LRN: adding generic functions for conversion of binary data to ascii and back: GNUNET_CRYPTO_string_to_data and GNUNET_CRYPTO_data_to_string

1 files changed, 118 insertions, 1 deletions

diff --git a/src/util/crypto_hash.c b/src/util/crypto_hash.c
index 63d965406..dce449ff0 100644
--- a/src/util/crypto_hash.c
+++ b/src/util/crypto_hash.c
@@ -241,7 +241,6 @@ GNUNET_CRYPTO_hash_file_cancel (struct GNUNET_CRYPTO_FileHashContext *fhc)
 }
 
 
-
 /* ***************** binary-ASCII encoding *************** */
 
 /**
@@ -262,6 +261,124 @@ getValue__ (unsigned char a)
 
 
 /**
+ * Convert binary data to ASCII encoding.  The ASCII encoding is rather
+ * GNUnet specific.  It was chosen such that it only uses characters
+ * in [0-9A-V], can be produced without complex arithmetics and uses a
+ * small number of characters.  The GNUnet encoding uses 103 characters.
+ * Does not append 0-terminator, but returns a pointer to the place where
+ * it should be placed, if needed.
+ *
+ * @param data data to encode
+ * @param size size of data (in bytes)
+ * @param out buffer to fill
+ * @param out_size size of the buffer. Must be large enough to hold
+ * ((size*8) + (((size*8) % 5) > 0 ? 5 - ((size*8) % 5) : 0)) / 5 bytes
+ * @return pointer to the next byte in 'out' or NULL on error.
+ */
+char *
+GNUNET_CRYPTO_data_to_string (unsigned char *data, size_t size, char *out, size_t out_size)
+{
+  /**
+   * 32 characters for encoding (GNUNET_CRYPTO_hash => 32 characters)
+   */
+  static char *encTable__ = "0123456789ABCDEFGHIJKLMNOPQRSTUV";
+  unsigned int wpos;
+  unsigned int rpos;
+  unsigned int bits;
+  unsigned int vbit;
+
+  GNUNET_assert (data != NULL);
+  GNUNET_assert (out != NULL);
+  GNUNET_assert (out_size >= (((size*8) + ((size*8) % 5)) % 5));
+  vbit = 0;
+  wpos = 0;
+  rpos = 0;
+  bits = 0;
+  while ((rpos < size) || (vbit > 0))
+  {
+    if ((rpos < size) && (vbit < 5))
+    {
+      bits = (bits << 8) | data[rpos++];   /* eat 8 more bits */
+      vbit += 8;
+    }
+    if (vbit < 5)
+    {
+      bits <<= (5 - vbit);      /* zero-padding */
+      GNUNET_assert (vbit == ((size * 8) % 5));
+      vbit = 5;
+    }
+    if (wpos >= out_size)
+      return NULL;
+    out[wpos++] = encTable__[(bits >> (vbit - 5)) & 31];
+    vbit -= 5;
+  }
+  if (wpos != out_size)
+    return NULL;
+  GNUNET_assert (vbit == 0);
+  return &out[wpos];
+}
+
+
+/**
+ * Convert ASCII encoding back to data
+ * out_size must match exactly the size of the data before it was encoded.
+ *
+ * @param enc the encoding
+ * @param enclen number of characters in 'enc' (without 0-terminator, which can be missing)
+ * @param out location where to store the decoded data
+ * @param out_size sizeof the output buffer
+ * @return GNUNET_OK on success, GNUNET_SYSERR if result has the wrong encoding
+ */
+int
+GNUNET_CRYPTO_string_to_data (const char *enc, size_t enclen,
+                              unsigned char *out, size_t out_size)
+{
+  unsigned int rpos;
+  unsigned int wpos;
+  unsigned int bits;
+  unsigned int vbit;
+  int ret;
+  int shift;
+  int encoded_len = out_size * 8;
+  if (encoded_len % 5 > 0)
+  {
+    vbit = encoded_len % 5; /* padding! */
+    shift = 5 - vbit;
+  }
+  else
+  {
+    vbit = 0;
+    shift = 0;
+  }
+  if ((encoded_len + shift) / 5 != enclen)
+    return GNUNET_SYSERR;
+
+  wpos = out_size;
+  rpos = enclen;
+  bits = (ret = getValue__ (enc[--rpos])) >> (5 - encoded_len % 5);
+  if (-1 == ret)
+    return GNUNET_SYSERR;
+  while (wpos > 0)
+  {
+    GNUNET_assert (rpos > 0);
+    bits = ((ret = getValue__ (enc[--rpos])) << vbit) | bits;
+    if (-1 == ret)
+      return GNUNET_SYSERR;
+    vbit += 5;
+    if (vbit >= 8)
+    {
+      out[--wpos] = (unsigned char) bits;
+      bits >>= 8;
+      vbit -= 8;
+    }
+  }
+  GNUNET_assert (rpos == 0);
+  GNUNET_assert (vbit == 0);
+  return GNUNET_OK;
+}
+
+
+/**
  * Convert GNUNET_CRYPTO_hash to ASCII encoding.  The ASCII encoding is rather
  * GNUnet specific.  It was chosen such that it only uses characters
  * in [0-9A-V], can be produced without complex arithmetics and uses a