From e3a483f4e420293968f37f1359d00fe8343725fc Mon Sep 17 00:00:00 2001
From: Unbewohnte <unbewohnte@gmail.com>
Date: Sat, 8 Jan 2022 12:38:56 +0300
Subject: [PATCH] Check entry contents in the middle and in the end of the file
 for less false-positives; though it is still not good enough

---
 src/broom.cpp |  8 +++++---
 src/entry.cpp | 17 +++++++++++------
 src/entry.hpp |  8 ++++----
 src/main.cpp  | 20 ++++++++++++++++----
 4 files changed, 36 insertions(+), 17 deletions(-)

diff --git a/src/broom.cpp b/src/broom.cpp
index b03fbb5..64933d6 100644
--- a/src/broom.cpp
+++ b/src/broom.cpp
@@ -49,18 +49,20 @@ void Broom::track(const std::filesystem::path path) {
     if (std::filesystem::is_directory(path)) {
         // it`s a directory. Track every regular file recursively
         std::filesystem::directory_options options =  (
-            std::filesystem::directory_options::skip_permission_denied
+        std::filesystem::directory_options::skip_permission_denied
         );
 
         for (auto dir_entry : std::filesystem::recursive_directory_iterator(path, options)) {
             if (!dir_entry.is_regular_file()) {
+                // skip everything that we cannot process so easily
                 continue;
             };
 
             Entry entry(dir_entry.path());
             m_tracked_entries.push_back(entry);
-        };
+        }
     } else if (std::filesystem::is_regular_file(path)) {
+        // just a file
         Entry entry(path);
         m_tracked_entries.push_back(entry);
     }
@@ -127,7 +129,7 @@ uintmax_t Broom::untrack_unique_contents() {
 
     for (auto entry_iter = m_tracked_entries.begin(); entry_iter != m_tracked_entries.end();) {
         // the same logic:
-        // check if contents of this entry is already in the map
+        // check if contents of this entry are already in the map
         // if yes --> increment occurences counter
         // if not --> add it to the map with a counter of 1
 
diff --git a/src/entry.cpp b/src/entry.cpp
index 68943fb..034ceb7 100644
--- a/src/entry.cpp
+++ b/src/entry.cpp
@@ -37,7 +37,8 @@ void Entry::get_size() {
     filesize = std::filesystem::file_size(path);
 };
 
-// reads 2 pieces from the beginning and the end of a file, converts them into
+
+// reads 2 pieces from the middle and the end of a file, converts them into
 // a convenient hex-encoded string
 void Entry::get_pieces() {
     std::fstream entry_file;
@@ -47,18 +48,20 @@ void Entry::get_pieces() {
         throw std::ifstream::failure("Could not open \"" + path.string() + "\"; reason: " + std::string(std::strerror(errno)) + "\n");
     }
 
-    // TODO(Properly test it)
     char pieces_buffer[PIECE_SIZE * 2];
     if (filesize <= PIECE_SIZE * 2) {
         // can`t take whole 2 pieces !
         // read the whole file then
         entry_file.read(pieces_buffer, filesize);
     } else {
-        // read CHUNK_SIZE bytes from the beginning of the file
-        char start_buf[PIECE_SIZE];
-        entry_file.read(start_buf, PIECE_SIZE);
+        uintmax_t middle_of_the_file = (double) filesize / 2.0 - PIECE_SIZE;
+
+        entry_file.seekg(middle_of_the_file, std::ios::beg);
+        // read CHUNK_SIZE bytes from the middle of the file
+        char middle_buf[PIECE_SIZE];
+        entry_file.read(middle_buf, PIECE_SIZE);
         for (uint8_t i = 0; i < PIECE_SIZE; i++) {
-            pieces_buffer[i] = start_buf[i];
+            pieces_buffer[i] = middle_buf[i];
         };
 
         // jump to the last CHUNK_SIZE bytes of the file and read the as well
@@ -78,6 +81,8 @@ void Entry::get_pieces() {
     };
 
     pieces = pieces_hex.str();
+
+    std::cout << pieces << std::endl;
 };
 
 // Remove entry from the disk
diff --git a/src/entry.hpp b/src/entry.hpp
index d73ac96..4ee69fe 100644
--- a/src/entry.hpp
+++ b/src/entry.hpp
@@ -25,15 +25,15 @@ along with broom.  If not, see <https://www.gnu.org/licenses/>.
 #include <sstream>
 #include <iomanip>
 
-// 2 pieces (beginning and end of the file)
-const uint8_t PIECE_SIZE = 24;
+// 2 pieces (middle and end of the file)
+const uint8_t PIECE_SIZE = 16;
 
 // A wrapper for every file with all necessary information
 class Entry {
 public:
     std::filesystem::path path;
     uintmax_t filesize;
-    std::string pieces; // 2 hex-represented pieces of file (beginning and end)
+    std::string pieces; // 2 hex-represented pieces of file
 
     Entry(const std::filesystem::path entry_path);
     ~Entry();
@@ -41,7 +41,7 @@ public:
     // sets this entry`s filesize
     void get_size();
 
-    // reads 2 pieces from the beginning and the end of a file, converts them into
+    // reads 2 pieces from the middle and the end of a file, converts them into
     // a convenient hex-encoded string
     void get_pieces();
 
diff --git a/src/main.cpp b/src/main.cpp
index d39c846..429b2ec 100644
--- a/src/main.cpp
+++ b/src/main.cpp
@@ -88,21 +88,33 @@ int main(int argc, char* argv[]) {
     };
 
     // no path was specified at all
-    if (tracked_path.string() == "") {
+    if (tracked_path.empty()) {
         print_help();
         return 1;
     };
 
 
     Broom broom(options);
-
     try {
         broom.track(tracked_path);
+        broom.find_duplicates();
     } catch(const std::invalid_argument& e) {
-        std::cout << e.what() << std::endl;
+        std::cerr
+        << "[ERROR] Invalid argument: " << std::endl
+        << e.what() << std::endl;
+        return 1;
+
+    } catch(const std::filesystem::filesystem_error& e) {
+        std::cerr
+        << "[ERROR] FS error: " << std::endl
+        << e.what() << std::endl;
+        return 1;
+
+    } catch(...) {
+        std::cerr
+        << "[ERROR] Unexpected exception" << std::endl;
         return 1;
     };
-    broom.find_duplicates();
 
     return 0;
 };