@matheusbsbs / matheus-bagliano / commits / af3abef94a
Fork 0

diffcore-delta.c: update the comment on the algorithm.

The comment at the top of the file described an old algorithm that was neutral to text/binary differences (it hashed sliding window of N-byte sequences and counted overlaps), but long time ago we switched to a new heuristics that are more suitable for line oriented (read: text) files that are much faster. Signed-off-by: Junio C Hamano <gitster@pobox.com>

Junio C Hamano committed Jun 28, 2007 at 23:11 UTC af3abef94af9c821a0c192c693c3e5342ab8729f
1 file changed +9 -12
diffcore-delta.c
+9 -12
@@ -5,23 +5,20 @@
5 /*
6 * Idea here is very simple.
7 *
8 - * We have total of (sz-N+1) N-byte overlapping sequences in buf whose
9 - * size is sz. If the same N-byte sequence appears in both source and
10 - * destination, we say the byte that starts that sequence is shared
11 - * between them (i.e. copied from source to destination).
8 + * Almost all data we are interested in are text, but sometimes we have
9 + * to deal with binary data. So we cut them into chunks delimited by
10 + * LF byte, or 64-byte sequence, whichever comes first, and hash them.
11 *
13 - * For each possible N-byte sequence, if the source buffer has more
14 - * instances of it than the destination buffer, that means the
15 - * difference are the number of bytes not copied from source to
16 - * destination. If the counts are the same, everything was copied
17 - * from source to destination. If the destination has more,
18 - * everything was copied, and destination added more.
12 + * For those chunks, if the source buffer has more instances of it
13 + * than the destination buffer, that means the difference are the
14 + * number of bytes not copied from source to destination. If the
15 + * counts are the same, everything was copied from source to
16 + * destination. If the destination has more, everything was copied,
17 + * and destination added more.
18 *
19 * We are doing an approximation so we do not really have to waste
20 * memory by actually storing the sequence. We just hash them into
21 * somewhere around 2^16 hashbuckets and count the occurrences.
23 - *
24 - * The length of the sequence is arbitrarily set to 8 for now.
22 */
23
24 /* Wild guess at the initial hash size */