Add a bad char heuristic to StringRef::find.

Based on Horspool's simplified version of Boyer-Moore. We use a constant-sized table of
uint8_ts to keep cache thrashing low, needles bigger than 255 bytes are uncommon anyways.

The worst case is still O(n*m) but we do a lot better on the average case now.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@142061 91177308-0d34-0410-b5e6-96231b3b80d8
This commit is contained in:
Benjamin Kramer
2011-10-15 10:08:31 +00:00
parent e9b58d0aac
commit 6e6a558ebc
2 changed files with 32 additions and 3 deletions

View File

@ -245,6 +245,12 @@ TEST(StringRefTest, Find) {
EXPECT_EQ(StringRef::npos, Str.find("zz"));
EXPECT_EQ(2U, Str.find("ll", 2));
EXPECT_EQ(StringRef::npos, Str.find("ll", 3));
EXPECT_EQ(0U, Str.find(""));
StringRef LongStr("hellx xello hell ello world foo bar hello");
EXPECT_EQ(36U, LongStr.find("hello"));
EXPECT_EQ(28U, LongStr.find("foo"));
EXPECT_EQ(12U, LongStr.find("hell", 2));
EXPECT_EQ(0U, LongStr.find(""));
EXPECT_EQ(3U, Str.rfind('l'));
EXPECT_EQ(StringRef::npos, Str.rfind('z'));