]> wimlib.net Git - wimlib/commitdiff
make_huffman_decode_table(): Fix comment v1.2.2
authorEric Biggers <ebiggers3@gmail.com>
Fri, 21 Dec 2012 18:52:14 +0000 (12:52 -0600)
committerEric Biggers <ebiggers3@gmail.com>
Fri, 21 Dec 2012 18:52:14 +0000 (12:52 -0600)
src/decompress.c

index ebbf33e3e9b481f2c4d0fcbdaa8d4bbc5f07ceb6..e97aab5c5c2f72cbf407ccacc65f8a750fb62f01 100644 (file)
@@ -38,7 +38,8 @@
  *                     (2**table_bits) + 2 * num_syms to guarantee
  *                     that there is enough space.
  *
  *                     (2**table_bits) + 2 * num_syms to guarantee
  *                     that there is enough space.
  *
- * @num_syms:          Total number of symbols in the Huffman tree.
+ * @num_syms:          Number of symbols in the alphabet, including symbols
+ *                     that do not appear in this particular input chunk.
  *
  * @table_bits:                Any symbols with a code length of table_bits or less can
  *                     be decoded in one lookup of the table.  2**table_bits
  *
  * @table_bits:                Any symbols with a code length of table_bits or less can
  *                     be decoded in one lookup of the table.  2**table_bits
@@ -46,7 +47,7 @@
  *                     any Huffman codes longer than @table_bits.
  *
  * @lens:              An array of length @num_syms, indexable by symbol, that
  *                     any Huffman codes longer than @table_bits.
  *
  * @lens:              An array of length @num_syms, indexable by symbol, that
- *                     gives the length of the Huffman codeward for that
+ *                     gives the length of the Huffman codeword for that
  *                     symbol.  Because the Huffman tree is in canonical form,
  *                     it can be reconstructed by only knowing the length of
  *                     the codeword for each symbol.  It is assumed, but not
  *                     symbol.  Because the Huffman tree is in canonical form,
  *                     it can be reconstructed by only knowing the length of
  *                     the codeword for each symbol.  It is assumed, but not