cell/source/blob.h

#ifndef BLOB_H
#define BLOB_H

#include <stddef.h>
#include <stdint.h>
#include <stdlib.h>
#include <string.h>

// -----------------------------------------------------------------------------
// A simple blob structure that can be in two states:
//   - antestone (mutable): writing is allowed
//   - stone (immutable): reading is allowed
//
// The blob is stored as an array of bits in memory, but for simplicity here,
// we store them in a dynamic byte array with a bit_length and capacity in bits.
// -----------------------------------------------------------------------------

typedef struct blob {
  // The actual buffer holding the bits (in multiples of 8 bits).
  uint8_t *data;

  // The total number of bits currently in use (the "length" of the blob).
  size_t bit_length;

  // The total capacity in bits that 'data' can currently hold without realloc.
  size_t bit_capacity;

  // 0 = antestone (mutable)
  // 1 = stone (immutable)
  int is_stone;
} blob;

// Initialize a new blob
void blob_init(blob *b);

// Create a new blob with specified bit capacity
blob *blob_new(size_t bit_capacity);

// Create a blob from another blob (copy bits from 'from' to 'to')
blob *blob_new_from_blob(const blob *src, size_t from, size_t to);

// Create a blob with length bits, filled with logical value or random function
blob *blob_new_with_fill(size_t length_bits, int logical_value);

// Destroy blob and free memory
void blob_destroy(blob *b);

// Turn a blob into stone (immutable)
void blob_make_stone(blob *b);

// Write operations (only work on antestone blobs)
int blob_write_bit(blob *b, int bit_val);
int blob_write_blob(blob *b, const blob *src);
int blob_write_dec64(blob *b, double d);
int blob_write_fit(blob *b, int64_t fit, int length);
int blob_write_kim(blob *b, int64_t value);
int blob_write_pad(blob *b, int block_size);
int blob_write_text(blob *b, const char *text);

int blob_write_bytes(blob *b, void *data, size_t length);

// Read operations (only work on stone blobs)
int blob_read_bit(const blob *b, size_t pos, int *out_bit);
blob *blob_read_blob(const blob *b, size_t from, size_t to);
int blob_read_dec64(const blob *b, size_t from, double *out_value);
int blob_read_fit(const blob *b, size_t from, int length, int64_t *out_value);
int blob_read_kim(const blob *b, size_t from, int64_t *out_value, size_t *bits_read);
int blob_read_text(const blob *b, size_t from, char **out_text, size_t *bits_read);
int blob_pad_check(const blob *b, size_t from, int block_size);

// Utility functions
int kim_length_for_fit(int64_t value);

#ifdef BLOB_IMPLEMENTATION

// Helper to ensure capacity for writing
static int blob_ensure_capacity(blob *b, size_t new_bits) {
  size_t need_bits = b->bit_length + new_bits;
  if (need_bits <= b->bit_capacity) return 0;

  // Increase capacity (in multiples of bytes). 
  // We can pick a growth strategy. For demonstration, double it:
  size_t new_capacity = b->bit_capacity == 0 ? 64 : b->bit_capacity * 2;
  while (new_capacity < need_bits) new_capacity *= 2;

  // Round up new_capacity to a multiple of 8 bits
  if (new_capacity % 8) {
    new_capacity += 8 - (new_capacity % 8);
  }

  size_t new_size_bytes = new_capacity / 8;
  uint8_t *new_ptr = realloc(b->data, new_size_bytes);
  if (!new_ptr) {
    return -1; // out of memory
  }
  // zero-fill the new area (only beyond the old capacity)
  size_t old_size_bytes = b->bit_capacity / 8;
  if (new_size_bytes > old_size_bytes) {
    memset(new_ptr + old_size_bytes, 0, new_size_bytes - old_size_bytes);
  }
  b->data = new_ptr;
  b->bit_capacity = new_capacity;
  return 0;
}

void blob_init(blob *b) {
  b->data = NULL;
  b->bit_length = 0;
  b->bit_capacity = 0;
  b->is_stone = 0;
}

blob *blob_new(size_t bit_capacity) {
  blob *b = calloc(1, sizeof(blob));
  if (!b) return NULL;
  
  blob_init(b);
  
  if (bit_capacity > 0) {
    // Round up to multiple of 8
    if (bit_capacity % 8) {
      bit_capacity += 8 - (bit_capacity % 8);
    }
    b->bit_capacity = bit_capacity;
    size_t bytes = bit_capacity / 8;
    b->data = calloc(bytes, 1);
    if (!b->data) {
      free(b);
      return NULL;
    }
  }
  
  return b;
}

blob *blob_new_from_blob(const blob *src, size_t from, size_t to) {
  if (!src) return NULL;
  if (to > src->bit_length) to = src->bit_length;
  if (from >= to) return blob_new(0);
  
  size_t copy_len = to - from;
  blob *b = blob_new(copy_len);
  if (!b) return NULL;
  
  b->bit_length = copy_len;
  
  // Copy bits
  for (size_t i = 0; i < copy_len; i++) {
    size_t src_bit_index = from + i;
    size_t src_byte = src_bit_index >> 3;
    size_t src_off = src_bit_index & 7;
    int bit_val = (src->data[src_byte] >> src_off) & 1;
    size_t dst_byte = i >> 3;
    size_t dst_off = i & 7;
    if (bit_val) {
      b->data[dst_byte] |= (1 << dst_off);
    }
  }
  
  return b;
}

blob *blob_new_with_fill(size_t length_bits, int logical_value) {
  blob *b = blob_new(length_bits);
  if (!b) return NULL;
  
  b->bit_length = length_bits;
  
  if (logical_value) {
    // Fill with 1s
    size_t bytes = b->bit_capacity / 8;
    memset(b->data, 0xff, bytes);
    // Clear unused bits in last byte
    size_t used_bits_in_last_byte = length_bits & 7;
    if (used_bits_in_last_byte && bytes > 0) {
      uint8_t mask = (1 << used_bits_in_last_byte) - 1;
      b->data[bytes - 1] &= mask;
    }
  }
  
  return b;
}

void blob_destroy(blob *b) {
  if (b) {
    if (b && b->data) {
      free(b->data);
      b->data = NULL;
      b->bit_length = 0;
      b->bit_capacity = 0;
    }
    free(b);
  }
}

void blob_make_stone(blob *b) {
  if (!b) return;
  b->is_stone = 1;
  // Optionally shrink the buffer to exactly bit_length in size
  if (b->bit_capacity > b->bit_length) {
    size_t size_in_bytes = (b->bit_length + 7) >> 3; // round up to full bytes
    uint8_t *new_ptr = NULL;
    if (size_in_bytes) {
      new_ptr = realloc(b->data, size_in_bytes);
      if (new_ptr) {
        b->data = new_ptr;
      }
    } else {
      // zero length
      free(b->data);
      b->data = NULL;
    }
    b->bit_capacity = b->bit_length; // capacity in bits now matches length
  }
}

int blob_write_bit(blob *b, int bit_val) {
  if (!b || b->is_stone) return -1;
  
  if (blob_ensure_capacity(b, 1) < 0) {
    return -1;
  }
  
  size_t bit_index = b->bit_length;
  size_t byte_index = bit_index >> 3;
  size_t offset_in_byte = bit_index & 7;

  // set or clear bit
  if (bit_val)
    b->data[byte_index] |= (1 << offset_in_byte);
  else
    b->data[byte_index] &= ~(1 << offset_in_byte);

  b->bit_length++;
  return 0;
}

int blob_write_blob(blob *b, const blob *src) {
  if (!b || !src || b->is_stone) return -1;
  
  // Append all bits from src blob
  for (size_t i = 0; i < src->bit_length; i++) {
    size_t byte_idx = i / 8;
    size_t bit_idx = i % 8;
    int bit = (src->data[byte_idx] >> bit_idx) & 1;
    if (blob_write_bit(b, bit) < 0) {
      return -1;
    }
  }
  
  return 0;
}

int blob_write_dec64(blob *b, double d) {
  if (!b || b->is_stone) return -1;
  
  // Simple DEC64 encoding: store as IEEE 754 double (64 bits)
  uint64_t bits;
  memcpy(&bits, &d, sizeof(bits));
  
  // Write 64 bits
  for (int i = 0; i < 64; i++) {
    if (blob_write_bit(b, (bits >> i) & 1) < 0) {
      return -1;
    }
  }
  
  return 0;
}

int blob_write_fit(blob *b, int64_t fit, int length) {
  if (!b || b->is_stone) return -1;
  if (length < 0 || length > 64) return -1;
  
  // Check if fit requires more bits than allowed
  if (length < 64) {
    int64_t max = (1LL << length) - 1;
    if (fit < 0 || fit > max) {
      return -1;
    }
  }
  
  // Write the bits
  for (int i = 0; i < length; i++) {
    if (blob_write_bit(b, (fit >> i) & 1) < 0) {
      return -1;
    }
  }
  
  return 0;
}

// Calculate the kim length in bits for a fit (int64) value
int kim_length_for_fit(int64_t value) {
  if (value >= -1 && value <= 127) return 8;
  if (value >= -127 && value <= 16383) return 16;
  if (value >= -16383 && value <= 2097151) return 24;
  if (value >= -2097151 && value <= 268435455) return 32;
  if (value >= -268435455 && value <= 34359738367LL) return 40;
  if (value >= -34359738367LL && value <= 4398046511103LL) return 48;
  if (value >= -4398046511103LL && value <= 562949953421311LL) return 56;
  if (value >= -562949953421311LL && value <= 36028797018963967LL) return 64;
  if (value >= -36028797018963967LL) return 72;
  return 80; // Maximum kim length
}

int blob_write_kim(blob *b, int64_t value) {
  if (!b || b->is_stone) return -1;
  
  int bits = kim_length_for_fit(value);
  int bytes = bits / 8;
  
  uint8_t kim_bytes[10] = {0}; // Max 10 bytes for kim
  
  if (value < 0) {
    // Negative number: first byte is 0x80, then encode -value-1
    kim_bytes[0] = 0x80;
    value = -value - 1;
    
    // Encode remaining bytes
    for (int i = bytes - 1; i > 0; i--) {
      kim_bytes[i] = value & 0x7F;
      value >>= 7;
      if (i > 1) kim_bytes[i] |= 0x80; // Set continuation bit
    }
  } else {
    // Positive number
    for (int i = bytes - 1; i >= 0; i--) {
      kim_bytes[i] = value & 0x7F;
      value >>= 7;
      if (i > 0) kim_bytes[i] |= 0x80; // Set continuation bit
    }
  }
  
  // Write the kim bytes as bits
  for (int i = 0; i < bytes; i++) {
    for (int j = 0; j < 8; j++) {
      if (blob_write_bit(b, (kim_bytes[i] >> j) & 1) < 0)
        return -1;
    }
  }
  
  return 0;
}

int blob_write_pad(blob *b, int block_size) {
  if (!b || b->is_stone) return -1;
  if (block_size <= 0) return -1;
  
  // Write a 1 bit
  if (blob_write_bit(b, 1) < 0) {
    return -1;
  }
  
  // Calculate how many 0 bits to add
  size_t current_len = b->bit_length;
  size_t remainder = current_len % block_size;
  if (remainder > 0) {
    size_t zeros_needed = block_size - remainder;
    for (size_t i = 0; i < zeros_needed; i++) {
      if (blob_write_bit(b, 0) < 0) {
        return -1;
      }
    }
  }
  
  return 0;
}

int blob_write_text(blob *b, const char *text) {
  if (!b || !text || b->is_stone) return -1;
  
  size_t len = strlen(text);
  
  // Write kim-encoded length
  if (blob_write_kim(b, len) < 0) {
    return -1;
  }
  
  // Write each character as kim-encoded UTF-32
  // For simplicity, assuming ASCII
  for (size_t i = 0; i < len; i++) {
    if (blob_write_kim(b, (unsigned char)text[i]) < 0) {
      return -1;
    }
  }
  
  return 0;
}

int blob_read_bit(const blob *b, size_t pos, int *out_bit) {
  if (!b || !b->is_stone || !out_bit) return -1;
  if (pos >= b->bit_length) return -1;
  
  size_t byte_index = pos >> 3;
  size_t offset_in_byte = pos & 7;
  *out_bit = (b->data[byte_index] & (1 << offset_in_byte)) ? 1 : 0;
  return 0;
}

blob *blob_read_blob(const blob *b, size_t from, size_t to) {
  if (!b || !b->is_stone) return NULL;
  return blob_new_from_blob(b, from, to);
}

int blob_read_dec64(const blob *b, size_t from, double *out_value) {
  if (!b || !b->is_stone || !out_value) return -1;
  if (from + 64 > b->bit_length) return -1;
  
  // Read 64 bits
  uint64_t bits = 0;
  for (int i = 0; i < 64; i++) {
    int bit;
    blob_read_bit(b, from + i, &bit);
    if (bit) bits |= (1ULL << i);
  }
  
  // Convert to double
  memcpy(out_value, &bits, sizeof(*out_value));
  return 0;
}

int blob_read_fit(const blob *b, size_t from, int length, int64_t *out_value) {
  if (!b || !b->is_stone || !out_value) return -1;
  if (length < 0 || length > 64) return -1;
  if (from + length > b->bit_length) return -1;
  
  // Read bits
  int64_t value = 0;
  for (int i = 0; i < length; i++) {
    int bit;
    blob_read_bit(b, from + i, &bit);
    if (bit) value |= (1LL << i);
  }
  
  *out_value = value;
  return 0;
}

int blob_read_kim(const blob *b, size_t from, int64_t *out_value, size_t *bits_read) {
  if (!b || !b->is_stone || !out_value || !bits_read) return -1;
  
  size_t pos = from;
  uint8_t bytes[10];
  int byte_count = 0;
  
  // Read bytes until we find one without continuation bit
  while (byte_count < 10) {
    if (pos + 8 > b->bit_length) return -1;
    
    uint8_t byte = 0;
    for (int i = 0; i < 8; i++) {
      int bit;
      if (blob_read_bit(b, pos + i, &bit) < 0)
        return -1;
      if (bit) byte |= (1 << i);
    }
    
    bytes[byte_count++] = byte;
    pos += 8;
    
    if (!(byte & 0x80)) break; // No continuation bit
  }
  
  if (byte_count == 0) return -1;
  
  // Decode the kim value
  int64_t value = 0;
  int is_negative = (bytes[0] == 0x80);
  
  if (is_negative) {
    // Skip the 0x80 byte and decode remaining
    for (int i = 1; i < byte_count; i++) {
      value = (value << 7) | (bytes[i] & 0x7F);
    }
    value = -value - 1;
  } else {
    for (int i = 0; i < byte_count; i++) {
      value = (value << 7) | (bytes[i] & 0x7F);
    }
  }
  
  *out_value = value;
  *bits_read = byte_count * 8;
  return 0;
}

int blob_read_text(const blob *b, size_t from, char **out_text, size_t *bits_read) {
  if (!b || !b->is_stone || !out_text || !bits_read) return -1;
  if (from >= b->bit_length) return -1;
  
  // Read kim-encoded length
  int64_t length;
  size_t len_bits;
  if (blob_read_kim(b, from, &length, &len_bits) < 0) {
    return -1;
  }
  
  if (length < 0) return -1;
  
  size_t pos = from + len_bits;
  
  // Read characters
  char *str = malloc(length + 1);
  if (!str) return -1;
  
  for (int64_t i = 0; i < length; i++) {
    int64_t ch;
    size_t ch_bits;
    if (blob_read_kim(b, pos, &ch, &ch_bits) < 0) {
      free(str);
      return -1;
    }
    // For simplicity, assuming ASCII
    str[i] = (char)(ch & 0xFF);
    pos += ch_bits;
  }
  str[length] = '\0';
  
  *out_text = str;
  *bits_read = pos - from;
  return 0;
}

int blob_pad_check(const blob *b, size_t from, int block_size) {
  if (!b || !b->is_stone) return 0;
  if (block_size <= 0) return 0;
  
  // Check if blob length is multiple of block_size
  if (b->bit_length % block_size != 0) {
    return 0;
  }
  
  // Check if difference between length and from is <= block_size
  int64_t diff = b->bit_length - from;
  if (diff <= 0 || diff > block_size) {
    return 0;
  }
  
  // Check if bit at from is 1
  int bit;
  if (blob_read_bit(b, from, &bit) < 0 || bit != 1) {
    return 0;
  }
  
  // Check if remaining bits are 0
  for (size_t i = from + 1; i < b->bit_length; i++) {
    if (blob_read_bit(b, i, &bit) < 0 || bit != 0) {
      return 0;
    }
  }
  
  return 1;
}

#endif /* BLOB_IMPLEMENTATION */

#endif /* BLOB_H */