From d42f7587e5972d7e787b5cf2a043c33fbe96091f Mon Sep 17 00:00:00 2001 From: "reed@google.com" Date: Tue, 23 Apr 2013 15:35:24 +0000 Subject: add block comment for using SkChecksum git-svn-id: http://skia.googlecode.com/svn/trunk@8827 2bbb7eff-a529-9590-31e7-b0007b416f81 --- include/core/SkChecksum.h | 9 +++++++++ 1 file changed, 9 insertions(+) (limited to 'include/core') diff --git a/include/core/SkChecksum.h b/include/core/SkChecksum.h index 9cb45c3307..287109fef9 100644 --- a/include/core/SkChecksum.h +++ b/include/core/SkChecksum.h @@ -10,6 +10,15 @@ #include "SkTypes.h" +/** + * Computes a 32bit checksum from a blob of 32bit aligned data. This is meant + * to be very very fast, as it is used internally by the font cache, in + * conjuction with the entire raw key. This algorithm does not generate + * unique values as well as others (e.g. MD5) but it performs much faster. + * Skia's use cases can survive non-unique values (since the entire key is + * always available). Clients should only be used in circumstances where speed + * over uniqueness is at a premium. + */ class SkChecksum : SkNoncopyable { private: /* -- cgit v1.2.3