Replacing printf, putchar and snprintf in the hot path with custom functions

Replacing snprintf(), localtime() and strftime() with custom formating
functions

Reworking progress_thread(), instead of composing the progress printing
with multiple printf() and putchar, we compose it in a buffer and write
it at onse using WriteConsole in Windows and write in Linux

CMakeLists.txt

@@ -1,7 +1,7 @@
 cmake_minimum_required(VERSION 3.20)
-project(filehasher
+project(dfin
     VERSION 1.0.0
-    DESCRIPTION "High-performance file hasher with I/O Ring/io_uring support"
+    DESCRIPTION "High-performance duplicate finder with I/O Ring/io_uring support"
     LANGUAGES C
 )
 
@@ -106,7 +106,7 @@ endif()
 # ---------------------------------------------------------------------------
 
 set(SOURCES
-    file_hasher.c
+    duplicate_finder.c
     xxhash.c
     xxh_x86dispatch.c
 )
@@ -116,7 +116,7 @@ set(HEADERS
     arena.h
     base.h
     xxhash.h
-    lf_mpmc.h
+    mt_mpmc.h
 )
 
 # ---------------------------------------------------------------------------

arena.c

@@ -196,7 +196,7 @@ mem_arena *arena_create(arena_params *params) { // mk create
       arena->free_list = arena_create(&(arena_params){
           .reserve_size = MiB(1),
           .commit_size = MiB(1),
-          .align = ARENA_ALIGN,
+          .align = ARENA_CACHE_ALIGN,
           .push_size = sizeof(arena_free_node),
           .allow_free_list = false,
           .free_list = NULL,
@@ -620,6 +620,119 @@ void *arena_swapback_pop(mem_arena **arena_ptr, u64 index) { // mk swapback
 /* ============================================================
    Utilities
    ============================================================ */
+typedef enum arena_trim_flags {
+  ARENA_TRIM_NONE = 0,
+
+  ARENA_TRIM_SPACE = 1 << 0,
+  ARENA_TRIM_TAB = 1 << 1,
+  ARENA_TRIM_LF = 1 << 2,
+  ARENA_TRIM_CR = 1 << 3,
+  ARENA_TRIM_NUL = 1 << 4,
+
+} arena_trim_flags;
+
+u64 arena_trim_string(mem_arena **arena_ptr, char *str, u8 termination_flags) {
+  ASSERT(arena_ptr);
+  ASSERT(*arena_ptr);
+  ASSERT(str);
+
+  if (!arena_ptr || !*arena_ptr || !str) {
+    return 0;
+  }
+
+  mem_arena *arena = *arena_ptr;
+
+  /* ------------------------------------------------------------
+     Find owning block
+     ------------------------------------------------------------ */
+
+  mem_arena *owner = arena_block_from_ptr(arena, (u8 *)str);
+
+  ASSERT(owner);
+  if (!owner) {
+    return 0;
+  }
+
+  /* ------------------------------------------------------------
+     Must be current block
+     ------------------------------------------------------------ */
+
+  if (owner != arena) {
+    fprintf(stderr, "arena_trim_string(): string is not "
+                    "in current arena block.\n");
+    return 0;
+  }
+
+  /* ------------------------------------------------------------
+     Compute string position
+     ------------------------------------------------------------ */
+
+  u64 str_pos = arena_pos_from_ptr(arena, str);
+
+  /* ------------------------------------------------------------
+     Original reserved size
+     ------------------------------------------------------------ */
+
+  u64 allocated_size = arena->pos - str_pos;
+
+  /* ------------------------------------------------------------
+     Compute sizes
+     ------------------------------------------------------------ */
+
+  u64 str_size = strlen(str);
+
+  char *dst = str + str_size;
+  u64 termination_size = 0;
+
+  if (termination_flags & ARENA_TRIM_SPACE) {
+    *dst++ = ' ';
+    termination_size++;
+  }
+
+  if (termination_flags & ARENA_TRIM_TAB) {
+    *dst++ = '\t';
+    termination_size++;
+  }
+
+  if (termination_flags & ARENA_TRIM_CR) {
+    *dst++ = '\r';
+    termination_size++;
+  }
+
+  if (termination_flags & ARENA_TRIM_LF) {
+    *dst++ = '\n';
+    termination_size++;
+  }
+
+  if (termination_flags & ARENA_TRIM_NUL) {
+    *dst++ = '\0';
+    termination_size++;
+  }
+  /* ------------------------------------------------------------
+     Final used size
+     ------------------------------------------------------------ */
+
+  u64 used_size = str_size + termination_size;
+
+  used_size = ALIGN_UP_POW2(used_size, arena->align);
+
+  /* ------------------------------------------------------------
+     Overflow detection
+     ------------------------------------------------------------ */
+
+  if (used_size > allocated_size) {
+    fprintf(stderr, "arena_trim_string(): string overflow "
+                    "detected.\n");
+  }
+
+  /* ------------------------------------------------------------
+     Update arena position
+     ------------------------------------------------------------ */
+
+  arena->pos = str_pos + used_size;
+
+  return used_size;
+}
 
 void *arena_clear(mem_arena **arena_ptr) { // mk clear
 
@@ -801,7 +914,7 @@ mem_arena_temp arena_scratch_get(mem_arena **conflicts, u32 num_conflicts) {
     arena_params params = {
         .reserve_size = MiB(64),
         .commit_size = MiB(1),
-        .align = ARENA_ALIGN,
+        .align = ARENA_CACHE_ALIGN,
         .push_size = 8,
         .allow_free_list = false,
         .allow_swapback = true,

arena.h

@@ -239,7 +239,7 @@ void *arena_ptr_from_index(mem_arena *arena, u64 index);
 */
 
 #define ARENA_HEADER_SIZE (sizeof(mem_arena))
-#define ARENA_ALIGN (sizeof(void *))
+#define ARENA_CACHE_ALIGN (sizeof(void *))
 
 // arena config
 typedef enum arena_growth_policy {

base.h

@@ -35,6 +35,7 @@
 #include <sys/stat.h>
 #include <sys/vfs.h>
 #include <unistd.h>
+#include <sys/mman.h>
 #endif
 
 #include <assert.h>
@@ -147,9 +148,6 @@ static void sleep_ms(int ms) { Sleep(ms); }
 #define _DEFAULT_SOURCE
 #endif
 
-#include <sys/mman.h>
-#include <unistd.h>
-
 static u32 plat_get_pagesize(void) { return (u32)sysconf(_SC_PAGESIZE); }
 
 static void *plat_mem_reserve(u64 size) {

build.bat

@@ -1,6 +1,8 @@
 @echo off
 setlocal enabledelayedexpansion
 
+set PROJECT_NAME=dfin
+
 :: ============================================================================
 :: build.bat
 :: ============================================================================
@@ -45,7 +47,7 @@ exit /b 1
 :main
 set BUILD_DIR=%SCRIPT_DIR%\build\windows\%BUILD_TYPE%
 
-echo === Building filehasher (%BUILD_TYPE%) ===
+echo === Building %PROJECT_NAME% (%BUILD_TYPE%) ===
 
 :: --------------------------------------------------------------------------
 :: Clean if requested
@@ -167,4 +169,4 @@ popd
 
 echo.
 echo === Build Complete ===
-echo Executable: %BUILD_DIR%\filehasher.exe
+echo Executable: %BUILD_DIR%\%PROJECT_NAME%.exe

build.sh

@@ -1,6 +1,6 @@
 #!/usr/bin/env bash
 # ============================================================================
-# build.sh - Build script for filehasher (Linux)
+# build.sh - Build script (Linux)
 # Usage: ./build.sh [Release|Debug] [clean]
 #
 # Compiler preference: gcc > clang
@@ -9,6 +9,8 @@
 
 set -euo pipefail
 
+PROJECT_NAME="dfin"
+
 # ---------------------------------------------------------------------------
 # Colors
 # ---------------------------------------------------------------------------
@@ -55,7 +57,7 @@ done
 readonly BUILD_DIR="build/linux/${BUILD_TYPE}"
 readonly SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
 
-echo -e "${GREEN}=== Building filehasher (${BUILD_TYPE}) ===${NC}"
+echo -e "${GREEN}=== Building ${PROJECT_NAME} (${BUILD_TYPE}) ===${NC}"
 echo "Project: ${SCRIPT_DIR}"
 
 # ---------------------------------------------------------------------------
@@ -226,18 +228,18 @@ echo
 # ---------------------------------------------------------------------------
 cd "${SCRIPT_DIR}"
 
-if [[ -f "${BUILD_DIR}/filehasher" ]]; then
+if [[ -f "${BUILD_DIR}/${PROJECT_NAME}" ]]; then
-    echo -e "${GREEN}Executable: ${BUILD_DIR}/filehasher${NC}"
+    echo -e "${GREEN}Executable: ${BUILD_DIR}/${PROJECT_NAME}${NC}"
 
     if command -v file &> /dev/null; then
-        echo -e "  Type: $(file -b ${BUILD_DIR}/filehasher)"
+        echo -e "  Type: $(file -b ${BUILD_DIR}/${PROJECT_NAME})"
     fi
 
     if command -v du &> /dev/null; then
-        echo -e "  Size: $(du -h ${BUILD_DIR}/filehasher | cut -f1)"
+        echo -e "  Size: $(du -h ${BUILD_DIR}/${PROJECT_NAME} | cut -f1)"
     fi
-elif [[ -f "${BUILD_DIR}/filehasher.exe" ]]; then
+elif [[ -f "${BUILD_DIR}/${PROJECT_NAME}.exe" ]]; then
-    echo -e "${GREEN}Executable: ${BUILD_DIR}/filehasher.exe${NC}"
+    echo -e "${GREEN}Executable: ${BUILD_DIR}/${PROJECT_NAME}.exe${NC}"
 else
     echo -e "${YELLOW}Note: Could not locate executable${NC}"
     echo "Checking build directory:"
@@ -269,4 +271,4 @@ if [[ "${EXPORT_COMPILE_COMMANDS}" == "ON" ]]; then
 fi
 
 echo
-echo -e "${GREEN}Ready to run: ./${BUILD_DIR}/filehasher${NC}"
+echo -e "${GREEN}Ready to run: ./${BUILD_DIR}/${PROJECT_NAME}${NC}"

duplicate_finder.c

@@ -32,7 +32,11 @@ int main(int argc, char **argv) {
     buf[strcspn(buf, "\r\n")] = 0;
 
     if (buf[0] == 0) {
-      strcpy(folders[0], ".");
+      if (!platform_get_current_directory(folders[0], sizeof(folders[0]))) {
+        fprintf(stderr, "Failed to get current directory\n");
+        return 1;
+      }
+      normalize_path(folders[0]);
       folder_count = 1;
     } else {
       folder_count = parse_paths(buf, folders, 64);
@@ -71,7 +75,19 @@ int main(int argc, char **argv) {
       .max_nbre_blocks = 1,
   };
 
-  mem_arena *gp_arena = arena_create(&params);
+  arena_params params_caligned = {
+      .reserve_size = GiB(1),
+      .commit_size = MiB(16),
+      .align = ARENA_CACHE_ALIGN,
+      .push_size = 0,
+      .allow_free_list = true,
+      .allow_swapback = false,
+      .growth_policy = ARENA_GROWTH_NORMAL,
+      .commit_policy = ARENA_COMMIT_LAZY,
+      .max_nbre_blocks = 1,
+  };
+
+  mem_arena *gp_arena = arena_create(&params_caligned);
 
   // -------------------------------
   // Detect hardware
@@ -119,7 +135,7 @@ int main(int argc, char **argv) {
   mpmc_init(&file_queue, MiB(1));
 
   // Starting hash threads
-  WorkerContext workers[num_hash_threads];
+  HasherContext workers[num_hash_threads];
   Thread *hash_threads =
       arena_push(&gp_arena, sizeof(Thread) * num_hash_threads, true);
 
@@ -155,8 +171,7 @@ int main(int argc, char **argv) {
 
   for (uint8_t i = 0; i < num_scan_threads; i++) {
     scanners[i].num_threads = num_scan_threads;
-    scanners[i].path_arena = arena_create(&params);
+    scanners[i].meta_arena = arena_create(&params_caligned);
-    scanners[i].meta_arena = arena_create(&params);
     scanners[i].dir_queue = &dir_queue;
     scanners[i].file_queue = &file_queue;
 
@@ -170,7 +185,7 @@ int main(int argc, char **argv) {
   // Initial folder push
   for (int i = 0; i < folder_count; i++) {
     size_t len = strlen(folders[i]) + 1;
-    char *path = arena_push(&scanners[0].path_arena, len, false);
+    char *path = arena_push(&scanners[0].meta_arena, len, false);
     memcpy(path, folders[i], len);
     mpmc_push_work(&dir_queue, path);
   }

mt_mpmc.h

@@ -214,19 +214,27 @@ static void mpmc_producers_finished(MPMCQueue *q, u8 consumer_count) {
 /* Done */
 /* ----------------------------------------------------------- */
 static void mpmc_task_done(MPMCQueue *q, u8 consumer_count) {
+
+  bool finished = false;
+
   mtx_lock(&q->lock);
 
   if (--q->work_count == 0) {
-     mpmc_producers_finished(q, consumer_count);
+    finished = true;
   }
 
   mtx_unlock(&q->lock);
+
+  if (finished) {
+    mpmc_producers_finished(q, consumer_count);
+  }
 }
 
 /* ----------------------------------------------------------- */
 /* MPMC Cleanup */
 /* ----------------------------------------------------------- */
-// static void mpmc_finish(MPMCQueue *q) { // Comment to prevent warning: unused function
+// static void mpmc_finish(MPMCQueue *q) { // Comment to prevent warning: unused
+// function
 //   if (!q) return;
 //
 //   if (q->slots) {