_dup("ca.pem");
  EXPECT_TRUE(validate_config(cfg));
  config_delete(cfg);
 }
@@ -13,6 +13,8 @@ static void test_data_compress_decompress_roundtrip() {
  size_t len = strlen(original);

  char* buf = malloc(len);
+  if (!buf)
+    return;
  memcpy(buf, original, len);
  Data* original_data = data_create(buf, len);
  EXPECT_NOT_NULL(original_data);
@@ -14,6 +14,8 @@
 static Data* random_data(int min_size, int max_size) {
  int size = min_size + rand() % (max_size - min_size + 1);
  char* buf = malloc(size);
+  if (!buf)
+    return NULL;
  for (int i = 0; i < size; i++)
    buf[i] = (char)(rand() % 256);
  return data_create(buf, size);
@@ -81,6 +83,8 @@ static void test_property_chunk_roundtrip() {

    int content_len = 1 + rand() % 4096;
    char* content = malloc(content_len);
+    if (!content)
+      return;
    for (int i = 0; i < content_len; i++)
      content[i] = (char)(rand() % 256);

@@ -122,6 +122,8 @@ static void test_queue_destroyer() {

  for (int i = 0; i < 3; i++) {
    int* val = malloc(sizeof(int));
+    if (!val)
+      break;
    *val = i;
    queue_enqueue(q, val);
  }
@@ -181,6 +183,8 @@ static void test_queue_multithreaded() {

  for (int i = 1; i <= 100; i++) {
    int* val = malloc(sizeof(int));
+    if (!val)
+      break;
    *val = i;
    queue_enqueue_multithreaded(q, val, &mutex, &cnd_empty, &cnd_full);
  }
@@ -385,6 +385,34 @@ static void test_scanner_no_patterns() {
  rmdir(dir);
 }

+static void test_parallel_scanner_root_chunks_without_workers() {
+  const char* dir = "test_parallel_scan_root";
+  const char* file1 = "test_parallel_scan_root/a.txt";
+  const char* file2 = "test_parallel_scan_root/b.txt";
+
+  EXPECT_EQ_INT(mkdir(dir, 0755), 0);
+  create_test_file(file1, "a");
+  create_test_file(file2, "b");
+
+  ParallelScanner* scanner = parallel_scanner_create(dir, false, 1, NULL, 0, NULL, 0, 0, 0, 0, 0,
+                                                     false, false, false, false, false);
+  EXPECT_NOT_NULL(scanner);
+
+  int total_files = 0;
+  Chunk* chunk;
+  while ((chunk = parallel_scanner_next(scanner)) != NULL) {
+    total_files += chunk->element_count;
+    chunk_destroy(chunk);
+  }
+  EXPECT_EQ_INT(total_files, 2);
+  EXPECT_FALSE(parallel_scanner_failed(scanner));
+
+  parallel_scanner_destroy(scanner);
+  unlink(file1);
+  unlink(file2);
+  rmdir(dir);
+}
+
 void test_scanner() {
  test_scanner_single_file();
  test_scanner_multiple_files();
@@ -399,4 +427,5 @@ void test_scanner() {
  test_scanner_size_range();
  test_scanner_mixed_patterns();
  test_scanner_no_patterns();
+  test_parallel_scanner_root_chunks_without_workers();
 }
@@ -32,6 +32,8 @@ static int mpmc_producer_func(void* arg) {
  ProducerCtx* ctx = (ProducerCtx*)arg;
  for (int i = 1; i <= ITEMS_PER_PRODUCER; i++) {
    int* val = malloc(sizeof(int));
+    if (!val)
+      return thrd_error;
    *val = ctx->producer_id * ITEMS_PER_PRODUCER + i;
    queue_enqueue_multithreaded(ctx->q, val, ctx->mutex, ctx->cnd_empty, ctx->cnd_full);
  }
@@ -121,6 +123,8 @@ static int bp_producer_func(void* arg) {
  BackpressureCtx* ctx = (BackpressureCtx*)arg;
  for (int i = 0; i < 5; i++) {
    int* val = malloc(sizeof(int));
+    if (!val)
+      return thrd_error;
    *val = i + 1;
    queue_enqueue_multithreaded(ctx->q, val, ctx->mutex, ctx->cnd_empty, ctx->cnd_full);
    ctx->items_sent++;
@@ -194,6 +198,8 @@ static void test_queue_rapid_create_destroy() {

    for (int j = 0; j < 3; j++) {
      int* val = malloc(sizeof(int));
+      if (!val)
+        break;
      *val = j;
      queue_enqueue(q, val);
    }
Author	SHA1	Message	Date
TapTap	6c4d0246bc	merge: resolve origin dev refactor conflicts CI / lint (pull_request) Successful in 28s Details CI / sanitizers (address) (pull_request) Successful in 36s Details CI / sanitizers (undefined) (pull_request) Successful in 37s Details CI / fuzz-build (pull_request) Successful in 14s Details CI / coverage (pull_request) Successful in 32s Details CI / build-and-test (pull_request) Successful in 1m15s Details CI / valgrind (pull_request) Successful in 33s Details	2026-08-16 09:56:15 +02:00
TapTap	1391564ce7	test: harden allocation checks CI / lint (pull_request) Successful in 31s Details CI / sanitizers (address) (pull_request) Successful in 37s Details CI / sanitizers (undefined) (pull_request) Successful in 37s Details CI / fuzz-build (pull_request) Successful in 14s Details CI / coverage (pull_request) Successful in 31s Details CI / build-and-test (pull_request) Successful in 1m14s Details CI / valgrind (pull_request) Successful in 33s Details	2026-08-16 09:37:28 +02:00
TapTap	d102622e28	fix: close remaining PR review gaps CI / lint (pull_request) Failing after 31s Details CI / build-and-test (pull_request) Has been skipped Details CI / sanitizers (address) (pull_request) Has been skipped Details CI / sanitizers (undefined) (pull_request) Has been skipped Details CI / fuzz-build (pull_request) Has been skipped Details CI / coverage (pull_request) Has been skipped Details CI / valgrind (pull_request) Has been skipped Details	2026-08-16 09:35:26 +02:00
TapTap	1abc17142f	Merge pull request 'refactor: split transfer and protocol responsibilities' (#212 ) from refactor/codebase-structure into dev CI / lint (push) Successful in 30s Details CI / sanitizers (undefined) (push) Successful in 38s Details CI / sanitizers (address) (push) Successful in 39s Details CI / fuzz-build (push) Successful in 14s Details CI / coverage (push) Successful in 31s Details CI / build-and-test (push) Successful in 1m15s Details CI / valgrind (push) Successful in 33s Details Reviewed-on: #212	2026-08-16 09:11:58 +02:00
TapTap	059dc2ac75	fix: close remaining PR review gaps CI / lint (pull_request) Successful in 32s Details CI / sanitizers (undefined) (pull_request) Successful in 36s Details CI / sanitizers (address) (pull_request) Successful in 37s Details CI / fuzz-build (pull_request) Successful in 13s Details CI / coverage (pull_request) Successful in 31s Details CI / build-and-test (pull_request) Successful in 1m15s Details CI / valgrind (pull_request) Successful in 33s Details	2026-08-15 20:38:04 +02:00
TapTap	7cffff0b8b	fix: fail closed on authorization setup failure CI / lint (pull_request) Successful in 30s Details CI / sanitizers (address) (pull_request) Successful in 36s Details CI / sanitizers (undefined) (pull_request) Successful in 35s Details CI / fuzz-build (pull_request) Successful in 15s Details CI / coverage (pull_request) Successful in 31s Details CI / build-and-test (pull_request) Successful in 1m15s Details CI / valgrind (pull_request) Successful in 33s Details	2026-08-15 20:02:09 +02:00
TapTap	d3b4c62f51	Merge branch 'dev' into refactor/codebase-structure CI / lint (pull_request) Successful in 30s Details CI / sanitizers (address) (pull_request) Successful in 36s Details CI / sanitizers (undefined) (pull_request) Successful in 37s Details CI / fuzz-build (pull_request) Successful in 15s Details CI / coverage (pull_request) Successful in 31s Details CI / build-and-test (pull_request) Successful in 1m15s Details CI / valgrind (pull_request) Successful in 32s Details	2026-08-15 15:09:00 +02:00
TapTap	2f804ecfdd	fix: address remaining PR 212 review issues CI / lint (pull_request) Successful in 31s Details CI / sanitizers (address) (pull_request) Successful in 37s Details CI / sanitizers (undefined) (pull_request) Successful in 36s Details CI / fuzz-build (pull_request) Successful in 15s Details CI / coverage (pull_request) Successful in 32s Details CI / build-and-test (pull_request) Successful in 1m15s Details CI / valgrind (pull_request) Successful in 33s Details	2026-08-15 13:52:38 +02:00
TapTap	0c24136052	Merge pull request 'docs: clarify rsync compatibility roadmap' (#215 ) from docs/rsync-compatible-readme-clean into dev CI / lint (push) Successful in 34s Details CI / sanitizers (address) (push) Successful in 35s Details CI / sanitizers (undefined) (push) Successful in 35s Details CI / fuzz-build (push) Successful in 15s Details CI / build-and-test (push) Successful in 1m15s Details CI / coverage (push) Successful in 31s Details CI / valgrind (push) Successful in 33s Details Reviewed-on: #215	2026-08-15 13:48:01 +02:00
TapTap	d19fc68803	Merge branch 'dev' into docs/rsync-compatible-readme-clean CI / lint (pull_request) Successful in 33s Details CI / sanitizers (address) (pull_request) Successful in 36s Details CI / sanitizers (undefined) (pull_request) Successful in 36s Details CI / fuzz-build (pull_request) Successful in 15s Details CI / coverage (pull_request) Successful in 32s Details CI / build-and-test (pull_request) Successful in 1m16s Details CI / valgrind (pull_request) Successful in 33s Details	2026-08-15 13:47:55 +02:00
TapTap	78876b110e	fix: harden remaining review findings CI / lint (pull_request) Successful in 34s Details CI / sanitizers (address) (pull_request) Successful in 37s Details CI / sanitizers (undefined) (pull_request) Successful in 37s Details CI / fuzz-build (pull_request) Successful in 15s Details CI / coverage (pull_request) Successful in 33s Details CI / build-and-test (pull_request) Successful in 1m16s Details CI / valgrind (pull_request) Successful in 33s Details	2026-08-15 13:44:31 +02:00
TapTap	6b7213db5c	docs: align compatibility guidance CI / lint (pull_request) Failing after 2s Details CI / build-and-test (pull_request) Has been skipped Details CI / sanitizers (address) (pull_request) Has been skipped Details CI / sanitizers (undefined) (pull_request) Has been skipped Details CI / fuzz-build (pull_request) Has been skipped Details CI / coverage (pull_request) Has been skipped Details CI / valgrind (pull_request) Has been skipped Details	2026-08-15 13:41:23 +02:00
TapTap	b3a724afc8	fix: preserve scanner file ownership on failure CI / lint (pull_request) Successful in 32s Details CI / sanitizers (address) (pull_request) Successful in 37s Details CI / sanitizers (undefined) (pull_request) Successful in 37s Details CI / fuzz-build (pull_request) Successful in 14s Details CI / coverage (pull_request) Successful in 31s Details CI / build-and-test (pull_request) Successful in 1m15s Details CI / valgrind (pull_request) Successful in 34s Details	2026-08-15 13:40:31 +02:00
TapTap	daf662cc2d	fix: address remaining PR review findings CI / lint (pull_request) Successful in 33s Details CI / sanitizers (address) (pull_request) Successful in 36s Details CI / sanitizers (undefined) (pull_request) Successful in 36s Details CI / fuzz-build (pull_request) Successful in 14s Details CI / coverage (pull_request) Successful in 31s Details CI / build-and-test (pull_request) Successful in 1m15s Details CI / valgrind (pull_request) Successful in 33s Details	2026-08-15 13:24:13 +02:00
TapTap	98f833980d	fix: handle pipeline cancellation failures CI / lint (pull_request) Successful in 30s Details CI / sanitizers (address) (pull_request) Successful in 37s Details CI / sanitizers (undefined) (pull_request) Successful in 38s Details CI / fuzz-build (pull_request) Successful in 16s Details CI / coverage (pull_request) Successful in 32s Details CI / build-and-test (pull_request) Successful in 1m16s Details CI / valgrind (pull_request) Successful in 32s Details	2026-08-15 13:20:50 +02:00
TapTap	de537810e5	docs: fix server destination root examples CI / lint (pull_request) Failing after 2s Details CI / build-and-test (pull_request) Has been skipped Details CI / sanitizers (address) (pull_request) Has been skipped Details CI / sanitizers (undefined) (pull_request) Has been skipped Details CI / fuzz-build (pull_request) Has been skipped Details CI / coverage (pull_request) Has been skipped Details CI / valgrind (pull_request) Has been skipped Details	2026-08-15 13:14:28 +02:00
TapTap	298bfe0d0f	fix: address delegated review findings CI / lint (pull_request) Successful in 33s Details CI / sanitizers (address) (pull_request) Successful in 35s Details CI / sanitizers (undefined) (pull_request) Successful in 36s Details CI / fuzz-build (pull_request) Successful in 14s Details CI / build-and-test (pull_request) Successful in 1m14s Details CI / coverage (pull_request) Successful in 30s Details CI / valgrind (pull_request) Successful in 33s Details	2026-08-15 13:13:57 +02:00
TapTap	604a14f0be	fix: close remaining PR review gaps CI / lint (pull_request) Successful in 30s Details CI / sanitizers (address) (pull_request) Successful in 36s Details CI / sanitizers (undefined) (pull_request) Successful in 36s Details CI / fuzz-build (pull_request) Successful in 15s Details CI / coverage (pull_request) Successful in 33s Details CI / build-and-test (pull_request) Successful in 1m15s Details CI / valgrind (pull_request) Successful in 33s Details	2026-08-15 13:13:50 +02:00
TapTap	def58554d9	docs: clarify rsync compatibility roadmap CI / lint (pull_request) Failing after 2s Details CI / build-and-test (pull_request) Has been skipped Details CI / sanitizers (address) (pull_request) Has been skipped Details CI / sanitizers (undefined) (pull_request) Has been skipped Details CI / fuzz-build (pull_request) Has been skipped Details CI / coverage (pull_request) Has been skipped Details CI / valgrind (pull_request) Has been skipped Details	2026-08-15 12:54:38 +02:00
TapTap	6f8847899c	fix: validate remaining wire booleans CI / lint (pull_request) Successful in 34s Details CI / sanitizers (address) (pull_request) Successful in 35s Details CI / fuzz-build (pull_request) Successful in 14s Details CI / sanitizers (undefined) (pull_request) Successful in 35s Details CI / build-and-test (pull_request) Successful in 1m14s Details CI / coverage (pull_request) Successful in 31s Details CI / valgrind (pull_request) Successful in 33s Details	2026-08-15 12:49:30 +02:00
TapTap	09454413d4	fix: address PR 212 review issues CI / lint (pull_request) Successful in 30s Details CI / sanitizers (address) (pull_request) Successful in 38s Details CI / sanitizers (undefined) (pull_request) Successful in 38s Details CI / fuzz-build (pull_request) Successful in 15s Details CI / coverage (pull_request) Successful in 32s Details CI / build-and-test (pull_request) Successful in 1m15s Details CI / valgrind (pull_request) Successful in 33s Details	2026-08-15 12:48:31 +02:00
TapTap	ff189aeef2	fix: harden network security boundaries CI / lint (pull_request) Successful in 33s Details CI / sanitizers (address) (pull_request) Successful in 37s Details CI / sanitizers (undefined) (pull_request) Successful in 37s Details CI / fuzz-build (pull_request) Successful in 13s Details CI / coverage (pull_request) Successful in 32s Details CI / build-and-test (pull_request) Successful in 1m16s Details CI / valgrind (pull_request) Successful in 32s Details	2026-08-15 12:34:41 +02:00
TapTap	41b624db5a	Merge pull request 'fix: satisfy clang-format in CLI test' (#211 ) from fix/dev-ci-clang-format into dev CI / lint (push) Successful in 34s Details CI / sanitizers (address) (push) Successful in 37s Details CI / sanitizers (undefined) (push) Successful in 37s Details CI / fuzz-build (push) Successful in 14s Details CI / coverage (push) Successful in 31s Details CI / build-and-test (push) Successful in 1m16s Details CI / valgrind (push) Successful in 34s Details Reviewed-on: #211	2026-08-15 11:55:41 +02:00