fix: set *failed on delta oversize branch; add -m RSF skip test

Review nits from independent review of the four fix branches:
- receive_delta_file STATUS_NEXT oversize branch now sets *failed=true
- receive_incremental_check oversize branch returns NULL (receiver sends the
  single STATUS_ERROR) instead of double-sending
- add multithreaded -m --ignore-existing --remove-source-files integration
  coverage so the writer-thread outcome path is exercised

README.md

@@ -484,10 +484,11 @@ defaults to the current directory. |
 
 ## Protocol and Security
 
-FastSync protocol version `2.4.0` is shared by the client and server. The
+FastSync protocol version `2.5.0` is shared by the client and server. The
 current protocol is sender-driven and includes configuration negotiation,
 including the maximum allocation limit, incremental checks, checksums,
-manifests, keep-alives, abort handling, and FastSync-native delta messages.
+manifests, keep-alives, abort handling, per-file remove-source results, and
+FastSync-native delta messages.
 Client and server versions must currently match exactly.
 
 TLS provides encrypted TCP transport. Supplying `--ca` enables certificate

src/client/client_cli.c

@@ -12,6 +12,7 @@
 #include "utils.h"
 #include <errno.h>
 #include <limits.h>
+#include <signal.h>
 #include <stdbool.h>
 #include <stddef.h>
 #include <stdio.h>
@@ -57,7 +58,7 @@ static bool parse_positive_int(const char* s, int* out_val) {
   return true;
 }
 
-/* Duplicate a string argument into *dest, freeing the old value. Returns true on success, false on
+/* Duplicate a string argument into *dest, freeing the old value. Returns 0 on success, -1 on
  * failure. */
 static int set_string_option(char** dest, const char* value, const char* option_name) {
   char* dup = str_dup(value);
@@ -70,7 +71,7 @@ static int set_string_option(char** dest, const char* value, const char* option_
   return 0;
 }
 
-/* Parse a string as a positive integer into *dest. Returns true on success, false on error. */
+/* Parse a string as a positive integer into *dest. Returns 0 on success, -1 on error. */
 static int set_positive_int_option(int* dest, const char* value, const char* option_name) {
   if (!parse_positive_int(value, dest)) {
     log_message(LOG_LEVEL_ERROR, "%s must be a positive integer", option_name);
@@ -102,7 +103,7 @@ static int set_compression_threads_option(int* dest, const char* value) {
   return 0;
 }
 
-/* Parse a string as a non-negative integer into *dest. Returns true on success, false on error. */
+/* Parse a string as a non-negative integer into *dest. Returns 0 on success, -1 on error. */
 static int set_nonneg_int_option(int* dest, const char* value, const char* option_name) {
   if (!parse_nonneg_int(value, dest)) {
     log_message(LOG_LEVEL_ERROR, "%s must be a non-negative integer", option_name);
@@ -237,7 +238,10 @@ static int parse_ull_arg(const char* val, unsigned long long* out, const char* o
   return 0;
 }
 
-static int parse_size_arg(const char* value, unsigned long long* out) {
+/* Parse a byte count with an optional single-letter binary suffix (K/M/G/T/P/E).
+ * When allow_zero is false, a bare 0 is rejected (size limits use true, since 0
+ * means "no limit"). Returns 0 on success, -1 on error. */
+static int parse_size_arg_allow_zero(const char* value, unsigned long long* out, bool allow_zero) {
   if (!value || *value < '0' || *value > '9')
     return -1;
   char* end;
@@ -281,12 +285,16 @@ static int parse_size_arg(const char* value, unsigned long long* out) {
       return -1;
     }
   }
-  if (number == 0 || number > ULLONG_MAX / multiplier)
+  if ((!allow_zero && number == 0) || number > ULLONG_MAX / multiplier)
     return -1;
   *out = number * multiplier;
   return 0;
 }
 
+static int parse_size_arg(const char* value, unsigned long long* out) {
+  return parse_size_arg_allow_zero(value, out, false);
+}
+
 /* Append a duplicated pattern to a growable pattern array. Returns 0 on success, -1 on error. */
 static int config_add_pattern(char*** patterns, int* count, const char* value,
                               const char* optname) {
@@ -450,6 +458,8 @@ static const OptionEntry* find_table_option(const char* arg) {
   return NULL;
 }
 
+/* Match a "--opt=value" argument against table options that take a value. Flags,
+ * no-ops, and unsupported options do not accept an inline "=" value. */
 static const OptionEntry* find_table_option_with_equals(const char* arg, const char** value) {
   const char* equals = strchr(arg, '=');
   if (!equals || equals == arg)
@@ -460,8 +470,8 @@ static const OptionEntry* find_table_option_with_equals(const char* arg, const c
     if ((strlen(entry->name) == name_len && strncmp(arg, entry->name, name_len) == 0) ||
         (entry->alias && strlen(entry->alias) == name_len &&
          strncmp(arg, entry->alias, name_len) == 0)) {
-      if (strcmp(entry->name, "--compress-choice") == 0 ||
-          strcmp(entry->name, "--compress-level") == 0) {
+      if (entry->kind == OPT_STRING || entry->kind == OPT_POS_INT ||
+          entry->kind == OPT_NONNEG_INT || entry->kind == OPT_ULL) {
         *value = equals + 1;
         return entry;
       }
@@ -516,8 +526,13 @@ static int apply_table_option(Config* config, const OptionEntry* entry, const ch
     return set_nonneg_int_option((int*)field, value, entry->name);
   case OPT_ULL: {
     unsigned long long v;
-    if (parse_ull_arg(value, &v, entry->name) != 0)
+    /* Size-limit options accept rsync-style suffixes (e.g. --max-size=2G); a
+     * plain byte count, including 0 ("no limit"), stays valid. */
+    if (parse_size_arg_allow_zero(value, &v, true) != 0) {
+      log_message(LOG_LEVEL_ERROR, "%s must be a non-negative size (B, K, M, G, T, P, or E)",
+                  entry->name);
       return -1;
+    }
     *(unsigned long long*)field = v;
     return 0;
   }
@@ -667,22 +682,38 @@ int parse_args(Config* config, int argc, char* argv[], int* positional_args,
       config->use_multithreading = true;
       config->use_metadata = true;
       log_info_message(LOG_INFO_MISC, "Enabled archive mode (-c -m -M)");
-    } else if (opt_is(argv[i], "-p", NULL) && i + 1 < argc) {
+    } else if (opt_is(argv[i], "-p", NULL)) {
+      if (i + 1 >= argc) {
+        log_message(LOG_LEVEL_ERROR, "missing argument for %s", argv[i]);
+        return -1;
+      }
       if (set_positive_int_option(&config->ssh_port, argv[++i], "-p") != 0)
         return -1;
       if (config->ssh_port > 65535) {
         log_message(LOG_LEVEL_ERROR, "SSH port must be 1-65535");
         return -1;
       }
-    } else if (opt_is(argv[i], "--exclude", NULL) && i + 1 < argc) {
+    } else if (opt_is(argv[i], "--exclude", NULL)) {
+      if (i + 1 >= argc) {
+        log_message(LOG_LEVEL_ERROR, "missing argument for %s", argv[i]);
+        return -1;
+      }
       if (config_add_pattern(&config->exclude_patterns, &config->exclude_count, argv[++i],
                              "--exclude") != 0)
         return -1;
-    } else if (opt_is(argv[i], "--include", NULL) && i + 1 < argc) {
+    } else if (opt_is(argv[i], "--include", NULL)) {
+      if (i + 1 >= argc) {
+        log_message(LOG_LEVEL_ERROR, "missing argument for %s", argv[i]);
+        return -1;
+      }
       if (config_add_pattern(&config->include_patterns, &config->include_count, argv[++i],
                              "--include") != 0)
         return -1;
-    } else if (opt_is(argv[i], "--delta-block", NULL) && i + 1 < argc) {
+    } else if (opt_is(argv[i], "--delta-block", NULL)) {
+      if (i + 1 >= argc) {
+        log_message(LOG_LEVEL_ERROR, "missing argument for %s", argv[i]);
+        return -1;
+      }
       unsigned long long val;
       if (parse_ull_arg(argv[++i], &val, "--delta-block") != 0)
         return -1;
@@ -690,7 +721,11 @@ int parse_args(Config* config, int argc, char* argv[], int* positional_args,
         config->delta_block_size = (uint32_t)val;
       else
         log_message(LOG_LEVEL_WARNING, "--delta-block value %llu out of range, using default", val);
-    } else if (opt_is(argv[i], "--delta-max", NULL) && i + 1 < argc) {
+    } else if (opt_is(argv[i], "--delta-max", NULL)) {
+      if (i + 1 >= argc) {
+        log_message(LOG_LEVEL_ERROR, "missing argument for %s", argv[i]);
+        return -1;
+      }
       unsigned long long val;
       if (parse_ull_arg(argv[++i], &val, "--delta-max") != 0)
         return -1;
@@ -731,7 +766,11 @@ int parse_args(Config* config, int argc, char* argv[], int* positional_args,
     } else if (opt_is(argv[i], "-s", NULL)) {
       config->use_chunk_serialization = true;
       log_info_message(LOG_INFO_MISC, "Enabled Chunk Serialization");
-    } else if (opt_is(argv[i], "--server-port", NULL) && i + 1 < argc) {
+    } else if (opt_is(argv[i], "--server-port", NULL)) {
+      if (i + 1 >= argc) {
+        log_message(LOG_LEVEL_ERROR, "missing argument for %s", argv[i]);
+        return -1;
+      }
       if (!parse_positive_int(argv[++i], &config->server_port)) {
         char* escaped = output_escape(argv[i], false);
         log_message(LOG_LEVEL_ERROR, "invalid --server-port value: %s",
@@ -743,7 +782,11 @@ int parse_args(Config* config, int argc, char* argv[], int* positional_args,
         log_message(LOG_LEVEL_ERROR, "server port must be 1-65535");
         return -1;
       }
-    } else if (opt_is(argv[i], "--bwlimit", NULL) && i + 1 < argc) {
+    } else if (opt_is(argv[i], "--bwlimit", NULL)) {
+      if (i + 1 >= argc) {
+        log_message(LOG_LEVEL_ERROR, "missing argument for %s", argv[i]);
+        return -1;
+      }
       unsigned long long kbps;
       if (parse_ull_arg(argv[++i], &kbps, "--bwlimit") != 0)
         return -1;
@@ -757,7 +800,11 @@ int parse_args(Config* config, int argc, char* argv[], int* positional_args,
       }
       io_set_bwlimit(kbps * 1024);
       log_info_message(LOG_INFO_MISC, "Set bandwidth limit to %llu KB/s", kbps);
-    } else if (opt_is(argv[i], "--chunk-size", NULL) && i + 1 < argc) {
+    } else if (opt_is(argv[i], "--chunk-size", NULL)) {
+      if (i + 1 >= argc) {
+        log_message(LOG_LEVEL_ERROR, "missing argument for %s", argv[i]);
+        return -1;
+      }
       unsigned long long val;
       if (parse_ull_arg(argv[++i], &val, "--chunk-size") != 0)
         return -1;
@@ -766,7 +813,11 @@ int parse_args(Config* config, int argc, char* argv[], int* positional_args,
         return -1;
       }
       config->chunk_size = val;
-    } else if (opt_is(argv[i], "--log-file", NULL) && i + 1 < argc) {
+    } else if (opt_is(argv[i], "--log-file", NULL)) {
+      if (i + 1 >= argc) {
+        log_message(LOG_LEVEL_ERROR, "missing argument for %s", argv[i]);
+        return -1;
+      }
       if (config->log_file) {
         fclose(config->log_file);
         config->log_file = NULL;
@@ -788,11 +839,19 @@ int parse_args(Config* config, int argc, char* argv[], int* positional_args,
     } else if (opt_is(argv[i], "--stderr", NULL)) {
       if (i + 1 >= argc || set_stderr_mode(argv[++i]) != 0)
         return -1;
-    } else if (opt_is(argv[i], "--exclude-from", NULL) && i + 1 < argc) {
+    } else if (opt_is(argv[i], "--exclude-from", NULL)) {
+      if (i + 1 >= argc) {
+        log_message(LOG_LEVEL_ERROR, "missing argument for %s", argv[i]);
+        return -1;
+      }
       if (read_patterns_from_file(argv[++i], &config->exclude_patterns, &config->exclude_count) !=
           0)
         return -1;
-    } else if (opt_is(argv[i], "--include-from", NULL) && i + 1 < argc) {
+    } else if (opt_is(argv[i], "--include-from", NULL)) {
+      if (i + 1 >= argc) {
+        log_message(LOG_LEVEL_ERROR, "missing argument for %s", argv[i]);
+        return -1;
+      }
       if (read_patterns_from_file(argv[++i], &config->include_patterns, &config->include_count) !=
           0)
         return -1;
@@ -817,16 +876,28 @@ int parse_args(Config* config, int argc, char* argv[], int* positional_args,
     } else if (opt_is(argv[i], "--info", NULL)) {
       if (i + 1 >= argc || parse_info_flags(argv[++i], config) != 0)
         return -1;
-    } else if (opt_is(argv[i], "-T", NULL) && i + 1 < argc) {
+    } else if (opt_is(argv[i], "-T", NULL)) {
+      if (i + 1 >= argc) {
+        log_message(LOG_LEVEL_ERROR, "missing argument for %s", argv[i]);
+        return -1;
+      }
       if (set_positive_int_option(&config->timeout, argv[++i], "-T") != 0)
         return -1;
     } else if (strncmp(argv[i], "--skip-compress=", 16) == 0) {
       if (parse_skip_compress(config, argv[i] + 16) != 0)
         return -1;
-    } else if (opt_is(argv[i], "--skip-compress", NULL) && i + 1 < argc) {
+    } else if (opt_is(argv[i], "--skip-compress", NULL)) {
+      if (i + 1 >= argc) {
+        log_message(LOG_LEVEL_ERROR, "missing argument for %s", argv[i]);
+        return -1;
+      }
       if (parse_skip_compress(config, argv[++i]) != 0)
         return -1;
-    } else if (opt_is(argv[i], "--compress-threads", NULL) && i + 1 < argc) {
+    } else if (opt_is(argv[i], "--compress-threads", NULL)) {
+      if (i + 1 >= argc) {
+        log_message(LOG_LEVEL_ERROR, "missing argument for %s", argv[i]);
+        return -1;
+      }
       if (set_compression_threads_option(&config->compression_threads, argv[++i]) != 0)
         return -1;
     } else if (opt_is(argv[i], "--checksum-choice", "--cc")) {
@@ -903,6 +974,10 @@ static int read_patterns_from_file(const char* filepath, char*** patterns, int*
 
 #ifndef FASTSYNC_TEST_BUILD
 int main(int argc, char* argv[]) {
+  /* The server may close a connection mid-stream (e.g. when it rejects an
+     oversized delta).  Ignore SIGPIPE so that a broken TCP connection
+     surfaces as a clean write error instead of killing the client. */
+  signal(SIGPIPE, SIG_IGN);
   const char* env_source = NULL;
   const char* env_dest = NULL;
   bool save_to_disk = false;

src/client/client_send.c

@@ -109,16 +109,13 @@ static bool add_chunk_to_manifest(ArrayList* manifest, const Chunk* chunk) {
   return true;
 }
 
-static bool finalize_transfer(Client* client) {
-  Status status;
-  return send_status(client->file_descriptor, STATUS_FINISHED) &&
-         receive_status(client->file_descriptor, &status) && status == STATUS_OK;
-}
+/* (finalize_transfer is defined after the SourceFile helpers below.) */
 
-typedef struct {
+typedef struct SourceFile {
   char* path;
   dev_t device;
   ino_t inode;
+  bool skipped; /* receiver reported the file was not written */
 } SourceFile;
 
 static void source_file_destroy(void* item) {
@@ -135,6 +132,8 @@ static void remove_transferred_sources(const Config* config, ArrayList* paths) {
     return;
   for (int i = 0; i < paths->size; i++) {
     SourceFile* source = paths->items[i];
+    if (source->skipped)
+      continue;
     const char* slash = strrchr(source->path, '/');
     const char* leaf = slash ? slash + 1 : source->path;
     char parent[PATH_MAX];
@@ -177,6 +176,7 @@ static SourceFile* source_file_create(const File* file) {
   source->path = str_dup(file->path);
   source->device = st.st_dev;
   source->inode = st.st_ino;
+  source->skipped = false;
   if (!source->path) {
     source_file_destroy(source);
     return NULL;
@@ -203,6 +203,33 @@ static void mark_sender_done(PipelineContextSender* context) {
   mtx_unlock(&context->mutex_progress);
 }
 
+/* Send the final STATUS_FINISHED frame and await the receiver's verdict.
+   When --remove-source-files is active the receiver acknowledges each data
+   file it processed, in send order: STATUS_NEXT means the file was written,
+   STATUS_OK means the file was skipped/unchanged.  Skipped sources are marked
+   so the later removal pass keeps them. */
+static bool finalize_transfer(Client* client, const Config* config, ArrayList* remove_sources) {
+  if (!send_status(client->file_descriptor, STATUS_FINISHED))
+    return false;
+  if (config->remove_source_files && remove_sources) {
+    for (int i = 0; i < remove_sources->size; i++) {
+      Status per_file;
+      if (!receive_status(client->file_descriptor, &per_file))
+        return false;
+      if (per_file == STATUS_ERROR)
+        return false;
+      if (per_file == STATUS_OK) {
+        ((SourceFile*)remove_sources->items[i])->skipped = true;
+      } else if (per_file != STATUS_NEXT) {
+        log_message(LOG_LEVEL_ERROR, "Unexpected per-file status from receiver");
+        return false;
+      }
+    }
+  }
+  Status status;
+  return receive_status(client->file_descriptor, &status) && status == STATUS_OK;
+}
+
 static void pipeline_cancel(PipelineContextSender* context) {
   mtx_lock(&context->mutex_scanner);
   mtx_lock(&context->mutex_loader);
@@ -575,7 +602,7 @@ static int send_chunks_multithreaded(void* pipeline_context) {
         if (send_delete_manifest(client->file_descriptor, context->manifest) != 0)
           goto send_fail;
       }
-      bool ok = finalize_transfer(client);
+      bool ok = finalize_transfer(client, context->config, context->remove_source_files);
       if (ok)
         remove_transferred_sources(context->config, context->remove_source_files);
       mtx_lock(&context->mutex_progress);
@@ -866,7 +893,7 @@ int send_files(Config* config) {
     array_list_delete(manifest);
     manifest = NULL;
   }
-  bool ok = finalize_transfer(client);
+  bool ok = finalize_transfer(client, config, remove_sources);
   if (ok)
     remove_transferred_sources(config, remove_sources);
   if (config->show_progress && !config->quiet)

src/server/receiver.c

@@ -1,6 +1,7 @@
 #include "receiver.h"
 
 #include "chunk.h"
+#include "file_receive.h"
 #include "log.h"
 #include "metadata.h"
 #include "protocol.h"
@@ -8,6 +9,48 @@
 #include <stdlib.h>
 #include <sys/stat.h>
 
+bool receiver_outcomes_append(ReceiverOutcomes* outcomes, unsigned char code) {
+  if (!outcomes)
+    return false;
+  if (outcomes->count == outcomes->capacity) {
+    size_t new_capacity = outcomes->capacity == 0 ? 64 : outcomes->capacity * 2;
+    if (new_capacity < outcomes->capacity)
+      return false;
+    unsigned char* grown = realloc(outcomes->entries, new_capacity);
+    if (!grown)
+      return false;
+    outcomes->entries = grown;
+    outcomes->capacity = new_capacity;
+  }
+  outcomes->entries[outcomes->count++] = code;
+  return true;
+}
+
+void receiver_outcomes_destroy(ReceiverOutcomes* outcomes) {
+  if (!outcomes)
+    return;
+  free(outcomes->entries);
+  outcomes->entries = NULL;
+  outcomes->count = 0;
+  outcomes->capacity = 0;
+}
+
+/* End-of-transfer success frame.  When --remove-source-files was negotiated
+   each processed data file is acknowledged first (STATUS_NEXT = written,
+   STATUS_OK = skipped) so the sender never removes a source the receiver did
+   not actually store.  The frame always ends with a plain STATUS_OK. */
+bool receiver_send_final_success(int fd, const Config* config, const ReceiverOutcomes* outcomes) {
+  if (!config->remove_source_files)
+    return send_status(fd, STATUS_OK);
+  size_t count = outcomes ? outcomes->count : 0;
+  for (size_t i = 0; i < count; i++) {
+    Status per_file = outcomes->entries[i] == FILE_SAVE_WRITTEN ? STATUS_NEXT : STATUS_OK;
+    if (!send_status(fd, per_file))
+      return false;
+  }
+  return send_status(fd, STATUS_OK);
+}
+
 static bool receiver_process_chunk(Chunk* chunk, const ReceiverSink* sink) {
   if (!chunk || !sink || !sink->store_file)
     return false;
@@ -52,7 +95,7 @@ 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) {
+    if (check_size > MAX_RECEIVE_WHOLE_FILE_SIZE) {
       free(check_path);
       send_status(file_descriptor, STATUS_ERROR);
       return false;
@@ -130,8 +173,14 @@ int receiver_process(Config* config, int file_descriptor, const ReceiverSink* si
     log_message(LOG_LEVEL_ERROR, "Did not receive FINISHED Status");
     goto receive_error;
   }
-  if (sink->send_success && !send_status(file_descriptor, STATUS_OK))
+  if (sink->send_success) {
+    if (sink->send_success_frame) {
+      if (!sink->send_success_frame(file_descriptor, sink->context))
         return -1;
+    } else if (!send_status(file_descriptor, STATUS_OK)) {
+      return -1;
+    }
+  }
   return 0;
 
 receive_error:
@@ -140,15 +189,41 @@ receive_error:
   return -1;
 }
 
-static bool receiver_save_file(File* file, void* context) {
-  Config* config = context;
-  bool success =
-      !config->save_to_disk || file_save_to_disk(config->receive_root_directory, file, config);
+/* ---- Single-threaded sink (used by receiver_receive_files) ---- */
+
+typedef struct {
+  Config* config;
+  ReceiverOutcomes outcomes;
+} ReceiverSaveContext;
+
+static bool receiver_save_file(File* file, void* context_pointer) {
+  ReceiverSaveContext* context = context_pointer;
+  FileSaveResult result = FILE_SAVE_ERROR;
+  if (!context->config->save_to_disk) {
+    /* Nothing is stored; report the file as not-written so a
+       --remove-source-files sender keeps its source. */
+    result = FILE_SAVE_SKIPPED;
+  } else {
+    result = file_save_to_disk_full(context->config->receive_root_directory, file, context->config);
+  }
+  if (result != FILE_SAVE_ERROR && context->config->remove_source_files &&
+      !receiver_outcomes_append(&context->outcomes, (unsigned char)result)) {
     file_destroy(file);
-  return success;
+    return false;
+  }
+  file_destroy(file);
+  return result != FILE_SAVE_ERROR;
+}
+
+static bool receiver_send_success_frame(int fd, void* context_pointer) {
+  ReceiverSaveContext* context = context_pointer;
+  return receiver_send_final_success(fd, context->config, &context->outcomes);
 }
 
 int receiver_receive_files(Config* config, int file_descriptor) {
-  ReceiverSink sink = {receiver_save_file, config, true, true};
-  return receiver_process(config, file_descriptor, &sink);
+  ReceiverSaveContext context = {.config = config, .outcomes = {0}};
+  ReceiverSink sink = {receiver_save_file, &context, true, true, receiver_send_success_frame};
+  int ret = receiver_process(config, file_descriptor, &sink);
+  receiver_outcomes_destroy(&context.outcomes);
+  return ret;
 }

src/server/receiver.h

@@ -3,16 +3,37 @@
 
 #include "config.h"
 #include "file.h"
+#include "file_receive.h"
 
 typedef bool (*ReceiverFileSink)(File* file, void* context);
 
+/* Ordered per-file save outcomes for one connection.  One entry is appended
+   for every data-bearing file the receiver processes (in the order the files
+   were sent) so the sender of a --remove-source-files transfer can be told
+   which sources were actually written versus skipped on the receiver. */
+typedef struct {
+  unsigned char* entries; /* FILE_SAVE_WRITTEN or FILE_SAVE_SKIPPED */
+  size_t count;
+  size_t capacity;
+} ReceiverOutcomes;
+
+typedef bool (*ReceiverSuccessFrame)(int fd, void* context);
+
 typedef struct {
   ReceiverFileSink store_file;
   void* context;
   bool send_error;
   bool send_success;
+  /* Emits the end-of-transfer success frame.  When the sender requested
+     --remove-source-files this includes one per-file status per processed
+     data file followed by the final STATUS_OK; otherwise just STATUS_OK. */
+  ReceiverSuccessFrame send_success_frame;
 } ReceiverSink;
 
+bool receiver_outcomes_append(ReceiverOutcomes* outcomes, unsigned char code);
+void receiver_outcomes_destroy(ReceiverOutcomes* outcomes);
+bool receiver_send_final_success(int fd, const Config* config, const ReceiverOutcomes* outcomes);
+
 int receiver_process(Config* config, int file_descriptor, const ReceiverSink* sink);
 int receiver_receive_files(Config* config, int file_descriptor);
 

src/server/server.c

@@ -1,14 +1,12 @@
 #include "config.h"
-#include "chunk.h"
 #include "file.h"
 #include "log.h"
-#include "metadata.h"
 #include "multiprocessing.h"
+#include "protocol.h"
 #include "queue.h"
 #include "receiver.h"
 #include "transport_tcp.h"
 #include "transport_tls.h"
-#include "unistd.h"
 #include "utils.h"
 #include <fcntl.h>
 #include <limits.h>
@@ -16,7 +14,6 @@
 #include <stdio.h>
 #include <stdlib.h>
 #include <string.h>
-#include <sys/stat.h>
 #include <unistd.h>
 #include <openssl/x509.h>
 
@@ -26,6 +23,13 @@ static bool allow_delete;
 static bool allow_unauthenticated;
 static const char* required_client_cn;
 
+/* Aggregate payload bytes the multithreaded receiver may buffer ahead of the
+   slow disk writer.  Receiving one more chunk adds up to ~2 * MAX_CHUNK_SIZE
+   of transient wire/decompression buffers on top of the queued payloads, so
+   this ceiling keeps total per-connection receive memory (decompressed and
+   per-file copied chunk buffers included) within MAX_CONNECTION_MEMORY. */
+#define RECEIVER_QUEUE_MAX_BYTES (MAX_CONNECTION_MEMORY - 2 * MAX_CHUNK_SIZE)
+
 static bool tls_client_identity_allowed(SSL* ssl) {
   if (!ssl || !required_client_cn)
     return false;
@@ -101,140 +105,6 @@ static bool __attribute__((unused)) configure_authorization(const char* root) {
   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;
-        }
-        long long check_mtime_nsec;
-        if (!receive_n_data(fd, &check_mtime_nsec, sizeof(check_mtime_nsec)) ||
-            check_mtime_nsec < 0 || check_mtime_nsec >= 1000000000LL) {
-          free(check_path);
-          send_status(fd, STATUS_ERROR);
-          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);
-        long long old_mtime_nsec = 0;
-        if (has_old) {
-#ifdef __linux__
-          old_mtime_nsec = st.st_mtim.tv_nsec;
-#endif
-        }
-        bool match = !config->ignore_times && has_old &&
-                     (unsigned long long)st.st_size == check_size &&
-                     metadata_mtime_matches(st.st_mtime, old_mtime_nsec, (time_t)check_mtime,
-                                            (long)check_mtime_nsec, config->modify_window);
-        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;
@@ -314,6 +184,7 @@ void handler(int file_descriptor) {
     protocol_session_set_max_alloc(&context->session, config->max_alloc);
     atomic_store(&context->session.total_allocated_bytes,
                  atomic_load(&session.total_allocated_bytes));
+    pipeline_context_receiver_set_queue_byte_limit(context, RECEIVER_QUEUE_MAX_BYTES);
     thrd_t receiver, writer;
     bool receiver_created = thrd_create(&receiver, receive_thread, context) == thrd_success;
     bool writer_created = false;
@@ -342,9 +213,15 @@ void handler(int file_descriptor) {
     int writer_result;
     thrd_join(receiver, &receiver_result);
     thrd_join(writer, &writer_result);
-    send_status(file_descriptor, receiver_result == thrd_success && writer_result == thrd_success
-                                     ? STATUS_OK
-                                     : STATUS_ERROR);
+    bool transfer_ok = receiver_result == thrd_success && writer_result == thrd_success;
+    if (transfer_ok) {
+      if (!receiver_send_final_success(file_descriptor, config, &context->outcomes))
+        transfer_ok = false;
+    } else {
+      send_status(file_descriptor, STATUS_ERROR);
+    }
+    if (!transfer_ok)
+      log_message(LOG_LEVEL_ERROR, "Transfer failed");
     pipeline_context_receiver_destroy(context);
   } else {
     if (receiver_receive_files(config, file_descriptor) != 0)

src/shared/config.c

@@ -139,8 +139,9 @@ static bool validate_received_config(const Config* config) {
          valid_wire_bool(config->use_delete) && valid_wire_bool(config->use_incremental) &&
          valid_wire_bool(config->size_only) && valid_wire_bool(config->ignore_times) &&
          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->remove_source_files) && 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->ignore_existing) &&
@@ -274,11 +275,11 @@ static bool send_delta_fields(int fd, const Config* c) {
 
 static bool send_file_options(int fd, const Config* c) {
   return send_int(fd, c->backup) && send_str(fd, c->backup_dir ? c->backup_dir : "") &&
-         send_int(fd, c->follow_symlinks) && send_int(fd, c->copy_links) &&
-         send_int(fd, c->safe_links) && send_int(fd, c->copy_unsafe_links) &&
-         send_int(fd, c->preserve_hard_links) && send_int(fd, c->preserve_acls) &&
-         send_int(fd, c->preserve_xattrs) && send_int(fd, c->preserve_devices) &&
-         send_int(fd, c->preserve_sparse);
+         send_int(fd, c->remove_source_files) && send_int(fd, c->follow_symlinks) &&
+         send_int(fd, c->copy_links) && send_int(fd, c->safe_links) &&
+         send_int(fd, c->copy_unsafe_links) && send_int(fd, c->preserve_hard_links) &&
+         send_int(fd, c->preserve_acls) && send_int(fd, c->preserve_xattrs) &&
+         send_int(fd, c->preserve_devices) && send_int(fd, c->preserve_sparse);
 }
 
 static bool send_selection_options(int fd, const Config* c) {
@@ -355,8 +356,17 @@ static bool receive_delta_fields(int fd, Config* c) {
 static bool receive_file_options(int fd, Config* c) {
   if (!receive_wire_bool(fd, &c->backup))
     return false;
-  c->backup_dir = receive_str(fd);
-  if (!c->backup_dir)
+  char* backup_dir = receive_str(fd);
+  if (!backup_dir)
+    return false;
+  if (*backup_dir != '\0') {
+    c->backup_dir = backup_dir;
+  } else {
+    /* The sender serializes an unset (NULL) string as "", so canonicalize the
+       empty wire value back to NULL to preserve NULL-vs-empty semantics. */
+    free(backup_dir);
+  }
+  if (!receive_wire_bool(fd, &c->remove_source_files))
     return false;
   bool* flags[] = {&c->follow_symlinks,   &c->copy_links,          &c->safe_links,
                    &c->copy_unsafe_links, &c->preserve_hard_links, &c->preserve_acls,
@@ -386,12 +396,37 @@ static bool receive_selection_options(int fd, Config* c) {
 }
 
 static bool receive_resume_options(int fd, Config* c) {
-  c->temp_dir = receive_str(fd);
-  if (!c->temp_dir || !receive_wire_bool(fd, &c->partial))
+  char* temp_dir = receive_str(fd);
+  if (!temp_dir)
     return false;
-  c->partial_dir = receive_str(fd);
-  c->suffix = c->partial_dir ? receive_str(fd) : NULL;
-  if (!c->partial_dir || !c->suffix || !receive_wire_bool(fd, &c->delete_before))
+  if (*temp_dir != '\0') {
+    c->temp_dir = temp_dir;
+  } else {
+    free(temp_dir);
+  }
+  if (!receive_wire_bool(fd, &c->partial))
+    return false;
+  /* These options have NULL client defaults, so the sender transmits an empty
+     string for "unset".  Canonicalize the empty wire value back to NULL so
+     receivers observe exactly what the client configured (plain --backup, for
+     example, must not look like --backup-dir ""). */
+  char* partial_dir = receive_str(fd);
+  if (!partial_dir)
+    return false;
+  if (*partial_dir != '\0') {
+    c->partial_dir = partial_dir;
+  } else {
+    free(partial_dir);
+  }
+  char* suffix = receive_str(fd);
+  if (!suffix)
+    return false;
+  if (*suffix != '\0') {
+    c->suffix = suffix;
+  } else {
+    free(suffix);
+  }
+  if (!receive_wire_bool(fd, &c->delete_before))
     return false;
   if (!receive_wire_bool(fd, &c->checksum))
     return false;

src/shared/config.h

@@ -147,7 +147,7 @@ typedef struct Config {
   bool skip_compress_set;
 } Config;
 
-#define PROTOCOL_VERSION "2.4.0"
+#define PROTOCOL_VERSION "2.5.0"
 #define DEFAULT_CHUNK_SIZE (10 * 1024 * 1024)
 
 Config* config_create(void);

src/shared/delta.c

@@ -216,6 +216,119 @@ static void free_instructions(DeltaInstruction* instrs, uint32_t count) {
   free(instrs);
 }
 
+/* Sentinel meaning "no signature block" in the lookup index chains.  Block
+ * counts are bounded well below UINT32_MAX, so it doubles as a null link. */
+#define DELTA_NO_BLOCK UINT32_MAX
+
+/* Avalanche mix for the rolling checksum so blocks do not cluster in the
+ * bucket table when the weak checksum has little entropy (e.g. all-zero or
+ * patterned files). */
+static uint32_t delta_adler_mix(uint32_t h) {
+  h ^= h >> 16;
+  h *= 0x7feb352dU;
+  h ^= h >> 15;
+  h *= 0x846ca68bU;
+  h ^= h >> 16;
+  return h;
+}
+
+/* Smallest power of two >= v.  v must be non-zero. */
+static uint32_t delta_next_pow2(uint32_t v) {
+  v--;
+  v |= v >> 1;
+  v |= v >> 2;
+  v |= v >> 4;
+  v |= v >> 8;
+  v |= v >> 16;
+  return v + 1;
+}
+
+/* Build a hash index over sig->blocks keyed by the (mixed) rolling checksum.
+ * All blocks sharing an Adler-32 value land in the same bucket; collisions
+ * are chained through a single contiguous allocation:
+ *
+ *   [0, bucket_count)                 heads (first block per bucket)
+ *   [bucket_count, 2*bucket_count)    tails (last block per bucket)
+ *   [2*bucket_count, ...)             per-block chain links
+ *
+ * Blocks are inserted in ascending index order so every bucket chain is
+ * ordered exactly like the historical linear scan.  Returns the base pointer
+ * (also the heads array) or NULL when no index could be allocated; callers
+ * then fall back to the linear scan. */
+static uint32_t* delta_build_index(const DeltaSignature* sig, uint32_t bucket_count) {
+  if (sig->block_count == 0 || bucket_count == 0)
+    return NULL;
+
+  size_t entries = (size_t)2 * bucket_count + sig->block_count;
+  if (entries > SIZE_MAX / sizeof(uint32_t))
+    return NULL;
+
+  uint32_t* index = protocol_alloc(entries * sizeof(uint32_t));
+  if (!index)
+    return NULL;
+
+  uint32_t* heads = index;
+  uint32_t* tails = index + bucket_count;
+  uint32_t* next = index + 2 * bucket_count;
+  uint32_t mask = bucket_count - 1;
+
+  memset(heads, 0xFF, (size_t)bucket_count * sizeof(uint32_t));
+  memset(tails, 0xFF, (size_t)bucket_count * sizeof(uint32_t));
+
+  for (uint32_t j = 0; j < sig->block_count; j++) {
+    uint32_t b = delta_adler_mix(sig->blocks[j].adler32) & mask;
+    if (heads[b] == DELTA_NO_BLOCK)
+      heads[b] = j;
+    else
+      next[tails[b]] = j;
+    tails[b] = j;
+    next[j] = DELTA_NO_BLOCK;
+  }
+  return index;
+}
+
+/* Locate the signature block matching the byte window at new_data[i].
+ *
+ * Mirrors the original per-window behaviour exactly: only a full block_size
+ * window can match, candidates are accepted only when the weak (Adler-32) and
+ * strong (xxHash32) checksums both agree, and the lowest block index wins so
+ * the emitted op stream is byte-identical to the linear scan.  When heads is
+ * non-NULL the candidate set is reached through the bucket index (expected
+ * O(1) per window); otherwise an exact linear scan is used. */
+static uint32_t delta_find_match(const uint8_t* window, uint32_t window_len, uint32_t adler,
+                                 bool full_window, const DeltaSignature* sig, const uint32_t* heads,
+                                 const uint32_t* next, uint32_t mask) {
+  if (!full_window || sig->block_count == 0)
+    return DELTA_NO_BLOCK;
+
+  if (heads) {
+    uint32_t b = delta_adler_mix(adler) & mask;
+    uint32_t window_xxh = 0;
+    bool have_xxh = false;
+    for (uint32_t j = heads[b]; j != DELTA_NO_BLOCK; j = next[j]) {
+      if (sig->blocks[j].adler32 != adler)
+        continue;
+      if (!have_xxh) {
+        window_xxh = delta_xxhash32(window, window_len);
+        have_xxh = true;
+      }
+      if (window_xxh == sig->blocks[j].xxhash)
+        return j;
+    }
+    return DELTA_NO_BLOCK;
+  }
+
+  /* Fallback used when the index could not be allocated. */
+  for (uint32_t j = 0; j < sig->block_count; j++) {
+    if (sig->blocks[j].adler32 == adler) {
+      uint32_t window_xxh = delta_xxhash32(window, window_len);
+      if (window_xxh == sig->blocks[j].xxhash)
+        return j;
+    }
+  }
+  return DELTA_NO_BLOCK;
+}
+
 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 || !sig->blocks || new_file_size == 0 || block_size == 0 ||
@@ -230,6 +343,25 @@ Delta* delta_compute(const void* new_file_data, uint64_t new_file_size, const De
   if (!instrs)
     return NULL;
 
+  /* Build a one-time bucket index over the signature blocks keyed by the weak
+   * checksum.  This turns the per-byte-window candidate lookup from an
+   * O(block_count) linear scan into an expected O(1) probe, which dominates
+   * the cost for large mostly-matching files (the diff steps one byte at a
+   * time through changed regions).  On allocation failure the probe falls back
+   * to the original linear scan, so behaviour is unchanged under memory
+   * pressure. */
+  uint32_t* index = NULL;
+  const uint32_t* chain_next = NULL;
+  uint32_t mask = 0;
+  if (sig->block_count > 0) {
+    uint32_t bucket_count = delta_next_pow2(sig->block_count);
+    index = delta_build_index(sig, bucket_count);
+    if (index) {
+      chain_next = index + 2 * bucket_count;
+      mask = bucket_count - 1;
+    }
+  }
+
   uint64_t literal_start = 0;
   bool has_literal = false;
 
@@ -264,13 +396,13 @@ Delta* delta_compute(const void* new_file_data, uint64_t new_file_size, const De
     }
 
     bool matched = false;
-    for (uint32_t j = 0; j < sig->block_count; j++) {
-      if (adler == sig->blocks[j].adler32 && full_window) {
-        uint32_t xxh = delta_xxhash32(new_data + i, window_len);
-        if (xxh == sig->blocks[j].xxhash) {
+    uint32_t match_block = delta_find_match(new_data + i, window_len, adler, full_window, sig,
+                                            index, chain_next, mask);
+    if (match_block != DELTA_NO_BLOCK) {
       if (has_literal) {
         if (!flush_literal(&instrs, &capacity, &count, new_data, literal_start, i)) {
           free_instructions(instrs, count);
+          free(index);
           return NULL;
         }
         has_literal = false;
@@ -278,10 +410,11 @@ Delta* delta_compute(const void* new_file_data, uint64_t new_file_size, const De
 
       if (!ensure_capacity(&instrs, &capacity, count)) {
         free_instructions(instrs, count);
+        free(index);
         return NULL;
       }
       instrs[count].type = DELTA_INSTR_BLOCK_MATCH;
-          instrs[count].match.block_index = j;
+      instrs[count].match.block_index = match_block;
       instrs[count].match.block_offset = 0;
       instrs[count].match.length = window_len;
       count++;
@@ -289,9 +422,6 @@ Delta* delta_compute(const void* new_file_data, uint64_t new_file_size, const De
       i += window_len;
       rolling_valid = false;
       matched = true;
-          break;
-        }
-      }
     }
 
     if (!matched) {
@@ -303,6 +433,8 @@ Delta* delta_compute(const void* new_file_data, uint64_t new_file_size, const De
     }
   }
 
+  free(index);
+
   if (has_literal) {
     if (!flush_literal(&instrs, &capacity, &count, new_data, literal_start, new_file_size)) {
       free_instructions(instrs, count);

src/shared/file.c

@@ -348,10 +348,25 @@ static bool file_to_disk_secure_impl(const char* path, const void* data,
       if (newer) {
         ok = true;
       } else {
-        if (!sparse || data_size == 0 || ftruncate(fd, (off_t)data_size) == 0)
+        /* In-place overwrites: pre-size sparse targets and always trim the
+           file to the new payload length afterwards so shorter payloads can
+           never leave stale trailing bytes from a previous version. */
+        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)
+        if (ok)
+          ok = ftruncate(fd, (off_t)data_size) == 0;
+        /* Normalize the mode: apply the metadata-derived safe mode when the
+           sender supplied metadata (setuid/setgid/sticky are never honored);
+           otherwise fall back to a safe default so dangerous bits on an
+           existing destination cannot survive an overwrite. */
+        if (ok) {
+          if (metadata)
             ok = file_restore_metadata_fd(fd, metadata, preserve_executability);
+          else if (fchmod(fd, S_IRUSR | S_IWUSR | S_IRGRP | S_IROTH) != 0)
+            ok = false;
+        }
         if (ok && use_fsync)
           ok = fsync(fd) == 0;
       }

src/shared/file_receive.c

@@ -20,9 +20,14 @@
 #include "utils.h"
 
 #define MAX_SERVER_DELETE_COUNT 100000U
-#define MAX_FILE_DATA_SIZE MAX_RECEIVE_FILE_SIZE
+#define MAX_FILE_DATA_SIZE MAX_RECEIVE_WHOLE_FILE_SIZE
 
 bool file_save_to_disk(const char* root_directory, const File* file, const Config* config) {
+  return file_save_to_disk_full(root_directory, file, config) != FILE_SAVE_ERROR;
+}
+
+FileSaveResult file_save_to_disk_full(const char* root_directory, const File* file,
+                                      const Config* config) {
   /* Backups are incompatible with ignore-existing: moving the entry first
      would make a concurrent no-replace commit overwrite its old name. */
   bool backup_enabled = config && config->backup && !config->ignore_existing;
@@ -32,6 +37,7 @@ 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;
+  bool use_partial_root = partial_dir && config && config->partial;
   char *confined_backup = NULL, *confined_partial = NULL, *disk_path = NULL;
   char* destination_path = NULL;
   char *backup_path = NULL, *parent_copy = NULL;
@@ -42,23 +48,22 @@ bool file_save_to_disk(const char* root_directory, const File* file, const Confi
        (!backup_suffix || backup_suffix[0] == '\0' || strchr(backup_suffix, '/') != NULL ||
         strcmp(backup_suffix, ".") == 0 || strcmp(backup_suffix, "..") == 0))) {
     log_message(LOG_LEVEL_ERROR, "Invalid file or path received");
-    return false;
+    return FILE_SAVE_ERROR;
   }
 
   /* These options arrive from the client.  They are names below the server
      root, never independent filesystem roots. */
   if ((backup_dir && (backup_dir[0] == '/' || has_path_traversal(backup_dir))) ||
       (partial_dir && (partial_dir[0] == '/' || has_path_traversal(partial_dir))))
-    return false;
+    return FILE_SAVE_ERROR;
   if (backup_dir && !(confined_backup = path_cat(root_directory, backup_dir)))
-    return false;
+    return FILE_SAVE_ERROR;
   if (partial_dir && !(confined_partial = path_cat(root_directory, partial_dir))) {
     free(confined_backup);
-    return false;
+    return FILE_SAVE_ERROR;
   }
 
-  const char* actual_root =
-      (partial_dir && config && config->partial) ? confined_partial : root_directory;
+  const char* actual_root = use_partial_root ? confined_partial : root_directory;
   destination_path = path_cat(root_directory, file->path);
   disk_path = path_cat(actual_root, file->path);
   if (destination_path == NULL || disk_path == NULL) {
@@ -66,7 +71,7 @@ bool file_save_to_disk(const char* root_directory, const File* file, const Confi
     free(confined_partial);
     free(destination_path);
     free(disk_path);
-    return false;
+    return FILE_SAVE_ERROR;
   }
 
   /* --existing checks the final destination, not a temporary partial path. */
@@ -75,7 +80,7 @@ bool file_save_to_disk(const char* root_directory, const File* file, const Confi
     free(confined_partial);
     free(destination_path);
     free(disk_path);
-    return true;
+    return FILE_SAVE_SKIPPED;
   }
 
   /* --ignore-existing checks the final destination before partial files or
@@ -87,34 +92,40 @@ bool file_save_to_disk(const char* root_directory, const File* file, const Confi
       free(confined_partial);
       free(destination_path);
       free(disk_path);
-      return true;
+      return FILE_SAVE_SKIPPED;
     }
   }
-  free(destination_path);
-  destination_path = NULL;
 
   /* --update is receiver-side policy: never replace a newer destination.
-     The secure stat does not require read permission on the destination. */
-  if (config && config->update && file_destination_is_newer_secure(disk_path, file->metadata)) {
+     In partial-dir mode the entry that would be replaced is the real
+     destination, not the temporary partial file.  The secure stat does not
+     require read permission on the destination. */
+  const char* update_target = use_partial_root ? destination_path : disk_path;
+  if (config && config->update && file_destination_is_newer_secure(update_target, file->metadata)) {
     free(confined_backup);
     free(confined_partial);
+    free(destination_path);
     free(disk_path);
-    return true;
+    return FILE_SAVE_SKIPPED;
   }
 
   if (backup_enabled) {
+    /* Back up the entry that the incoming write will replace.  When writing
+       through a partial dir the pre-existing destination file is the one to
+       preserve; any stale partial file is overwritten without a backup. */
+    const char* replace_target = use_partial_root ? destination_path : disk_path;
     struct stat backup_stat;
-    if (file_stat_secure(disk_path, &backup_stat)) {
+    if (file_stat_secure(replace_target, &backup_stat)) {
       if (backup_dir) {
         backup_path = path_cat(confined_backup, file->path);
       } else {
-        size_t path_len = strlen(disk_path);
+        size_t path_len = strlen(replace_target);
         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, replace_target, path_len);
           memcpy(backup_path + path_len, backup_suffix, suffix_len + 1);
         }
       }
@@ -125,7 +136,7 @@ bool file_save_to_disk(const char* root_directory, const File* file, const Confi
         goto fail;
       free(parent_copy);
       parent_copy = NULL;
-      if (!file_rename_secure(disk_path, backup_path))
+      if (!file_rename_secure(replace_target, backup_path))
         goto fail;
       free(backup_path);
       backup_path = NULL;
@@ -149,13 +160,25 @@ bool file_save_to_disk(const char* root_directory, const File* file, const Confi
                 : file_to_disk_secure_with_fsync(disk_path, file->data->data, file->data->size,
                                                  inplace, sparse, metadata, preserve_executability,
                                                  config && config->use_fsync);
+  if (!ok)
+    goto fail;
+
+  /* --partial --partial-dir writes the complete file under the partial dir so
+     interrupted transfers leave a resumable copy there.  Once the file is
+     fully written it must be atomically installed at the real destination;
+     otherwise completed transfers would linger under the partial dir. */
+  if (use_partial_root) {
+    if (!file_rename_secure(disk_path, destination_path))
+      goto fail;
+  }
+
   free(parent_copy);
   free(backup_path);
   free(confined_backup);
   free(confined_partial);
   free(destination_path);
   free(disk_path);
-  return ok;
+  return FILE_SAVE_WRITTEN;
 
 fail:
   free(parent_copy);
@@ -164,13 +187,15 @@ fail:
   free(confined_partial);
   free(destination_path);
   free(disk_path);
-  return false;
+  return FILE_SAVE_ERROR;
 }
 
 static File* receive_delta_file(int fd, const Config* config, const char* check_path,
                                 void* old_data, unsigned long long old_size, bool* failed) {
-  if (!old_data)
+  if (!old_data) {
+    *failed = true;
     return NULL;
+  }
 
   DeltaSignature* sig = delta_signature_create(old_data, old_size, config->delta_block_size);
   if (!sig) {
@@ -206,7 +231,7 @@ static File* receive_delta_file(int fd, const Config* config, const char* check_
   }
 
   if (resp == STATUS_DELTA_DATA) {
-    Data* delta_data = receive_data_limited(fd, MAX_RECEIVE_FILE_SIZE);
+    Data* delta_data = receive_data_limited(fd, MAX_RECEIVE_WHOLE_FILE_SIZE);
     if (!delta_data) {
       delta_signature_destroy(sig);
       free(old_data);
@@ -219,7 +244,7 @@ static File* receive_delta_file(int fd, const Config* config, const char* check_
         !compression_should_skip_with_suffixes(
             check_path, config->skip_compress_suffixes,
             config->skip_compress_set ? config->skip_compress_count : -1)) {
-      raw_delta = data_decompress_limited(delta_data, MAX_RECEIVE_FILE_SIZE);
+      raw_delta = data_decompress_limited(delta_data, MAX_RECEIVE_WHOLE_FILE_SIZE);
       data_destroy(delta_data);
       if (!raw_delta) {
         free(old_data);
@@ -239,11 +264,12 @@ static File* receive_delta_file(int fd, const Config* config, const char* check_
     }
 
     uint64_t new_size = delta->new_file_size;
-    if (new_size > MAX_RECEIVE_FILE_SIZE || new_size > SIZE_MAX) {
+    if (new_size > MAX_RECEIVE_WHOLE_FILE_SIZE || new_size > SIZE_MAX) {
       delta_destroy(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);
@@ -284,6 +310,7 @@ static File* receive_delta_file(int fd, const Config* config, const char* check_
       free(old_data);
       delta_signature_destroy(sig);
       send_status(fd, STATUS_ERROR);
+      *failed = true;
       return NULL;
     }
     data_destroy(file->data);
@@ -314,7 +341,7 @@ static File* receive_delta_file(int fd, const Config* config, const char* check_
       }
     }
 
-    Data* file_data = receive_data_limited(fd, MAX_RECEIVE_FILE_SIZE);
+    Data* file_data = receive_data_limited(fd, MAX_RECEIVE_WHOLE_FILE_SIZE);
     if (file_data == NULL) {
       file_destroy(file);
       *failed = true;
@@ -325,7 +352,7 @@ static File* receive_delta_file(int fd, const Config* config, const char* check_
         !compression_should_skip_with_suffixes(
             file->path, config->skip_compress_suffixes,
             config->skip_compress_set ? config->skip_compress_count : -1)) {
-      Data* uncompressed = data_decompress_limited(file_data, MAX_RECEIVE_FILE_SIZE);
+      Data* uncompressed = data_decompress_limited(file_data, MAX_RECEIVE_WHOLE_FILE_SIZE);
       data_destroy(file_data);
       if (uncompressed == NULL) {
         file_destroy(file);
@@ -336,6 +363,7 @@ static File* receive_delta_file(int fd, const Config* config, const char* check_
         data_destroy(uncompressed);
         file_destroy(file);
         send_status(fd, STATUS_ERROR);
+        *failed = true;
         return NULL;
       }
       file_data = uncompressed;
@@ -384,7 +412,7 @@ File* receive_incremental_check(int fd, const Config* config, bool* skipped) {
     return NULL;
   }
 
-  if (check_size > MAX_RECEIVE_FILE_SIZE) {
+  if (check_size > MAX_RECEIVE_WHOLE_FILE_SIZE) {
     free(check_path);
     send_status(fd, STATUS_ERROR);
     return NULL;
@@ -405,6 +433,9 @@ File* receive_incremental_check(int fd, const Config* config, bool* skipped) {
     send_status(fd, STATUS_ERROR);
     return NULL;
   }
+
+  /* Open the existing destination entry (if any) once and keep the descriptor
+     until the quick-check below decides whether the old contents are needed. */
   struct stat st;
   bool has_old_file = false;
   int old_fd = -1;
@@ -416,9 +447,34 @@ File* receive_incremental_check(int fd, const Config* config, bool* skipped) {
     close(parent_fd);
     has_old_file = old_fd >= 0 && fstat(old_fd, &st) == 0 && S_ISREG(st.st_mode);
   }
+  if (!has_old_file && old_fd >= 0) {
+    close(old_fd);
+    old_fd = -1;
+  }
   unsigned long long old_size = has_old_file ? (unsigned long long)st.st_size : 0;
+
+  /* Decide from metadata alone whether the receiver already holds the file
+     the sender is offering.  The old contents are only read into memory when
+     a checksum comparison or a delta transfer actually requires them. */
+  bool size_equal = has_old_file && old_size == check_size;
+  bool match_by_metadata = false;
+  if (size_equal && !config->ignore_times && !config->size_only) {
+    long long old_mtime_nsec = 0;
+#ifdef __linux__
+    old_mtime_nsec = st.st_mtim.tv_nsec;
+#endif
+    match_by_metadata = metadata_mtime_matches(st.st_mtime, old_mtime_nsec, (time_t)check_mtime,
+                                               (long)check_mtime_nsec, config->modify_window);
+  }
+
+  bool try_delta = config->use_delta && !config->whole_file && has_old_file &&
+                   delta_should_attempt(old_size, check_size, config->delta_max_file_size);
+  bool checksum_needs_read = size_equal && !config->ignore_times && config->checksum;
+  bool need_old_data = checksum_needs_read || try_delta;
+
   void* old_data = NULL;
-  if (has_old_file && old_size > 0 && old_size <= MAX_RECEIVE_FILE_SIZE && old_size <= SIZE_MAX) {
+  if (need_old_data && has_old_file && old_size > 0 && old_size <= MAX_RECEIVE_WHOLE_FILE_SIZE &&
+      old_size <= SIZE_MAX) {
     old_data = protocol_alloc((size_t)old_size);
     if (old_data) {
       size_t got = 0;
@@ -433,73 +489,63 @@ File* receive_incremental_check(int fd, const Config* config, bool* skipped) {
       }
     }
   }
-  if (old_fd >= 0) {
-    close(old_fd);
-  }
 
-  bool match =
-      !config->ignore_times && has_old_file && (unsigned long long)st.st_size == check_size;
-  if (match && config->checksum) {
-    uint64_t old_checksum = old_size == 0 ? delta_xxhash64("", 0) : 0;
-    if (old_data)
-      old_checksum = delta_xxhash64(old_data, (size_t)old_size);
-    match = (old_size == 0 || old_data) && old_checksum == check_checksum;
-    free(old_data);
-    old_data = NULL;
-  } else if (match && !config->size_only) {
-    long long old_mtime_nsec = 0;
-#ifdef __linux__
-    old_mtime_nsec = st.st_mtim.tv_nsec;
-#endif
-    match = metadata_mtime_matches(st.st_mtime, old_mtime_nsec, (time_t)check_mtime,
-                                   (long)check_mtime_nsec, config->modify_window);
+  /* Quick-skip decision.  If no content comparison is required this is final
+     and the old file was never read; if the read failed the file is not
+     skipped and the transfer proceeds with the full new contents. */
+  bool match = false;
+  if (checksum_needs_read) {
+    if (old_size == 0)
+      match = delta_xxhash64("", 0) == check_checksum;
+    else
+      match = old_data != NULL && delta_xxhash64(old_data, (size_t)old_size) == check_checksum;
+  } else if (size_equal && !config->ignore_times) {
+    match = config->size_only || match_by_metadata;
   }
 
   if (match) {
     free(old_data);
     if (!send_status(fd, STATUS_OK)) {
+      close(old_fd);
       free(full_path);
       free(check_path);
       return NULL;
     }
+    close(old_fd);
     free(full_path);
     free(check_path);
     *skipped = true;
     return NULL;
   }
 
-  bool try_delta = config->use_delta && !config->whole_file && has_old_file && old_data != NULL &&
-                   delta_should_attempt(old_size, check_size, config->delta_max_file_size);
-
-  if (try_delta) {
+  if (try_delta && old_data != NULL) {
     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) {
+      close(old_fd);
       free(full_path);
       free(check_path);
       return delta_file;
     }
     if (delta_failed) {
+      close(old_fd);
       free(full_path);
       free(check_path);
       return NULL;
     }
+  }
   free(old_data);
   old_data = NULL;
-    try_delta = false;
-  }
 
-  if (!try_delta) {
-    free(old_data);
-    old_data = NULL;
   if (!send_status(fd, STATUS_NEXT)) {
+    close(old_fd);
     free(full_path);
     free(check_path);
     return NULL;
   }
-  }
+  close(old_fd);
 
   File* file = file_create(check_path);
   free(check_path);
@@ -517,7 +563,7 @@ File* receive_incremental_check(int fd, const Config* config, bool* skipped) {
     }
   }
 
-  Data* file_data = receive_data_limited(fd, MAX_RECEIVE_FILE_SIZE);
+  Data* file_data = receive_data_limited(fd, MAX_RECEIVE_WHOLE_FILE_SIZE);
   if (file_data == NULL) {
     file_destroy(file);
     return NULL;
@@ -527,7 +573,7 @@ File* receive_incremental_check(int fd, const Config* config, bool* skipped) {
       !compression_should_skip_with_suffixes(file->path, config->skip_compress_suffixes,
                                              config->skip_compress_set ? config->skip_compress_count
                                                                        : -1)) {
-    Data* uncompressed = data_decompress_limited(file_data, MAX_RECEIVE_FILE_SIZE);
+    Data* uncompressed = data_decompress_limited(file_data, MAX_RECEIVE_WHOLE_FILE_SIZE);
     data_destroy(file_data);
     if (uncompressed == NULL) {
       file_destroy(file);
@@ -536,7 +582,6 @@ File* receive_incremental_check(int fd, const Config* config, bool* skipped) {
     if (uncompressed->size > MAX_FILE_DATA_SIZE) {
       data_destroy(uncompressed);
       file_destroy(file);
-      send_status(fd, STATUS_ERROR);
       return NULL;
     }
     file_data = uncompressed;
@@ -571,7 +616,7 @@ File* file_receive(const Config* config, int file_descriptor) {
       return NULL;
     }
   }
-  Data* file_data = receive_data_limited(file_descriptor, MAX_RECEIVE_FILE_SIZE);
+  Data* file_data = receive_data_limited(file_descriptor, MAX_RECEIVE_WHOLE_FILE_SIZE);
   if (file_data == NULL) {
     file_destroy(file);
     return NULL;
@@ -580,7 +625,7 @@ File* file_receive(const Config* config, int file_descriptor) {
       !compression_should_skip_with_suffixes(file->path, config->skip_compress_suffixes,
                                              config->skip_compress_set ? config->skip_compress_count
                                                                        : -1)) {
-    Data* file_data_uncompressed = data_decompress_limited(file_data, MAX_RECEIVE_FILE_SIZE);
+    Data* file_data_uncompressed = data_decompress_limited(file_data, MAX_RECEIVE_WHOLE_FILE_SIZE);
     data_destroy(file_data);
     if (file_data_uncompressed == NULL) {
       file_destroy(file);

src/shared/file_receive.h

@@ -10,6 +10,14 @@
 File* file_receive(const Config* config, int file_descriptor);
 File* receive_incremental_check(int fd, const Config* config, bool* skipped);
 int receive_manifest(int fd, const Config* config, int* next_status);
+
+/* Outcome of a single file_save_to_disk operation.  The receiver needs to
+   distinguish "written" from "skipped" so --remove-source-files can be told
+   which sources were actually stored. */
+typedef enum { FILE_SAVE_ERROR = 0, FILE_SAVE_WRITTEN = 1, FILE_SAVE_SKIPPED = 2 } FileSaveResult;
+
+FileSaveResult file_save_to_disk_full(const char* root_directory, const File* file,
+                                      const Config* config);
 bool file_save_to_disk(const char* root_directory, const File* file, const Config* config);
 
 #endif

src/shared/multiprocessing.c

@@ -105,9 +105,14 @@ PipelineContextReceiver* pipeline_context_receiver_create(Config* config, Queue*
   context->queue = queue;
   context->file_descriptor = file_descriptor;
   context->ssl = ssl;
+  context->outcomes.entries = NULL;
+  context->outcomes.count = 0;
+  context->outcomes.capacity = 0;
   protocol_session_init(&context->session, file_descriptor, file_descriptor);
   protocol_session_set_ssl(&context->session, ssl);
   context->receiver_done = false;
+  context->queued_bytes = 0;
+  context->max_queue_bytes = 0;
   atomic_init(&context->cancelled, false);
   int init = 0;
   if (mtx_init(&context->mutex, mtx_plain) != thrd_success)
@@ -137,21 +142,81 @@ fail:
 void pipeline_context_receiver_destroy(PipelineContextReceiver* context) {
   config_delete(context->config);
   queue_destroy(context->queue);
+  receiver_outcomes_destroy(&context->outcomes);
   mtx_destroy(&context->mutex);
   cnd_destroy(&context->condition_not_full);
   cnd_destroy(&context->condition_not_empty);
   free(context);
 }
 
-static bool receiver_enqueue_file(File* file, void* context_pointer) {
-  PipelineContextReceiver* context = context_pointer;
-  if (queue_enqueue_multithreaded_cancel(context->queue, file, &context->mutex,
-                                         &context->condition_not_empty,
-                                         &context->condition_not_full, &context->cancelled))
-    return true;
+void pipeline_context_receiver_set_queue_byte_limit(PipelineContextReceiver* context,
+                                                    size_t max_bytes) {
+  if (context == NULL)
+    return;
+  mtx_lock(&context->mutex);
+  context->max_queue_bytes = max_bytes;
+  context->queued_bytes = 0;
+  cnd_broadcast(&context->condition_not_full);
+  mtx_unlock(&context->mutex);
+}
+
+void pipeline_context_receiver_note_bytes_released(PipelineContextReceiver* context,
+                                                   size_t released_bytes) {
+  if (context == NULL || context->max_queue_bytes == 0 || released_bytes == 0)
+    return;
+  mtx_lock(&context->mutex);
+  if (released_bytes >= context->queued_bytes)
+    context->queued_bytes = 0;
+  else
+    context->queued_bytes -= released_bytes;
+  cnd_signal(&context->condition_not_full);
+  mtx_unlock(&context->mutex);
+}
+
+bool pipeline_context_receiver_enqueue_file(PipelineContextReceiver* context, File* file) {
+  if (context == NULL || file == NULL)
+    return false;
+  size_t file_bytes = file->data ? file->data->size : 0;
+  mtx_lock(&context->mutex);
+  while (!atomic_load(&context->cancelled)) {
+    bool blocked_by_count = queue_is_full(context->queue);
+    bool blocked_by_budget = false;
+    if (context->max_queue_bytes > 0) {
+      size_t budget = context->max_queue_bytes;
+      size_t used = context->queued_bytes;
+      if (used >= budget) {
+        blocked_by_budget = true;
+      } else if (file_bytes > budget - used) {
+        /* A single payload larger than the whole budget (not possible with
+           the per-file receive cap) is only admitted to an empty pipeline so
+           the wait can never deadlock. */
+        blocked_by_budget = used != 0;
+      }
+    }
+    if (!blocked_by_count && !blocked_by_budget)
+      break;
+    cnd_wait(&context->condition_not_full, &context->mutex);
+  }
+  if (atomic_load(&context->cancelled)) {
+    mtx_unlock(&context->mutex);
     file_destroy(file);
     return false;
   }
+  if (!queue_enqueue(context->queue, file)) {
+    mtx_unlock(&context->mutex);
+    file_destroy(file);
+    return false;
+  }
+  context->queued_bytes += file_bytes;
+  cnd_signal(&context->condition_not_empty);
+  mtx_unlock(&context->mutex);
+  return true;
+}
+
+static bool receiver_enqueue_file(File* file, void* context_pointer) {
+  PipelineContextReceiver* context = (PipelineContextReceiver*)context_pointer;
+  return pipeline_context_receiver_enqueue_file(context, file);
+}
 
 static void receiver_thread_fail(PipelineContextReceiver* context) {
   mtx_lock(&context->mutex);
@@ -170,7 +235,7 @@ int receive_thread(void* pipeline_context) {
   const Config* config = context->config;
   mtx_unlock(&context->mutex);
 
-  ReceiverSink sink = {receiver_enqueue_file, context, false, false};
+  ReceiverSink sink = {receiver_enqueue_file, context, false, false, NULL};
   if (receiver_process((Config*)config, file_descriptor, &sink) != 0) {
     receiver_thread_fail(context);
     protocol_session_unbind();
@@ -211,8 +276,30 @@ int write_thread(void* pipeline_context) {
       protocol_session_unbind();
       return thrd_success;
     }
-    if (save_to_disk && !file_save_to_disk(root_directory, file, context->config)) {
+    size_t file_bytes = file->data ? file->data->size : 0;
+    FileSaveResult result = FILE_SAVE_SKIPPED;
+    if (save_to_disk) {
+      result = file_save_to_disk_full(root_directory, file, context->config);
+      if (result == FILE_SAVE_ERROR) {
         file_destroy(file);
+        pipeline_context_receiver_note_bytes_released(context, file_bytes);
+        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);
+        free(root_directory);
+        protocol_session_unbind();
+        return thrd_error;
+      }
+    }
+    /* Record the per-file outcome so a --remove-source-files sender learns
+       which sources were actually written versus skipped on the receiver. */
+    if (context->config->remove_source_files &&
+        !receiver_outcomes_append(&context->outcomes, (unsigned char)result)) {
+      file_destroy(file);
+      pipeline_context_receiver_note_bytes_released(context, file_bytes);
       mtx_lock(&context->mutex);
       atomic_store(&context->cancelled, true);
       context->receiver_done = true;
@@ -224,5 +311,6 @@ int write_thread(void* pipeline_context) {
       return thrd_error;
     }
     file_destroy(file);
+    pipeline_context_receiver_note_bytes_released(context, file_bytes);
   }
 }

src/shared/multiprocessing.h

@@ -9,6 +9,7 @@
 #include "file.h"
 #include "protocol.h"
 #include "queue.h"
+#include "receiver.h"
 #include <openssl/ssl.h>
 
 typedef struct {
@@ -40,11 +41,20 @@ typedef struct PipelineContextReceiver {
   int file_descriptor;
   SSL* ssl;
   ProtocolSession session;
+  ReceiverOutcomes outcomes;
   mtx_t mutex;
   cnd_t condition_not_full;
   cnd_t condition_not_empty;
   bool receiver_done;
   atomic_bool cancelled;
+  /* Aggregate payload bytes that have been received but not yet released by
+     the disk writer (queued or in the writer's hand).  Guarded by `mutex`.
+     When `max_queue_bytes` is non-zero the receiver blocks before enqueuing
+     once this total would exceed it, so decompressed/copied file payloads
+     buffered ahead of a slow disk writer respect the per-connection memory
+     budget instead of growing without bound. */
+  size_t queued_bytes;
+  size_t max_queue_bytes;
 } PipelineContextReceiver;
 
 PipelineContextSender* pipeline_context_sender_create(Config* config, Queue* queue_scanner,
@@ -53,6 +63,18 @@ void pipeline_context_sender_destroy(PipelineContextSender* context);
 PipelineContextReceiver* pipeline_context_receiver_create(Config* config, Queue* queue_receiver,
                                                           int file_descriptor, SSL* ssl);
 void pipeline_context_receiver_destroy(PipelineContextReceiver* context);
+/* Bound the bytes buffered ahead of the disk writer (see max_queue_bytes). */
+void pipeline_context_receiver_set_queue_byte_limit(PipelineContextReceiver* context,
+                                                    size_t max_bytes);
+/* Blocking enqueue used by the receive pipeline sink.  Blocks while the queue
+   is full by element count or when adding `file` would push queued_bytes over
+   the configured byte limit; waits until the disk writer releases bytes.
+   Takes ownership of `file` on success and destroys it on failure/cancel. */
+bool pipeline_context_receiver_enqueue_file(PipelineContextReceiver* context, File* file);
+/* Account for `released_bytes` of payload memory that has been freed by the
+   disk writer, unblocking a receiver that is waiting on the byte limit. */
+void pipeline_context_receiver_note_bytes_released(PipelineContextReceiver* context,
+                                                   size_t released_bytes);
 int receive_thread(void* pipeline_context);
 int write_thread(void* pipeline_context);
 #endif

src/shared/protocol.c

@@ -14,7 +14,6 @@
 
 #define RECEIVE_TIMEOUT_SEC 60 /* 60 second per-message timeout */
 #define SEND_TIMEOUT_SEC 60
-#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;

src/shared/protocol.h

@@ -9,10 +9,16 @@
 /* Maximum allowed string size for receive_str (64 KB) */
 #define MAX_STRING_SIZE (64 * 1024)
 
-/* 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 uncompressed file payload accepted by the receiver's whole-file
+ * paths.  A single whole file is charged against the per-connection memory
+ * reservation (MAX_CONNECTION_MEMORY) and against the server allocation
+ * ceiling (MAX_SERVER_ALLOC), so this mirrors those 256 MB bounds rather than
+ * the older 64 MB chunk-era cap.  Chunk-serialized payloads keep their own
+ * 64 MB cap (MAX_CHUNK_SIZE). */
+#define MAX_RECEIVE_WHOLE_FILE_SIZE (256ULL * 1024 * 1024)
+
+/* Maximum allowed data payload size for receive_data (whole-file bound) */
+#define MAX_DATA_PAYLOAD_SIZE MAX_RECEIVE_WHOLE_FILE_SIZE
 
 /* Maximum chunk size (64 MB) — prevents unbounded allocation from the wire */
 #define MAX_CHUNK_SIZE (64ULL * 1024 * 1024)
@@ -22,6 +28,10 @@
 #define DEFAULT_MAX_ALLOC (1ULL * 1024 * 1024 * 1024)
 /* Server policy ceiling for a client-provided allocation limit. */
 #define MAX_SERVER_ALLOC (256ULL * 1024 * 1024)
+/* Bounded cumulative per-connection receive budget.  In-flight wire buffers,
+   decompression buffers and queued (not yet written) file payloads for a
+   connection must stay within this ceiling. */
+#define MAX_CONNECTION_MEMORY (256ULL * 1024 * 1024)
 
 typedef struct ssl_st SSL;
 

src/shared/utils.c

@@ -1,6 +1,5 @@
 #include "utils.h"
 #include "array_list.h"
-#include "libgen.h"
 #include "log.h"
 #include <dirent.h>
 #include <errno.h>
@@ -83,45 +82,6 @@ static int open_authorized_destination(const char* dest_root) {
   return dirfd;
 }
 
-bool mkdir_r(const char* path) {
-  if (!path || *path == '\0')
-    return false;
-  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;
-  }
-  bool ok = true;
-  char* saveptr = NULL;
-  char* component = strtok_r(duplicate, "/", &saveptr);
-  while (component) {
-    if (strcmp(component, "..") == 0) {
-      ok = false;
-      break;
-    }
-    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;
-    }
-    component = strtok_r(NULL, "/", &saveptr);
-  }
-  close(dirfd);
-  free(duplicate);
-  return ok;
-}
-
 char* str_dup(const char* string) {
   if (string == NULL)
     return NULL;

src/shared/utils.h

@@ -5,7 +5,6 @@
 #include <stddef.h>
 #include <stdbool.h>
 
-bool mkdir_r(const char* path);
 char* str_dup(const char* string);
 char* output_escape(const char* string, bool eight_bit_output);
 char* path_cat(const char* path1, const char* path2);

tests/integration/test_features.py

@@ -1,4 +1,5 @@
 """Feature tests: incremental sync, bandwidth limiting, dry run, metadata, filters."""
+import filecmp
 import os
 import shutil
 import sys
@@ -813,3 +814,210 @@ class TestBandwidthLimit:
         mismatches, missing = verify_transfer(SOURCE_DIR, received)
         assert not missing, f"Missing: {missing}"
         assert not mismatches, f"Mismatch: {mismatches}"
+
+
+def _read_file(path):
+    with open(path, "rb") as fh:
+        return fh.read()
+
+
+class TestRemoveSourceFilesSkips:
+    """--remove-source-files must not delete sources the receiver skipped
+    (rsync reference behavior)."""
+
+    def test_existing_first_sync_keeps_new_source(self, shared_server):
+        source = os.path.join(TEST_DATA_DIR, "remove_rsf_existing_src")
+        dest = os.path.join(TEST_DATA_DIR, "remove_rsf_existing_dst")
+        clean_dir(source)
+        clean_dir(dest)
+        with open(os.path.join(source, "only.txt"), "wb") as f:
+            f.write(b"keep me")
+
+        result, _ = run_client(source, dest, flags=["--remove-source-files", "--existing"],
+                               port=shared_server.port)
+        assert result.returncode == 0, f"Sync failed: {result.stderr[:200]}"
+        # The file exists only on the source side, so --existing makes the
+        # receiver skip it; the source must therefore not be removed.
+        assert os.path.isfile(os.path.join(source, "only.txt"))
+        received = get_dest_received_dir(dest, source)
+        assert not os.path.exists(os.path.join(received, "only.txt"))
+
+    def test_ignore_existing_keeps_skipped_source(self, shared_server):
+        source = os.path.join(TEST_DATA_DIR, "remove_rsf_ignore_src")
+        dest = os.path.join(TEST_DATA_DIR, "remove_rsf_ignore_dst")
+        clean_dir(source)
+        clean_dir(dest)
+        source_file = os.path.join(source, "file.txt")
+        with open(source_file, "wb") as f:
+            f.write(b"payload")
+
+        result, _ = run_client(source, dest, port=shared_server.port)
+        assert result.returncode == 0
+
+        result, _ = run_client(source, dest, flags=["--remove-source-files", "--ignore-existing"],
+                               port=shared_server.port)
+        assert result.returncode == 0, f"Sync failed: {result.stderr[:200]}"
+        # Destination already has the file, so the second run is a receiver
+        # skip; the source file must survive.
+        assert os.path.isfile(source_file)
+        received = get_dest_received_dir(dest, source)
+        assert _read_file(os.path.join(received, "file.txt")) == b"payload"
+
+    def test_update_newer_destination_keeps_source(self, shared_server):
+        source = os.path.join(TEST_DATA_DIR, "remove_rsf_update_src")
+        dest = os.path.join(TEST_DATA_DIR, "remove_rsf_update_dst")
+        clean_dir(source)
+        clean_dir(dest)
+        source_file = os.path.join(source, "file.txt")
+        with open(source_file, "wb") as f:
+            f.write(b"source payload")
+
+        result, _ = run_client(source, dest, port=shared_server.port)
+        assert result.returncode == 0
+
+        received = get_dest_received_dir(dest, source)
+        received_file = os.path.join(received, "file.txt")
+        with open(received_file, "wb") as f:
+            f.write(b"newer destination payload")
+        os.utime(received_file, ns=(time.time_ns() + 10**9, time.time_ns() + 10**9))
+
+        result, _ = run_client(source, dest, flags=["--remove-source-files", "--update"],
+                               port=shared_server.port)
+        assert result.returncode == 0, f"Sync failed: {result.stderr[:200]}"
+        # --update skips a destination that is newer than the source, so the
+        # source must not be removed.
+        assert os.path.isfile(source_file)
+        assert _read_file(received_file) == b"newer destination payload"
+
+    def test_multithreaded_ignore_existing_keeps_skipped_source(self, shared_server):
+        """The multithreaded writer path must also report per-file outcomes so a
+        --remove-source-files sender does not delete skipped sources."""
+        source = os.path.join(TEST_DATA_DIR, "remove_rsf_mt_ignore_src")
+        dest = os.path.join(TEST_DATA_DIR, "remove_rsf_mt_ignore_dst")
+        clean_dir(source)
+        clean_dir(dest)
+        source_file = os.path.join(source, "file.txt")
+        with open(source_file, "wb") as f:
+            f.write(b"payload")
+
+        result, _ = run_client(source, dest, port=shared_server.port)
+        assert result.returncode == 0
+
+        result, _ = run_client(source, dest,
+                               flags=["--remove-source-files", "--ignore-existing", "-m"],
+                               port=shared_server.port)
+        assert result.returncode == 0, f"Sync failed: {result.stderr[:200]}"
+        # Destination already has the file, so the receiver (writer thread)
+        # skips it; the source must survive.
+        assert os.path.isfile(source_file)
+        received = get_dest_received_dir(dest, source)
+        assert _read_file(os.path.join(received, "file.txt")) == b"payload"
+
+
+class TestBackup:
+    def _sync(self, source, dest, flags, port):
+        return run_client(source, dest, flags=flags, port=port)
+
+    def test_plain_backup_keeps_previous_version(self, shared_server):
+        source = os.path.join(TEST_DATA_DIR, "backup_src")
+        dest = os.path.join(TEST_DATA_DIR, "backup_dst")
+        clean_dir(source)
+        clean_dir(dest)
+        source_file = os.path.join(source, "f.txt")
+        with open(source_file, "wb") as f:
+            f.write(b"AAAA")
+
+        result, _ = self._sync(source, dest, ["--backup"], shared_server.port)
+        assert result.returncode == 0, f"Backup sync failed: {result.stderr[:200]}"
+
+        with open(source_file, "wb") as f:
+            f.write(b"BBBB")
+        result, _ = self._sync(source, dest, ["--backup"], shared_server.port)
+        assert result.returncode == 0, f"Backup sync failed: {result.stderr[:200]}"
+
+        received = get_dest_received_dir(dest, source)
+        assert _read_file(os.path.join(received, "f.txt")) == b"BBBB"
+        # rsync default suffix "~" keeps the overwritten version.
+        assert _read_file(os.path.join(received, "f.txt~")) == b"AAAA"
+
+    def test_backup_custom_suffix(self, shared_server):
+        source = os.path.join(TEST_DATA_DIR, "backup_suffix_src")
+        dest = os.path.join(TEST_DATA_DIR, "backup_suffix_dst")
+        clean_dir(source)
+        clean_dir(dest)
+        source_file = os.path.join(source, "f.txt")
+        with open(source_file, "wb") as f:
+            f.write(b"AAAA")
+
+        flags = ["--backup", "--suffix", ".bak"]
+        result, _ = self._sync(source, dest, flags, shared_server.port)
+        assert result.returncode == 0, f"Backup sync failed: {result.stderr[:200]}"
+        with open(source_file, "wb") as f:
+            f.write(b"BBBB")
+        result, _ = self._sync(source, dest, flags, shared_server.port)
+        assert result.returncode == 0, f"Backup sync failed: {result.stderr[:200]}"
+
+        received = get_dest_received_dir(dest, source)
+        assert _read_file(os.path.join(received, "f.txt")) == b"BBBB"
+        assert _read_file(os.path.join(received, "f.txt.bak")) == b"AAAA"
+
+    def test_backup_dir_stores_backups_separately(self, shared_server):
+        source = os.path.join(TEST_DATA_DIR, "backup_dir_src")
+        dest = os.path.join(TEST_DATA_DIR, "backup_dir_dst")
+        clean_dir(source)
+        clean_dir(dest)
+        source_file = os.path.join(source, "f.txt")
+        with open(source_file, "wb") as f:
+            f.write(b"AAAA")
+
+        flags = ["--backup", "--backup-dir", "backups"]
+        result, _ = self._sync(source, dest, flags, shared_server.port)
+        assert result.returncode == 0, f"Backup sync failed: {result.stderr[:200]}"
+        with open(source_file, "wb") as f:
+            f.write(b"BBBB")
+        result, _ = self._sync(source, dest, flags, shared_server.port)
+        assert result.returncode == 0, f"Backup sync failed: {result.stderr[:200]}"
+
+        received = get_dest_received_dir(dest, source)
+        assert _read_file(os.path.join(received, "f.txt")) == b"BBBB"
+        backup = os.path.join(dest, "backups", os.path.relpath(source_file, os.path.sep))
+        assert _read_file(backup) == b"AAAA"
+
+
+class TestPartialDir:
+    def test_completed_transfer_installed_in_destination(self, shared_server):
+        source = os.path.join(TEST_DATA_DIR, "partial_src")
+        dest = os.path.join(TEST_DATA_DIR, "partial_dst")
+        clean_dir(source)
+        clean_dir(dest)
+        source_file = os.path.join(source, "f.txt")
+        with open(source_file, "wb") as f:
+            f.write(b"partial payload")
+
+        result, _ = run_client(source, dest, flags=["--partial", "--partial-dir", ".partial"],
+                               port=shared_server.port)
+        assert result.returncode == 0, f"Partial sync failed: {result.stderr[:200]}"
+
+        received = get_dest_received_dir(dest, source)
+        assert _read_file(os.path.join(received, "f.txt")) == b"partial payload"
+        # A completed transfer must not remain under the partial directory.
+        partial = os.path.join(dest, ".partial", os.path.relpath(source_file, os.path.sep))
+        assert not os.path.exists(partial)
+
+
+class TestLargeFile:
+    def test_transfer_100mb_file(self, shared_server):
+        source = os.path.join(TEST_DATA_DIR, "large_src")
+        dest = os.path.join(TEST_DATA_DIR, "large_dst")
+        clean_dir(source)
+        clean_dir(dest)
+        source_file = os.path.join(source, "big.bin")
+        chunk = os.urandom(1024 * 1024)
+        with open(source_file, "wb") as f:
+            for _ in range(100):
+                f.write(chunk)
+
+        result, _ = run_client(source, dest, port=shared_server.port)
+        assert result.returncode == 0, f"Large-file sync failed: {result.stderr[:200]}"
+        received = get_dest_received_dir(dest, source)
+        assert filecmp.cmp(source_file, os.path.join(received, "big.bin"), shallow=False)

tests/test_client_cli.c

@@ -1088,6 +1088,70 @@ static void test_parse_args_rejects_invalid_compression_choice() {
   config_delete(cfg);
 }
 
+/* Every value-taking table option accepts an inline "--opt=value" form. */
+static void test_parse_args_table_equals_size_options() {
+  Config* cfg = config_create();
+  char* argv[] = {"fastsync", "--max-size=2G", "--min-size=1K", "/src", "/dst"};
+  int positional_args[2];
+  int positional_count = 0;
+
+  EXPECT_EQ_INT(parse_args(cfg, 5, argv, positional_args, &positional_count), 0);
+  EXPECT_TRUE(cfg->max_size == 2ULL * 1024 * 1024 * 1024);
+  EXPECT_TRUE(cfg->min_size == 1024ULL);
+  EXPECT_EQ_INT(positional_count, 2);
+  config_delete(cfg);
+}
+
+static void test_parse_args_table_equals_string_and_int_options() {
+  Config* cfg = config_create();
+  char* argv[] = {"fastsync", "--suffix=.bak", "--timeout=30", "--max-depth=5", "/src", "/dst"};
+  int positional_args[2];
+  int positional_count = 0;
+
+  EXPECT_EQ_INT(parse_args(cfg, 6, argv, positional_args, &positional_count), 0);
+  EXPECT_EQ_STR(cfg->suffix, ".bak");
+  EXPECT_EQ_INT(cfg->timeout, 30);
+  EXPECT_EQ_INT(cfg->max_depth, 5);
+  EXPECT_EQ_INT(positional_count, 2);
+  config_delete(cfg);
+
+  cfg = config_create();
+  char* backup_argv[] = {"fastsync", "--backup-dir=/tmp/bak", "/src", "/dst"};
+  positional_count = 0;
+  EXPECT_EQ_INT(parse_args(cfg, 4, backup_argv, positional_args, &positional_count), 0);
+  EXPECT_EQ_STR(cfg->backup_dir, "/tmp/bak");
+  config_delete(cfg);
+}
+
+/* Options that take a separate value must report "missing argument", not the
+ * generic "Unknown option", when they are the final argv entry. */
+static void test_parse_args_missing_argument_diagnostic() {
+  static const char* const options[] = {"--exclude", "--server-port", "--skip-compress", "-T"};
+
+  for (size_t i = 0; i < sizeof(options) / sizeof(options[0]); i++) {
+    Config* cfg = config_create();
+    char* argv[] = {"fastsync", (char*)options[i]};
+    int positional_args[2];
+    int positional_count = 0;
+    FILE* log_file = tmpfile();
+    char log_buffer[512] = {0};
+
+    EXPECT_NOT_NULL(log_file);
+    log_set_file(log_file);
+
+    EXPECT_EQ_INT(parse_args(cfg, 2, argv, positional_args, &positional_count), -1);
+    fflush(log_file);
+    rewind(log_file);
+    EXPECT_TRUE(fread(log_buffer, 1, sizeof(log_buffer) - 1, log_file) > 0);
+    EXPECT_TRUE(strstr(log_buffer, "missing argument") != NULL);
+    EXPECT_TRUE(strstr(log_buffer, "Unknown option") == NULL);
+
+    log_set_file(NULL);
+    fclose(log_file);
+    config_delete(cfg);
+  }
+}
+
 void test_client_cli() {
   test_validate_config_required_paths();
   test_validate_config_incompatible_options();
@@ -1155,6 +1219,9 @@ void test_client_cli() {
   test_parse_args_compression_alias_equals();
   test_parse_args_rejects_invalid_compression_level_equals();
   test_parse_args_rejects_invalid_compression_choice();
+  test_parse_args_table_equals_size_options();
+  test_parse_args_table_equals_string_and_int_options();
+  test_parse_args_missing_argument_diagnostic();
   test_parse_args_partial_progress();
   test_parse_args_checksum_choice_aliases();
   test_parse_args_checksum_choice_requires_value();

tests/test_config.c

@@ -282,6 +282,105 @@ static void test_config_receive_truncated() {
   close(p[1]);
 }
 
+static bool config_string_roundtrip_matches(const Config* send_cfg, Config* recv) {
+  /* The sender serializes NULL strings as "" on the wire.  Receivers must
+     canonicalize those empty values back to NULL for the options whose client
+     default is NULL (backup_dir, temp_dir, partial_dir, suffix), while a real
+     non-empty value round-trips unchanged. */
+  const char* fields[4];
+  char* const* recv_fields[4];
+  fields[0] = send_cfg->backup_dir;
+  recv_fields[0] = &recv->backup_dir;
+  fields[1] = send_cfg->temp_dir;
+  recv_fields[1] = &recv->temp_dir;
+  fields[2] = send_cfg->partial_dir;
+  recv_fields[2] = &recv->partial_dir;
+  fields[3] = send_cfg->suffix;
+  recv_fields[3] = &recv->suffix;
+  for (int i = 0; i < 4; i++) {
+    const char* sent = fields[i];
+    const char* got = *recv_fields[i];
+    if (sent == NULL || sent[0] == '\0') {
+      if (got != NULL)
+        return false;
+    } else if (got == NULL || strcmp(sent, got) != 0) {
+      return false;
+    }
+  }
+  return true;
+}
+
+static bool roundtrip_config_ok(const Config* send_cfg) {
+  int p[2];
+  if (socketpair(AF_UNIX, SOCK_STREAM, 0, p) != 0)
+    return false;
+  pid_t pid = fork();
+  if (pid == 0) {
+    close(p[1]);
+    io_set_fds(p[0], p[0]);
+    Config* recv = config_receive(p[0]);
+    bool ok = recv != NULL;
+    if (ok) {
+      ok = recv->version != NULL && strcmp(recv->version, PROTOCOL_VERSION) == 0;
+      ok = ok && recv->send_directory && recv->receive_root_directory;
+      ok = ok && config_string_roundtrip_matches(send_cfg, recv);
+    }
+    config_delete(recv);
+    close(p[0]);
+    _exit(ok ? 0 : 1);
+  } else {
+    close(p[0]);
+    io_set_fds(p[1], p[1]);
+    bool sent = config_send(p[1], send_cfg);
+    int status;
+    waitpid(pid, &status, 0);
+    close(p[1]);
+    return sent && WIFEXITED(status) && WEXITSTATUS(status) == 0;
+  }
+}
+
+/* Issue #252: NULL-vs-empty must survive the wire for backup_dir, temp_dir,
+   partial_dir, and suffix.  NULL and explicitly-empty client values are both
+   serialized as "" and must be reconstructed as NULL so plain --backup (with
+   no --suffix/--backup-dir) works exactly like the client configured it. */
+static void test_config_string_null_vs_empty_roundtrip() {
+  if (is_running_under_valgrind())
+    return;
+
+  /* NULL values on the wire must come back as NULL. */
+  Config* a = config_create();
+  EXPECT_NOT_NULL(a);
+  a->send_directory = str_dup("/src");
+  a->receive_root_directory = str_dup("/dst");
+  EXPECT_TRUE(roundtrip_config_ok(a));
+  config_delete(a);
+
+  /* Explicitly empty strings (indistinguishable on the wire from NULL) must
+     be canonicalized to NULL by the receiver. */
+  Config* b = config_create();
+  EXPECT_NOT_NULL(b);
+  b->send_directory = str_dup("/src");
+  b->receive_root_directory = str_dup("/dst");
+  b->backup_dir = str_dup("");
+  b->temp_dir = str_dup("");
+  b->partial_dir = str_dup("");
+  b->suffix = str_dup("");
+  EXPECT_TRUE(roundtrip_config_ok(b));
+  config_delete(b);
+
+  /* Non-empty values must round-trip unchanged. */
+  Config* c = config_create();
+  EXPECT_NOT_NULL(c);
+  c->send_directory = str_dup("/src");
+  c->receive_root_directory = str_dup("/dst");
+  c->backup_dir = str_dup("backups");
+  c->temp_dir = str_dup("/tmp/fast");
+  c->partial_dir = str_dup(".partial");
+  c->suffix = str_dup(".bak");
+  EXPECT_TRUE(roundtrip_config_ok(c));
+  config_delete(c);
+}
+
 static void test_config_is_remote_dest() {
   /* Valid SSH-style destinations */
   EXPECT_TRUE(config_is_remote_dest("user@host:/path"));
@@ -315,6 +414,7 @@ void test_config() {
     test_config_send_receive();
     test_config_send_receive_version_mismatch();
     test_config_receive_truncated();
+    test_config_string_null_vs_empty_roundtrip();
   }
   test_config_is_remote_dest();
 }

tests/test_delta.c

@@ -343,6 +343,297 @@ static void test_delta_apply_rejects_output_overflow() {
   EXPECT_TRUE(delta_apply(old_data, sizeof(old_data), &delta, 1) == NULL);
 }
 
+/* ---------------------------------------------------------------------------
+ * Hash-index lookup differential tests.
+ *
+ * delta_compute buckets signature blocks by their weak checksum.  These tests
+ * prove the bucket-indexed candidate lookup is behaviour-identical to the
+ * original per-window linear scan: the emitted instruction stream (types,
+ * lengths, literal bytes and chosen block indices) must match a naive linear
+ * reference exactly, and the delta must reconstruct the new buffer.
+ * ------------------------------------------------------------------------- */
+
+#define REF_NO_MATCH UINT32_MAX
+
+typedef struct {
+  DeltaInstruction* items;
+  uint32_t count;
+  uint32_t cap;
+} RefDelta;
+
+static void ref_delta_free(RefDelta* ref) {
+  if (!ref->items)
+    return;
+  for (uint32_t i = 0; i < ref->count; i++)
+    if (ref->items[i].type == DELTA_INSTR_LITERAL)
+      free(ref->items[i].literal.data);
+  free(ref->items);
+  ref->items = NULL;
+  ref->count = 0;
+  ref->cap = 0;
+}
+
+static bool ref_delta_push(RefDelta* ref, DeltaInstruction instr) {
+  if (ref->count == ref->cap) {
+    uint32_t new_cap = ref->cap ? ref->cap * 2 : 16;
+    DeltaInstruction* tmp = realloc(ref->items, (size_t)new_cap * sizeof(DeltaInstruction));
+    if (!tmp)
+      return false;
+    ref->items = tmp;
+    ref->cap = new_cap;
+  }
+  ref->items[ref->count++] = instr;
+  return true;
+}
+
+static bool ref_delta_flush_literal(RefDelta* ref, const uint8_t* data, uint64_t start,
+                                    uint64_t end) {
+  if (start >= end)
+    return true;
+  uint8_t* lit = malloc((size_t)(end - start));
+  if (!lit)
+    return false;
+  memcpy(lit, data + start, (size_t)(end - start));
+  DeltaInstruction instr = {
+      .type = DELTA_INSTR_LITERAL,
+      .literal = {.data = lit, .length = (uint32_t)(end - start)},
+  };
+  return ref_delta_push(ref, instr);
+}
+
+/* Naive O(windows x blocks) re-implementation of the historical delta_compute
+ * candidate scan: only full windows may match, a candidate needs both the weak
+ * (Adler-32) and strong (xxHash32) checksums to agree, and the lowest matching
+ * block index is selected. */
+static bool ref_delta_build(RefDelta* ref, const uint8_t* new_data, uint64_t new_size,
+                            const DeltaSignature* sig) {
+  uint32_t block_size = sig->block_size;
+  uint64_t i = 0;
+  uint64_t literal_start = 0;
+  bool has_literal = false;
+
+  while (i < new_size) {
+    uint32_t window_len = (uint32_t)((new_size - i < block_size) ? (new_size - i) : block_size);
+    bool full_window = (window_len == block_size);
+
+    uint32_t matched = REF_NO_MATCH;
+    if (full_window) {
+      uint32_t adler = delta_adler32(new_data + i, window_len);
+      for (uint32_t j = 0; j < sig->block_count; j++) {
+        if (sig->blocks[j].adler32 == adler &&
+            delta_xxhash32(new_data + i, window_len) == sig->blocks[j].xxhash) {
+          matched = j;
+          break;
+        }
+      }
+    }
+
+    if (matched != REF_NO_MATCH) {
+      if (has_literal) {
+        if (!ref_delta_flush_literal(ref, new_data, literal_start, i))
+          return false;
+        has_literal = false;
+      }
+      DeltaInstruction instr = {
+          .type = DELTA_INSTR_BLOCK_MATCH,
+          .match = {.block_index = matched, .block_offset = 0, .length = window_len},
+      };
+      if (!ref_delta_push(ref, instr))
+        return false;
+      i += window_len;
+    } else {
+      if (!has_literal) {
+        literal_start = i;
+        has_literal = true;
+      }
+      i++;
+    }
+  }
+
+  if (has_literal && !ref_delta_flush_literal(ref, new_data, literal_start, new_size))
+    return false;
+  return true;
+}
+
+static bool ref_delta_matches(const RefDelta* ref, const Delta* delta) {
+  if (ref->count != delta->instruction_count)
+    return false;
+  for (uint32_t i = 0; i < ref->count; i++) {
+    const DeltaInstruction* a = &ref->items[i];
+    const DeltaInstruction* b = &delta->instructions[i];
+    if (a->type != b->type)
+      return false;
+    if (a->type == DELTA_INSTR_BLOCK_MATCH) {
+      if (a->match.block_index != b->match.block_index ||
+          a->match.block_offset != b->match.block_offset || a->match.length != b->match.length)
+        return false;
+    } else {
+      if (a->literal.length != b->literal.length ||
+          memcmp(a->literal.data, b->literal.data, a->literal.length) != 0)
+        return false;
+    }
+  }
+  return true;
+}
+
+static void expect_linear_reference_match(const uint8_t* old_data, uint64_t old_size,
+                                          const uint8_t* new_data, uint64_t new_size,
+                                          uint32_t block_size, const char* label) {
+  DeltaSignature* sig = delta_signature_create(old_data, old_size, block_size);
+  if (!sig) {
+    printf("      [FAIL] %s: signature creation failed\n", label);
+    EXPECT_NOT_NULL(sig);
+    return;
+  }
+  Delta* delta = delta_compute(new_data, new_size, sig, block_size);
+  if (!delta) {
+    printf("      [FAIL] %s: delta_compute returned NULL\n", label);
+    delta_signature_destroy(sig);
+    EXPECT_NOT_NULL(delta);
+    return;
+  }
+  RefDelta ref = {0};
+  bool ok = ref_delta_build(&ref, new_data, new_size, sig);
+  if (ok)
+    ok = ref_delta_matches(&ref, delta);
+  if (!ok) {
+    printf("      [FAIL] %s: instruction stream differs from linear reference "
+           "(linear=%u indexed=%u)\n",
+           label, ref.count, delta->instruction_count);
+  }
+  ref_delta_free(&ref);
+  delta_destroy(delta);
+  delta_signature_destroy(sig);
+  EXPECT_TRUE(ok);
+}
+
+static void fill_delta_pattern(uint8_t* buf, uint64_t size, uint32_t seed) {
+  uint32_t x = seed ? seed : 1;
+  for (uint64_t i = 0; i < size; i++) {
+    x ^= x << 13;
+    x ^= x >> 17;
+    x ^= x << 5;
+    buf[i] = (uint8_t)(x >> 24);
+  }
+}
+
+static void test_delta_hash_index_matches_linear_reference() {
+  /* Identical file (full block alignment). */
+  uint8_t old_a[32768];
+  uint8_t new_a[32768];
+  fill_delta_pattern(old_a, sizeof(old_a), 42);
+  memcpy(new_a, old_a, sizeof(old_a));
+  expect_linear_reference_match(old_a, sizeof(old_a), new_a, sizeof(new_a), 2048,
+                                "identical 32KiB @ 2KiB");
+
+  /* Scattered single-byte edits in the middle of each block. */
+  uint8_t new_b[32768];
+  memcpy(new_b, old_a, sizeof(old_a));
+  for (size_t p = 100; p < sizeof(new_b); p += 4096)
+    new_b[p] ^= 0x5A;
+  expect_linear_reference_match(old_a, sizeof(old_a), new_b, sizeof(new_b), 2048,
+                                "32KiB scattered single-byte edits @ 2KiB");
+
+  /* Non-aligned old file (partial final block) with a single edit. */
+  uint8_t old_c[30000];
+  uint8_t new_c[30000];
+  fill_delta_pattern(old_c, sizeof(old_c), 7);
+  memcpy(new_c, old_c, sizeof(old_c));
+  new_c[15000] ^= 0x3C;
+  expect_linear_reference_match(old_c, sizeof(old_c), new_c, sizeof(new_c), 2048,
+                                "30KiB partial-tail single edit @ 2KiB");
+
+  /* Growth: appended data after an identical prefix. */
+  uint8_t old_d[24576];
+  uint8_t new_d[34576];
+  fill_delta_pattern(old_d, sizeof(old_d), 11);
+  memcpy(new_d, old_d, sizeof(old_d));
+  fill_delta_pattern(new_d + sizeof(old_d), sizeof(new_d) - sizeof(old_d), 23);
+  expect_linear_reference_match(old_d, sizeof(old_d), new_d, sizeof(new_d), 2048,
+                                "24KiB -> 34KiB appended @ 2KiB");
+
+  /* Insertion shifting everything after the edit point (rsync re-sync). */
+  uint8_t old_e[65536];
+  uint8_t new_e[65536 + 3000];
+  fill_delta_pattern(old_e, sizeof(old_e), 99);
+  memcpy(new_e, old_e, 20000);
+  fill_delta_pattern(new_e + 20000, 3000, 101);
+  memcpy(new_e + 23000, old_e + 20000, sizeof(old_e) - 20000);
+  expect_linear_reference_match(old_e, sizeof(old_e), new_e, sizeof(new_e), 2048,
+                                "64KiB + 3KiB insertion @ 2KiB");
+
+  /* Deletion shrinking the file. */
+  uint8_t new_f[sizeof(old_e) - 5000];
+  memcpy(new_f, old_e, 30000);
+  memcpy(new_f + 30000, old_e + 35000, sizeof(old_e) - 35000);
+  expect_linear_reference_match(old_e, sizeof(old_e), new_f, sizeof(new_f), 2048,
+                                "64KiB - 5KiB deletion @ 2KiB");
+
+  /* Repeated identical blocks must resolve to the lowest block index. */
+  uint8_t old_g[4 * 4096];
+  uint8_t new_g[4 * 4096];
+  for (uint32_t b = 0; b < 4; b++)
+    fill_delta_pattern(old_g + b * 4096, 4096, b % 2 == 0 ? 500 : 501); /* block0==block2 */
+  memcpy(new_g, old_g, sizeof(old_g));
+  new_g[4096 + 5] ^= 0x11; /* edit inside the second (duplicated) chunk */
+  expect_linear_reference_match(old_g, sizeof(old_g), new_g, sizeof(new_g), 4096,
+                                "duplicated chunks @ 4KiB");
+
+  /* Block larger than the file: nothing can match, all literal. */
+  uint8_t old_h[1000];
+  uint8_t new_h[1000];
+  fill_delta_pattern(old_h, sizeof(old_h), 3);
+  memcpy(new_h, old_h, sizeof(old_h));
+  expect_linear_reference_match(old_h, sizeof(old_h), new_h, sizeof(new_h), 4096,
+                                "1KiB file @ 4KiB block");
+}
+
+static void test_delta_hash_index_large_mostly_matching() {
+  const uint64_t size = 4ULL * 1024 * 1024;
+  const uint32_t block_size = 8192;
+
+  uint8_t* old_data = malloc((size_t)size);
+  uint8_t* new_data = malloc((size_t)size);
+  EXPECT_TRUE(old_data != NULL && new_data != NULL);
+
+  fill_delta_pattern(old_data, size, 1234);
+  memcpy(new_data, old_data, (size_t)size);
+
+  /* Scattered single-byte changes across the whole buffer.  Each change forces
+   * the diff to re-synchronise by walking one byte at a time through the
+   * affected block, which is exactly the case that used to cost O(bytes x
+   * blocks) with the linear scan. */
+  const uint64_t nchanges = 64;
+  for (uint64_t c = 0; c < nchanges; c++) {
+    uint64_t pos = (c * (size / nchanges)) + (c % 17);
+    new_data[pos] ^= (uint8_t)(0xA0 + (c % 16));
+  }
+
+  DeltaSignature* sig = delta_signature_create(old_data, size, block_size);
+  EXPECT_NOT_NULL(sig);
+  EXPECT_EQ_INT((int)sig->block_count, (int)(size / block_size));
+
+  Delta* delta = delta_compute(new_data, size, sig, block_size);
+  EXPECT_NOT_NULL(delta);
+  EXPECT_EQ_INT((int)delta->new_file_size, (int)size);
+
+  uint32_t match_count = 0;
+  for (uint32_t i = 0; i < delta->instruction_count; i++)
+    if (delta->instructions[i].type == DELTA_INSTR_BLOCK_MATCH)
+      match_count++;
+  EXPECT_TRUE(match_count > 0);
+
+  void* result = delta_apply(old_data, size, delta, block_size);
+  EXPECT_NOT_NULL(result);
+  EXPECT_EQ_INT(memcmp(result, new_data, (size_t)size), 0);
+
+  free(result);
+  delta_destroy(delta);
+  delta_signature_destroy(sig);
+  free(old_data);
+  free(new_data);
+}
+
 void test_delta() {
   test_adler32_basic();
   test_adler32_different_data();
@@ -360,4 +651,6 @@ void test_delta() {
   test_is_worthwhile();
   test_large_file_delta();
   test_delta_apply_rejects_output_overflow();
+  test_delta_hash_index_matches_linear_reference();
+  test_delta_hash_index_large_mostly_matching();
 }

tests/test_file.c

@@ -5,10 +5,12 @@
 #include "utils.h"
 #include "protocol.h"
 #include "test_utils.h"
+#include <fcntl.h>
 #include <stdlib.h>
 #include <string.h>
 #include <sys/stat.h>
 #include <sys/wait.h>
+#include <time.h>
 #include <unistd.h>
 
 static void test_file_create() {
@@ -250,6 +252,111 @@ static void test_file_save_to_disk_ignore_existing_entry_types() {
   rmdir(root);
 }
 
+/* Issue #253: with --partial --partial-dir a completed write must be installed
+   at the real destination rather than left under the partial directory. */
+static void test_file_save_to_disk_partial_install() {
+  const char* root = "test_partial_install_tmp";
+  const char* dest_file = "test_partial_install_tmp/file.txt";
+  const char* partial_file = "test_partial_install_tmp/.partial/file.txt";
+  unlink(dest_file);
+  unlink(partial_file);
+  rmdir("test_partial_install_tmp/.partial");
+  rmdir(root);
+
+  File* f = file_create("file.txt");
+  EXPECT_NOT_NULL(f);
+  const char* content = "partial-dir content";
+  f->data->data = malloc(strlen(content));
+  EXPECT_NOT_NULL(f->data->data);
+  memcpy(f->data->data, content, strlen(content));
+  f->data->size = strlen(content);
+
+  Config* config = config_create();
+  EXPECT_NOT_NULL(config);
+  config->partial = true;
+  config->partial_dir = str_dup(".partial");
+
+  EXPECT_EQ_INT(file_save_to_disk_full(root, f, config), FILE_SAVE_WRITTEN);
+
+  FILE* fp = fopen(dest_file, "rb");
+  EXPECT_NOT_NULL(fp);
+  // cppcheck-suppress knownConditionTrueFalse
+  if (fp) {
+    char buf[64] = {0};
+    size_t nread = fread(buf, 1, sizeof(buf) - 1, fp);
+    fclose(fp);
+    EXPECT_EQ_INT((int)nread, (int)strlen(content));
+    EXPECT_EQ_INT(memcmp(buf, content, strlen(content)), 0);
+  }
+  /* A completed transfer must not linger under the partial dir. */
+  EXPECT_EQ_INT(access(partial_file, F_OK), -1);
+
+  file_destroy(f);
+  config_delete(config);
+  unlink(dest_file);
+  rmdir(root);
+}
+
+/* Issue #251: file_save_to_disk_full must distinguish receiver-side skips
+   (--existing/--ignore-existing/--update) from real writes so the sender can
+   decide whether --remove-source-files may unlink its source. */
+static void test_file_save_to_disk_reports_skips() {
+  const char* root = "test_save_skip_tmp";
+  const char* existing_path = "test_save_skip_tmp/existing.txt";
+  unlink(existing_path);
+  rmdir(root);
+  EXPECT_TRUE(file_write_to_disk(existing_path, "old", 3, false, false));
+
+  Config* cfg = config_create();
+  EXPECT_NOT_NULL(cfg);
+
+  File* new_file = file_create("missing.txt");
+  EXPECT_NOT_NULL(new_file);
+  new_file->data->data = malloc(7);
+  EXPECT_NOT_NULL(new_file->data->data);
+  memcpy(new_file->data->data, "skipped", 7);
+  new_file->data->size = 7;
+
+  /* --existing: destination is missing -> skipped, not an error. */
+  cfg->existing = true;
+  EXPECT_EQ_INT(file_save_to_disk_full(root, new_file, cfg), FILE_SAVE_SKIPPED);
+  cfg->existing = false;
+
+  /* --ignore-existing: destination present -> skipped. */
+  File* present = file_create("existing.txt");
+  EXPECT_NOT_NULL(present);
+  present->data->data = malloc(3);
+  EXPECT_NOT_NULL(present->data->data);
+  memcpy(present->data->data, "new", 3);
+  present->data->size = 3;
+  cfg->ignore_existing = true;
+  EXPECT_EQ_INT(file_save_to_disk_full(root, present, cfg), FILE_SAVE_SKIPPED);
+  cfg->ignore_existing = false;
+
+  /* A normal overwrite of an existing file is a real write. */
+  EXPECT_EQ_INT(file_save_to_disk_full(root, present, cfg), FILE_SAVE_WRITTEN);
+
+  /* --update: a newer destination is skipped. */
+  struct stat st;
+  EXPECT_EQ_INT(stat(existing_path, &st), 0);
+  time_t now = time(NULL);
+  FileMetadata metadata = {.mode = st.st_mode,
+                           .uid = st.st_uid,
+                           .gid = st.st_gid,
+                           .mtime_sec = now - 100,
+                           .mtime_nsec = 0};
+  present->metadata = &metadata;
+  cfg->update = true;
+  EXPECT_EQ_INT(file_save_to_disk_full(root, present, cfg), FILE_SAVE_SKIPPED);
+  present->metadata = NULL;
+
+  file_destroy(new_file);
+  file_destroy(present);
+  config_delete(cfg);
+  unlink(existing_path);
+  rmdir(root);
+}
+
 static void test_file_write_to_disk_basic() {
   const char* content = "Basic file_write_to_disk test";
   EXPECT_TRUE(file_write_to_disk("test_file_write_to_disk_basic.txt", content, strlen(content),
@@ -624,6 +731,161 @@ static void test_file_send_single_calls_metadata_and_path() {
   }
 }
 
+static void test_inplace_overwrite_clears_special_mode_bits() {
+  const char* root = "test_inplace_tmp";
+  const char* path = "test_inplace_tmp/priv.txt";
+  const char* content = "olddata";
+  unlink(path);
+  rmdir(root);
+  EXPECT_EQ_INT(mkdir(root, 0700), 0);
+
+  /* Create a destination carrying setuid + sticky bits. */
+  int fd = open(path, O_WRONLY | O_CREAT | O_TRUNC | O_CLOEXEC, 0644);
+  EXPECT_TRUE(fd >= 0);
+  // cppcheck-suppress knownConditionTrueFalse
+  if (fd < 0) {
+    rmdir(root);
+    return;
+  }
+  EXPECT_EQ_INT((int)write(fd, content, strlen(content)), (int)strlen(content));
+  EXPECT_EQ_INT(fchmod(fd, S_ISUID | S_ISVTX | 0755), 0);
+  EXPECT_EQ_INT(close(fd), 0);
+
+  /* Overwrite in place without metadata: the mode must be normalized to a
+     safe default (0644) and the setuid/sticky bits must be gone. */
+  File* f = file_create("priv.txt");
+  EXPECT_NOT_NULL(f);
+  const char* new_content = "newdata";
+  f->data->data = malloc(strlen(new_content));
+  EXPECT_NOT_NULL(f->data->data);
+  memcpy(f->data->data, new_content, strlen(new_content));
+  f->data->size = strlen(new_content);
+
+  Config* cfg = config_create();
+  EXPECT_NOT_NULL(cfg);
+  cfg->inplace = true;
+  EXPECT_TRUE(file_save_to_disk(root, f, cfg));
+  file_destroy(f);
+  config_delete(cfg);
+
+  struct stat st;
+  EXPECT_EQ_INT(stat(path, &st), 0);
+  EXPECT_EQ_INT((int)(st.st_mode & (S_ISUID | S_ISGID | S_ISVTX)), 0);
+  EXPECT_EQ_INT((int)(st.st_mode & 0777), 0644);
+  FILE* stream = fopen(path, "rb");
+  char buf[16] = {0};
+  EXPECT_NOT_NULL(stream);
+  // cppcheck-suppress knownConditionTrueFalse
+  if (stream) {
+    size_t nread = fread(buf, 1, sizeof(buf) - 1, stream);
+    fclose(stream);
+    EXPECT_EQ_INT((int)nread, (int)strlen(new_content));
+  }
+  EXPECT_EQ_STR(buf, new_content);
+
+  unlink(path);
+  rmdir(root);
+}
+
+static void test_inplace_overwrite_metadata_strips_special_bits() {
+  const char* root = "test_inplace_meta_tmp";
+  const char* path = "test_inplace_meta_tmp/meta.txt";
+  const char* source = "test_inplace_meta_source.txt";
+  unlink(path);
+  unlink(source);
+  rmdir(root);
+  EXPECT_EQ_INT(mkdir(root, 0700), 0);
+
+  /* Existing destination with setuid+sticky set. */
+  int fd = open(path, O_WRONLY | O_CREAT | O_TRUNC | O_CLOEXEC, 0644);
+  EXPECT_TRUE(fd >= 0);
+  // cppcheck-suppress knownConditionTrueFalse
+  if (fd < 0) {
+    rmdir(root);
+    return;
+  }
+  EXPECT_EQ_INT((int)write(fd, "olddata", 7), 7);
+  EXPECT_EQ_INT(fchmod(fd, S_ISUID | S_ISVTX | 0755), 0);
+  EXPECT_EQ_INT(close(fd), 0);
+
+  /* Build source metadata carrying a plain executable mode (no specials). */
+  EXPECT_TRUE(file_write_to_disk(source, "source", 6, false, false));
+  EXPECT_EQ_INT(chmod(source, 0755), 0);
+  struct stat source_st;
+  EXPECT_EQ_INT(stat(source, &source_st), 0);
+
+  File* f = file_create("meta.txt");
+  EXPECT_NOT_NULL(f);
+  const char* new_content = "meta";
+  f->data->data = malloc(strlen(new_content));
+  EXPECT_NOT_NULL(f->data->data);
+  memcpy(f->data->data, new_content, strlen(new_content));
+  f->data->size = strlen(new_content);
+  f->metadata = file_metadata_create(&source_st);
+  EXPECT_NOT_NULL(f->metadata);
+
+  Config* cfg = config_create();
+  EXPECT_NOT_NULL(cfg);
+  cfg->inplace = true;
+  EXPECT_TRUE(file_save_to_disk(root, f, cfg));
+  file_destroy(f);
+  config_delete(cfg);
+  unlink(source);
+
+  struct stat st;
+  EXPECT_EQ_INT(stat(path, &st), 0);
+  /* Metadata-derived mode is applied and never includes setuid/setgid/sticky. */
+  EXPECT_EQ_INT((int)(st.st_mode & (S_ISUID | S_ISGID | S_ISVTX)), 0);
+  EXPECT_EQ_INT((int)(st.st_mode & 0777), 0755);
+
+  unlink(path);
+  rmdir(root);
+}
+
+static void test_inplace_overwrite_truncates_shorter_payload() {
+  const char* root = "test_inplace_trunc_tmp";
+  const char* path = "test_inplace_trunc_tmp/big.txt";
+  unlink(path);
+  rmdir(root);
+  EXPECT_EQ_INT(mkdir(root, 0700), 0);
+
+  const char* old_content = "0123456789abcdef"; /* 16 bytes */
+  EXPECT_TRUE(file_write_to_disk(path, old_content, strlen(old_content), false, false));
+
+  File* f = file_create("big.txt");
+  EXPECT_NOT_NULL(f);
+  const char* new_content = "hi";
+  f->data->data = malloc(strlen(new_content));
+  EXPECT_NOT_NULL(f->data->data);
+  memcpy(f->data->data, new_content, strlen(new_content));
+  f->data->size = strlen(new_content);
+
+  Config* cfg = config_create();
+  EXPECT_NOT_NULL(cfg);
+  cfg->inplace = true;
+  EXPECT_TRUE(file_save_to_disk(root, f, cfg));
+  file_destroy(f);
+  config_delete(cfg);
+
+  /* A shorter payload must truncate the file: no stale trailing bytes. */
+  struct stat st;
+  EXPECT_EQ_INT(stat(path, &st), 0);
+  EXPECT_EQ_INT((int)st.st_size, (int)strlen(new_content));
+  FILE* stream = fopen(path, "rb");
+  char buf[32] = {0};
+  EXPECT_NOT_NULL(stream);
+  // cppcheck-suppress knownConditionTrueFalse
+  if (stream) {
+    size_t nread = fread(buf, 1, sizeof(buf) - 1, stream);
+    fclose(stream);
+    EXPECT_EQ_INT((int)nread, (int)strlen(new_content));
+  }
+  EXPECT_EQ_STR(buf, new_content);
+
+  unlink(path);
+  rmdir(root);
+}
+
 void test_file() {
   test_file_create();
   test_file_destroy_null();
@@ -635,6 +897,8 @@ void test_file() {
   test_file_save_to_disk_existing();
   test_file_save_to_disk_ignore_existing();
   test_file_save_to_disk_ignore_existing_entry_types();
+  test_file_save_to_disk_partial_install();
+  test_file_save_to_disk_reports_skips();
   test_file_write_to_disk_basic();
   test_file_write_to_disk_with_fsync();
   test_file_write_to_disk_creates_dirs();
@@ -654,4 +918,7 @@ void test_file() {
     test_file_send_single_calls_metadata_and_path();
   }
   test_file_metadata_create();
+  test_inplace_overwrite_clears_special_mode_bits();
+  test_inplace_overwrite_metadata_strips_special_bits();
+  test_inplace_overwrite_truncates_shorter_payload();
 }

tests/test_multiprocessing.c

@@ -250,6 +250,87 @@ static void test_write_thread_done() {
   config_delete(cfg);
 }
 
+typedef struct {
+  PipelineContextReceiver* context;
+  File* file;
+  atomic_bool* done;
+  atomic_bool* result;
+} ByteBudgetEnqueueArg;
+
+static int byte_budget_enqueue_worker(void* arg) {
+  ByteBudgetEnqueueArg* worker = arg;
+  bool ok = pipeline_context_receiver_enqueue_file(worker->context, worker->file);
+  atomic_store(worker->result, ok);
+  atomic_store(worker->done, true);
+  return thrd_success;
+}
+
+/* A receiver must not buffer more decompressed/copied payload bytes ahead of
+   the (slow) disk writer than the configured byte budget: an enqueue that
+   would exceed the budget blocks until the writer releases bytes. */
+static void test_receiver_enqueue_byte_budget() {
+  Config* cfg = config_create();
+  EXPECT_NOT_NULL(cfg);
+  free(cfg->version);
+  cfg->version = str_dup(PROTOCOL_VERSION);
+  cfg->send_directory = str_dup("/src");
+  cfg->receive_root_directory = str_dup("/dst");
+  cfg->save_to_disk = false;
+
+  Queue* q = queue_create(16, file_destroy);
+  EXPECT_NOT_NULL(q);
+  PipelineContextReceiver* ctx = pipeline_context_receiver_create(cfg, q, -1, NULL);
+  EXPECT_NOT_NULL(ctx);
+  pipeline_context_receiver_set_queue_byte_limit(ctx, 3000);
+  ctx->receiver_done = false;
+
+  File* first = file_create("budget_file_1");
+  EXPECT_NOT_NULL(first);
+  first->data->size = 2000;
+  EXPECT_TRUE(pipeline_context_receiver_enqueue_file(ctx, first));
+  EXPECT_EQ_INT((int)ctx->queued_bytes, 2000);
+
+  /* Second 2000-byte payload would push the pipeline to 4000 > 3000 budget,
+     so the enqueue must block until the first payload is released. */
+  File* second = file_create("budget_file_2");
+  EXPECT_NOT_NULL(second);
+  second->data->size = 2000;
+  atomic_bool done;
+  atomic_bool result;
+  atomic_init(&done, false);
+  atomic_init(&result, false);
+  ByteBudgetEnqueueArg arg = {ctx, second, &done, &result};
+  thrd_t enqueuer;
+  EXPECT_EQ_INT(thrd_create(&enqueuer, byte_budget_enqueue_worker, &arg), thrd_success);
+
+  /* Give a broken (unbounded) implementation every chance to enqueue. */
+  struct timespec wait = {0, 200 * 1000000L};
+  thrd_sleep(&wait, NULL);
+  EXPECT_FALSE(atomic_load(&done));
+  EXPECT_EQ_INT((int)ctx->queued_bytes, 2000); /* budget still honored */
+
+  /* Simulate the disk writer: dequeue + destroy + release the first file. */
+  File* drained = queue_dequeue_multithreaded(q, &ctx->mutex, &ctx->condition_not_empty,
+                                              &ctx->condition_not_full, &ctx->receiver_done);
+  EXPECT_NOT_NULL(drained);
+  file_destroy(drained);
+  pipeline_context_receiver_note_bytes_released(ctx, 2000);
+  EXPECT_EQ_INT((int)ctx->queued_bytes, 0);
+
+  EXPECT_EQ_INT(thrd_join(enqueuer, NULL), thrd_success);
+  EXPECT_TRUE(atomic_load(&done));
+  EXPECT_TRUE(atomic_load(&result));
+  EXPECT_EQ_INT((int)ctx->queued_bytes, 2000); /* second payload now in flight */
+
+  /* Tear down: the second file is still queued and is freed by queue_destroy. */
+  mtx_destroy(&ctx->mutex);
+  cnd_destroy(&ctx->condition_not_full);
+  cnd_destroy(&ctx->condition_not_empty);
+  free(ctx);
+  queue_destroy(q);
+  config_delete(cfg);
+}
+
 void test_multiprocessing() {
   test_sender_create_destroy();
   test_receiver_create_destroy();
@@ -261,4 +342,5 @@ void test_multiprocessing() {
     test_receive_thread_failure_wakes_writer();
   }
   test_write_thread_done();
+  test_receiver_enqueue_byte_budget();
 }

tests/test_server.c

@@ -1,14 +1,18 @@
 #include "test_server.h"
 #include "config.h"
+#include "delta.h"
 #include "file.h"
 #include "protocol.h"
 #include "test_utils.h"
 #include "utils.h"
+#include <fcntl.h>
 #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>
 
 #include "receiver.h"
@@ -210,6 +214,249 @@ static void test_receive_incremental_check_rejects_invalid_nanoseconds() {
   config_delete(cfg);
 }
 
+static char* make_check_root(const char* tag) {
+  char tmpl[128];
+  snprintf(tmpl, sizeof(tmpl), "/tmp/fastsync_%s_XXXXXX", tag);
+  char* path = str_dup(tmpl);
+  if (!path)
+    return NULL;
+  if (!mkdtemp(path)) {
+    free(path);
+    return NULL;
+  }
+  return path;
+}
+
+static void write_check_file(const char* dir, const char* name, const char* content) {
+  char path[1024];
+  snprintf(path, sizeof(path), "%s/%s", dir, name);
+  int fd = open(path, O_WRONLY | O_CREAT | O_TRUNC, 0644);
+  if (fd >= 0) {
+    size_t len = strlen(content);
+    if (write(fd, content, len) != (ssize_t)len) {
+      /* intentionally ignored in tests */
+    }
+    close(fd);
+  }
+}
+
+/* Issue #255: a same-size/mtime match is decided from metadata alone, so the
+   receiver answers STATUS_OK (skip) and never asks for a data body. */
+static void test_incremental_check_quick_skip_by_mtime() {
+  Config* cfg = config_create();
+  EXPECT_NOT_NULL(cfg);
+  char* root = make_check_root("qskip");
+  EXPECT_NOT_NULL(root);
+  cfg->receive_root_directory = str_dup(root);
+  write_check_file(root, "file.txt", "0123456789abcdef");
+
+  char path[1024];
+  snprintf(path, sizeof(path), "%s/file.txt", root);
+  struct stat st;
+  EXPECT_EQ_INT(stat(path, &st), 0);
+
+  int p[2];
+  EXPECT_EQ_INT(socketpair(AF_UNIX, SOCK_STREAM, 0, p), 0);
+  io_set_fds(p[0], p[1]);
+  io_set_bwlimit(0);
+
+  pid_t pid = fork();
+  if (pid == 0) {
+    alarm(30);
+    close(p[1]);
+    io_set_fds(p[0], p[0]);
+    bool skipped = false;
+    File* file = receive_incremental_check(p[0], cfg, &skipped);
+    bool ok = file == NULL && skipped;
+    file_destroy(file);
+    config_delete(cfg);
+    close(p[0]);
+    _exit(ok ? 0 : 1);
+  } else {
+    close(p[0]);
+    io_set_fds(p[1], p[1]);
+    EXPECT_TRUE(send_str(p[1], "file.txt"));
+    unsigned long long size = (unsigned long long)st.st_size;
+    long long mtime = (long long)st.st_mtime;
+    long long mtime_nsec = 0;
+#ifdef __linux__
+    mtime_nsec = (long long)st.st_mtim.tv_nsec;
+#endif
+    EXPECT_TRUE(send_n_data(p[1], &size, sizeof(size)));
+    EXPECT_TRUE(send_n_data(p[1], &mtime, sizeof(mtime)));
+    EXPECT_TRUE(send_n_data(p[1], &mtime_nsec, sizeof(mtime_nsec)));
+    Status s;
+    EXPECT_TRUE(receive_status(p[1], &s));
+    EXPECT_EQ_INT(s, STATUS_OK);
+
+    int status;
+    waitpid(pid, &status, 0);
+    close(p[1]);
+    config_delete(cfg);
+    unlink(path);
+    rmdir(root);
+    free(root);
+    EXPECT_TRUE(WIFEXITED(status) && WEXITSTATUS(status) == 0);
+  }
+}
+
+/* Issue #255: a size mismatch cannot be a skip, so the receiver answers
+   STATUS_NEXT and consumes the full data body that follows. */
+static void test_incremental_check_size_mismatch_full_transfer() {
+  Config* cfg = config_create();
+  EXPECT_NOT_NULL(cfg);
+  char* root = make_check_root("qnext");
+  EXPECT_NOT_NULL(root);
+  cfg->receive_root_directory = str_dup(root);
+  write_check_file(root, "file.txt", "0123456789abcdef");
+
+  char path[1024];
+  snprintf(path, sizeof(path), "%s/file.txt", root);
+  struct stat st;
+  EXPECT_EQ_INT(stat(path, &st), 0);
+
+  int p[2];
+  EXPECT_EQ_INT(socketpair(AF_UNIX, SOCK_STREAM, 0, p), 0);
+  io_set_fds(p[0], p[1]);
+  io_set_bwlimit(0);
+
+  pid_t pid = fork();
+  if (pid == 0) {
+    alarm(30);
+    close(p[1]);
+    io_set_fds(p[0], p[0]);
+    bool skipped = false;
+    File* file = receive_incremental_check(p[0], cfg, &skipped);
+    bool ok = file != NULL && !skipped && file->path != NULL && strcmp(file->path, "file.txt") == 0;
+    file_destroy(file);
+    config_delete(cfg);
+    close(p[0]);
+    _exit(ok ? 0 : 1);
+  } else {
+    close(p[0]);
+    io_set_fds(p[1], p[1]);
+    EXPECT_TRUE(send_str(p[1], "file.txt"));
+    unsigned long long size = (unsigned long long)st.st_size + 1;
+    long long mtime = (long long)st.st_mtime;
+    long long mtime_nsec = 0;
+#ifdef __linux__
+    mtime_nsec = (long long)st.st_mtim.tv_nsec;
+#endif
+    EXPECT_TRUE(send_n_data(p[1], &size, sizeof(size)));
+    EXPECT_TRUE(send_n_data(p[1], &mtime, sizeof(mtime)));
+    EXPECT_TRUE(send_n_data(p[1], &mtime_nsec, sizeof(mtime_nsec)));
+    Status s;
+    EXPECT_TRUE(receive_status(p[1], &s));
+    EXPECT_EQ_INT(s, STATUS_NEXT);
+
+    Data* body = data_create_reserve(8);
+    EXPECT_NOT_NULL(body);
+    body->data = malloc(8);
+    EXPECT_NOT_NULL(body->data);
+    memcpy(body->data, "replaced", 8);
+    body->size = 8;
+    EXPECT_TRUE(send_data(p[1], body));
+    data_destroy(body);
+
+    int status;
+    waitpid(pid, &status, 0);
+    close(p[1]);
+    config_delete(cfg);
+    unlink(path);
+    rmdir(root);
+    free(root);
+    EXPECT_TRUE(WIFEXITED(status) && WEXITSTATUS(status) == 0);
+  }
+}
+
+/* Issue #256: when a received delta claims a result above the whole-file cap,
+   receive_delta_file must mark the operation failed so the caller aborts with
+   STATUS_ERROR instead of emitting STATUS_NEXT and waiting for a body that
+   never arrives. */
+static void test_incremental_check_delta_oversize_reports_failure() {
+  Config* cfg = config_create();
+  EXPECT_NOT_NULL(cfg);
+  char* root = make_check_root("qdelta");
+  EXPECT_NOT_NULL(root);
+  cfg->receive_root_directory = str_dup(root);
+  cfg->use_delta = true;
+
+  char content[20000];
+  memset(content, 'a', sizeof(content));
+  content[sizeof(content) - 1] = '\0';
+  write_check_file(root, "file.txt", content);
+
+  char path[1024];
+  snprintf(path, sizeof(path), "%s/file.txt", root);
+  struct stat st;
+  EXPECT_EQ_INT(stat(path, &st), 0);
+  EXPECT_EQ_INT((int)st.st_size, 19999);
+
+  int p[2];
+  EXPECT_EQ_INT(socketpair(AF_UNIX, SOCK_STREAM, 0, p), 0);
+  io_set_fds(p[0], p[1]);
+  io_set_bwlimit(0);
+
+  pid_t pid = fork();
+  if (pid == 0) {
+    alarm(30);
+    close(p[1]);
+    io_set_fds(p[0], p[0]);
+    bool skipped = false;
+    File* file = receive_incremental_check(p[0], cfg, &skipped);
+    bool ok = file == NULL && !skipped;
+    if (ok)
+      send_status(p[0], STATUS_ERROR); /* mirror the server error path */
+    file_destroy(file);
+    config_delete(cfg);
+    close(p[0]);
+    _exit(ok ? 0 : 1);
+  } else {
+    close(p[0]);
+    io_set_fds(p[1], p[1]);
+    EXPECT_TRUE(send_str(p[1], "file.txt"));
+    unsigned long long size = (unsigned long long)st.st_size;
+    long long mtime = 1; /* different from the file mtime: force a transfer */
+    long long mtime_nsec = 0;
+    EXPECT_TRUE(send_n_data(p[1], &size, sizeof(size)));
+    EXPECT_TRUE(send_n_data(p[1], &mtime, sizeof(mtime)));
+    EXPECT_TRUE(send_n_data(p[1], &mtime_nsec, sizeof(mtime_nsec)));
+
+    Status s;
+    EXPECT_TRUE(receive_status(p[1], &s));
+    EXPECT_EQ_INT(s, STATUS_DELTA_SIGNATURE);
+    Data* sig_data = receive_data(p[1]);
+    EXPECT_NOT_NULL(sig_data);
+    DeltaSignature* sig = delta_signature_deserialize(sig_data);
+    EXPECT_NOT_NULL(sig);
+    delta_signature_destroy(sig);
+    data_destroy(sig_data);
+
+    /* Send a delta whose claimed output size exceeds the whole-file cap. */
+    Delta delta;
+    memset(&delta, 0, sizeof(delta));
+    delta.new_file_size = MAX_RECEIVE_WHOLE_FILE_SIZE + 1;
+    Data* bogus = delta_serialize(&delta);
+    EXPECT_NOT_NULL(bogus);
+    EXPECT_TRUE(send_status(p[1], STATUS_DELTA_DATA));
+    EXPECT_TRUE(bogus != NULL && send_data(p[1], bogus));
+    data_destroy(bogus);
+
+    /* The receiver must answer with an error, never with STATUS_NEXT. */
+    EXPECT_TRUE(receive_status(p[1], &s));
+    EXPECT_EQ_INT(s, STATUS_ERROR);
+
+    int status;
+    waitpid(pid, &status, 0);
+    close(p[1]);
+    config_delete(cfg);
+    unlink(path);
+    rmdir(root);
+    free(root);
+    EXPECT_TRUE(WIFEXITED(status) && WEXITSTATUS(status) == 0);
+  }
+}
+
 void test_server() {
   if (!is_running_under_valgrind()) {
     test_receive_files_finished();
@@ -217,5 +464,8 @@ void test_server() {
     test_receive_files_abort();
     test_receive_manifest_rejects_traversal();
     test_receive_incremental_check_rejects_invalid_nanoseconds();
+    test_incremental_check_quick_skip_by_mtime();
+    test_incremental_check_size_mismatch_full_transfer();
+    test_incremental_check_delta_oversize_reports_failure();
   }
 }