Add a C encoder

It only works on little endian systems atm, because handling of
endianess is a pain in C. The C# and Python encoders handle it fine.

rif.h

@@ -0,0 +1,30 @@
+typedef struct rif {
+  // "lifV0001"
+  // "rifV0001"
+  char     magic_num[8]; // Big Endian. UTF-8 Encoded. 6C 69 66 56 30 30 30 31
+  uint64_t width;        // Big Endian
+  uint8_t  format;
+  uint8_t  data[];
+} rif_t;
+
+#define RIF_MAGIC_LITTLE {'l', 'i', 'f', 'V', '0', '0', '0', '1'}
+#define RIF_MAGIC_BIG    {'r', 'i', 'f', 'V', '0', '0', '0', '1'}
+
+#define RIF_FORMAT_R8G8B8   0
+#define RIF_FORMAT_R8G8B8A8 1
+
+// TODO detect endianess - this breaks on big endian systems.
+void write_rif_little(char* path, uint32_t width, uint32_t size, uint8_t format, void* data) {
+  // TODO is this better than just hardcoding 17?
+  const uint32_t header_size = sizeof(rif_t) - sizeof(uint8_t*);
+  const uint64_t img_size = header_size + size;
+  rif_t* img = malloc(img_size);
+  char magic[] = RIF_MAGIC_LITTLE;
+  memcpy(&img->magic_num, magic, 8);
+  img->width = width;
+  img->format = format;
+  memcpy(&img->data, data, size);
+  FILE* output_file = fopen(path, "w");
+  fwrite(img, 1, img_size, output_file);
+  free(img);
+}