cuberite/libdeflate
1
0
Fork 0
mirror of https://github.com/cuberite/libdeflate.git synced 2025-08-04 02:06:31 -04:00
Code Issues Packages Projects Releases Wiki Activity

lib: fix some typos in comments

Browse Source
This commit is contained in:
Eric Biggers 2019-08-24 17:38:50 -07:00
parent ab6ba9d521
commit 2a2e24dc8b
3 changed files with 4 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
lib/deflate_compress.c
View File

@ -1857,7 +1857,7 @@ deflate_finish_sequence(struct deflate_sequence *seq, u32 litrunlen)
* but rather we combine many symbols into a single "observation type". For
* literals we only look at the high bits and low bits, and for matches we only
* look at whether the match is long or not. The assumption is that for typical
* "real" data, places that are good block boundaries will tend to be noticable
* "real" data, places that are good block boundaries will tend to be noticeable
* based only on changes in these aggregate frequencies, without looking for
* subtle differences in individual symbols. For example, a change from ASCII
* bytes to non-ASCII bytes, or from few matches (generally less compressible)

2
lib/deflate_decompress.c
View File

@ -336,7 +336,7 @@ do { \
* pointer", meaning that the entry references a subtable that must be indexed
* using more bits of the compressed data to decode the symbol.
*
* Each decode table (a main table along with with its subtables, if any) is
* Each decode table (a main table along with its subtables, if any) is
* associated with a Huffman code. Logically, the result of a decode table
* lookup is a symbol from the alphabet from which the corresponding Huffman
* code was constructed. A symbol with codeword length n <= TABLEBITS is

4
lib/x86/crc32_pclmul_template.h
View File

@ -92,8 +92,8 @@
* When <= 512 bits remain in the message, we finish up by folding across
* smaller distances. This works similarly; the distance D is just different,
* so different constant multipliers must be used. Finally, once the remaining
* message is just 64 bits, it is is reduced to the CRC-32 using Barrett
* reduction (explained later).
* message is just 64 bits, it is reduced to the CRC-32 using Barrett reduction
* (explained later).
*
* For more information see the original paper from Intel:
* "Fast CRC Computation for Generic Polynomials Using PCLMULQDQ Instruction"