From d2a47b1a5ae6548a55c038c72b4af1549c827cf5 Mon Sep 17 00:00:00 2001
From: Claudio Maggioni <claudio@maggioni.xyz>
Date: Fri, 18 Oct 2019 09:43:32 +0200
Subject: [PATCH] grad_hw1: done

---
 .gitignore          |   4 +
 splitter/splitter.c | 193 ++++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 197 insertions(+)
 create mode 100644 splitter/splitter.c

diff --git a/.gitignore b/.gitignore
index 6bc2222..9feebf5 100644
--- a/.gitignore
+++ b/.gitignore
@@ -1,3 +1,7 @@
 sortlines/sortlines
 **/.ccls-cache
 **/.gdb_history
+
+**/*.dSYM/**/*
+splitter/splitter
+*.o
diff --git a/splitter/splitter.c b/splitter/splitter.c
new file mode 100644
index 0000000..4a4699b
--- /dev/null
+++ b/splitter/splitter.c
@@ -0,0 +1,193 @@
+// vim: set ts=2 sw=2 et tw=80:
+
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <math.h>
+
+const int MAX_SPLIT = 1024;
+const int BUFFER_SIZE = 1024;
+
+int usage();
+int split(const char* filename, long max_size);
+int merge(const char* index, const char* merged);
+
+inline int usage(const char* s) {
+  fprintf(stderr, "Usage: %s split [filename] [max-size]\n"
+      "       %s merge [index] [merged]\n", s, s);
+  return 129;
+}
+
+int main(int argc, char** const argv) {
+  if (argc != 4) return usage(argv[0]);
+  if (strcmp(argv[1], "split") == 0) {
+    long max = strtol(argv[3], NULL, 10);
+    if (max <= 0) {
+      return usage(argv[0]);
+    }
+    return split(argv[2], max);
+  } else if (strcmp(argv[1], "merge") == 0) {
+    return merge(argv[2], argv[3]);
+  } else {
+    return usage(argv[0]);
+  }
+}
+
+unsigned char checksum(unsigned char start,
+    const unsigned char* buffer, long size) {
+  unsigned char sum = start;
+  for (long i = 0; i < size; i++) {
+    sum ^= buffer[i];
+  }
+  return sum;
+}
+
+int split(const char* filename, long max_size) {
+  FILE* source = fopen(filename, "r");
+  if (source == NULL) {
+    perror("Cannot open source file");
+    return 1;
+  }
+
+  const int fnlen = strlen(filename);
+  char index_name[fnlen + 6]; // VLA
+  sprintf(index_name, "%s.index", filename);
+  FILE* index = fopen(index_name, "wx");
+
+  if (index == NULL) {
+    perror("Cannot create index file");
+    fclose(source);
+    return 2;
+  }
+
+  char part_name[fnlen + 10];
+  unsigned char buffer[BUFFER_SIZE];
+
+  for(int part = 1; !feof(source); part++) {
+    // assign file name to part
+    sprintf(part_name, "%s.part%d", filename, part);
+
+    FILE* part_file = fopen(part_name, "wx");
+
+    if (part_file == NULL) {
+      perror("Cannot create part file");
+      fclose(source);
+      fclose(index);
+      return 4;
+    }
+
+    unsigned char sum = 0x00;
+
+    for (long i = max_size; i > 0 && !feof(source);) {
+      const long to_read = i > BUFFER_SIZE ? BUFFER_SIZE : i;
+      const long bytes = fread(buffer, 1, to_read, source);
+
+      if (bytes == 0 || ferror(source)) {
+        perror("Cannot read source file");
+        fclose(source);
+        fclose(index);
+        return 3;
+      }
+
+      sum = checksum(sum, buffer, bytes);
+
+      const long w_bytes = fwrite(buffer, 1, bytes, part_file);
+
+      if (w_bytes == 0) {
+        perror("Cannot write on part file");
+        fclose(source);
+        fclose(index);
+        fclose(part_file);
+        return 5;
+      }
+
+      i -= bytes;
+    }
+
+    if (fprintf(index, "%s %02x\n", part_name, sum) < 0) {
+      fprintf(stderr, "Cannot update index file");
+      fclose(source);
+      fclose(index);
+      fclose(part_file);
+      return 6;
+    }
+
+    fclose(part_file);
+  }
+
+  fclose(source);
+  fclose(index);
+  return 0;
+}
+
+int merge(const char* n_index, const char* n_merged) {
+  FILE* index = fopen(n_index, "r");
+
+  if (index == NULL) {
+    perror("Cannot open index file");
+    return 7;
+  }
+
+  FILE* merged = fopen(n_merged, "wx");
+
+  if (merged == NULL) {
+    perror("Cannot create merged file");
+    fclose(index);
+    return 8;
+  }
+
+  char part_name[strlen(n_index) + 100];
+
+  while(!feof(index)) {
+    unsigned int sum;
+    fscanf(index, "%s %x\n", part_name, &sum);
+
+    FILE* part = fopen(part_name, "r");
+
+    if (part == NULL) {
+      perror("Cannot open part");
+      fclose(index);
+      fclose(merged);
+      return 9;
+    }
+
+    unsigned char check = 0x00;
+    while (!feof(part)) {
+      unsigned char buffer[1000];
+      long bytes = fread(buffer, 1, 1000, part);
+
+      if (ferror(part)) {
+        perror("Cannot read part");
+        fclose(index);
+        fclose(merged);
+        fclose(part);
+        return 10;
+      }
+
+      check = checksum(check, buffer, bytes);
+      fwrite(buffer, 1, bytes, merged);
+
+      if (ferror(part)) {
+        perror("Cannot wrtite on merged");
+        fclose(index);
+        fclose(merged);
+        fclose(part);
+        return 11;
+      }
+    }
+
+    fclose(part);
+
+    if ((unsigned char) sum != check) {
+      perror("Checksums differ for part");
+      fclose(index);
+      fclose(merged);
+      fclose(part);
+      return 12;
+    }
+  }
+
+  fclose(index);
+  fclose(merged);
+  return 0;
+}