Typo and grammar fixes
This commit is contained in:
@@ -14,7 +14,7 @@
|
||||
* files would be small relative to the size of the file.
|
||||
*
|
||||
* Various 'diffing' algorithms utilize this notion of edit distance and
|
||||
* the corrensponding concept of a minimal edit script between two
|
||||
* the corresponding concept of a minimal edit script between two
|
||||
* sequences to identify the regions within two files where they differ.
|
||||
* The core algorithm used in this match finder is described in:
|
||||
*
|
||||
@@ -28,12 +28,12 @@
|
||||
*
|
||||
* Note: after some experimentation, this approach proved to not provide enough
|
||||
* utility to justify the additional CPU used in finding matches. The one area
|
||||
* where this approach consistenly outperforms Zstandard even on level 19 is
|
||||
* when compressing small files (<10 KB) using a equally small dictionary that
|
||||
* where this approach consistently outperforms Zstandard even on level 19 is
|
||||
* when compressing small files (<10 KB) using an equally small dictionary that
|
||||
* is very similar to the source file. For the use case that this was intended,
|
||||
* (large similar files) this approach by itself took 5-10X longer than zstd-19 and
|
||||
* generally resulted in 2-3X larger files. The core advantage that zstd-19 has
|
||||
* over this appraoch for match finding is the overlapping matches. This approach
|
||||
* over this approach for match finding is the overlapping matches. This approach
|
||||
* cannot find any.
|
||||
*
|
||||
* I'm leaving this in the contrib section in case this ever becomes interesting
|
||||
|
||||
@@ -12,7 +12,7 @@
|
||||
* Dependencies
|
||||
***************************************/
|
||||
|
||||
/* Currently relies on qsort when combining contiguous matches. This can probabily
|
||||
/* Currently relies on qsort when combining contiguous matches. This can probably
|
||||
* be avoided but would require changes to the algorithm. The qsort is far from
|
||||
* the bottleneck in this algorithm even for medium sized files so it's probably
|
||||
* not worth trying to address */
|
||||
@@ -26,7 +26,7 @@
|
||||
* Constants
|
||||
***************************************/
|
||||
|
||||
/* Just a sential for the entires of the diagnomal matrix */
|
||||
/* Just a sential for the entires of the diagonal matrix */
|
||||
#define ZSTD_EDIST_DIAG_MAX (S32)(1 << 30)
|
||||
|
||||
/* How large should a snake be to be considered a 'big' snake.
|
||||
@@ -39,7 +39,7 @@
|
||||
#define ZSTD_EDIST_SNAKE_ITER_THRESH 200
|
||||
|
||||
/* After how many iterations should be just give up and take
|
||||
* the best availabe edit script for this round */
|
||||
* the best available edit script for this round */
|
||||
#define ZSTD_EDIST_EXPENSIVE_THRESH 1024
|
||||
|
||||
/*-*************************************
|
||||
@@ -192,7 +192,7 @@ static void ZSTD_eDist_diag(ZSTD_eDist_state* state,
|
||||
if (!useHeuristics)
|
||||
continue;
|
||||
|
||||
/* Everything under this point is a heuritic. Using these will
|
||||
/* Everything under this point is a heuristic. Using these will
|
||||
* substantially speed up the match finding. In some cases, taking
|
||||
* the total match finding time from several minutes to seconds.
|
||||
* Of course, the caveat is that the edit script found may no longer
|
||||
@@ -366,8 +366,8 @@ static int ZSTD_eDist_compare(ZSTD_eDist_state* state,
|
||||
}
|
||||
} else if (srcLow == srcHigh) {
|
||||
while (dictLow < dictHigh) {
|
||||
/* Reaching this point means deleteing dict[dictLow] from
|
||||
* the current positino of dict */
|
||||
/* Reaching this point means deleting dict[dictLow] from
|
||||
* the current position of dict */
|
||||
dictLow++;
|
||||
}
|
||||
} else {
|
||||
@@ -395,7 +395,7 @@ static int ZSTD_eDist_matchComp(const void* p, const void* q)
|
||||
}
|
||||
|
||||
/* The matches from the approach above will all be of the form
|
||||
* (dictIdx, srcIdx, 1). this method combines contiguous matches
|
||||
* (dictIdx, srcIdx, 1). This method combines contiguous matches
|
||||
* of length MINMATCH or greater. Matches less than MINMATCH
|
||||
* are discarded */
|
||||
static void ZSTD_eDist_combineMatches(ZSTD_eDist_state* state)
|
||||
|
||||
@@ -21,7 +21,7 @@
|
||||
* files would be small relative to the size of the file.
|
||||
*
|
||||
* Various 'diffing' algorithms utilize this notion of edit distance and
|
||||
* the corrensponding concept of a minimal edit script between two
|
||||
* the corresponding concept of a minimal edit script between two
|
||||
* sequences to identify the regions within two files where they differ.
|
||||
* The core algorithm used in this match finder is described in:
|
||||
*
|
||||
@@ -35,12 +35,12 @@
|
||||
*
|
||||
* Note: after some experimentation, this approach proved to not provide enough
|
||||
* utility to justify the additional CPU used in finding matches. The one area
|
||||
* where this approach consistenly outperforms Zstandard even on level 19 is
|
||||
* when compressing small files (<10 KB) using a equally small dictionary that
|
||||
* where this approach consistently outperforms Zstandard even on level 19 is
|
||||
* when compressing small files (<10 KB) using an equally small dictionary that
|
||||
* is very similar to the source file. For the use case that this was intended,
|
||||
* (large similar files) this approach by itself took 5-10X longer than zstd-19 and
|
||||
* generally resulted in 2-3X larger files. The core advantage that zstd-19 has
|
||||
* over this appraoch for match finding is the overlapping matches. This approach
|
||||
* over this approach for match finding is the overlapping matches. This approach
|
||||
* cannot find any.
|
||||
*
|
||||
* I'm leaving this in the contrib section in case this ever becomes interesting
|
||||
|
||||
Reference in New Issue
Block a user