| ... | ... | @@ -853,9 +853,7 @@ pub fn indexOf(comptime T: type, haystack: []const T, needle: []const T) ?usize |
| 853 | 853 | |
| 854 | 854 | /// Find the index in a slice of a sub-slice, searching from the end backwards. |
| 855 | 855 | /// To start looking at a different index, slice the haystack first. |
| 856 | | fn lastIndexOfNaive(comptime T: type, haystack: []const T, needle: []const T) ?usize { |
| 857 | | if (needle.len > haystack.len) return null; |
| 858 | | |
| 856 | fn lastIndexOfLinear(comptime T: type, haystack: []const T, needle: []const T) ?usize { |
| 859 | 857 | var i: usize = haystack.len - needle.len; |
| 860 | 858 | while (true) : (i -= 1) { |
| 861 | 859 | if (mem.eql(T, haystack[i .. i + needle.len], needle)) return i; |
| ... | ... | @@ -863,9 +861,7 @@ fn lastIndexOfNaive(comptime T: type, haystack: []const T, needle: []const T) ?u |
| 863 | 861 | } |
| 864 | 862 | } |
| 865 | 863 | |
| 866 | | fn indexOfPosNaive(comptime T: type, haystack: []const T, start_index: usize, needle: []const T) ?usize { |
| 867 | | if (needle.len > haystack.len) return null; |
| 868 | | |
| 864 | fn indexOfPosLinear(comptime T: type, haystack: []const T, start_index: usize, needle: []const T) ?usize { |
| 869 | 865 | var i: usize = start_index; |
| 870 | 866 | const end = haystack.len - needle.len; |
| 871 | 867 | while (i <= end) : (i += 1) { |
| ... | ... | @@ -874,6 +870,17 @@ fn indexOfPosNaive(comptime T: type, haystack: []const T, start_index: usize, ne |
| 874 | 870 | return null; |
| 875 | 871 | } |
| 876 | 872 | |
| 873 | fn boyerMooreHorspoolPreprocessReverse(pattern: []const u8, table: []usize) void { |
| 874 | for (table) |*c| { |
| 875 | c.* = pattern.len; |
| 876 | } |
| 877 | |
| 878 | var i: usize = pattern.len - 1; |
| 879 | while (i > 0) : (i -= 1) { |
| 880 | table[pattern[i]] = i; |
| 881 | } |
| 882 | } |
| 883 | |
| 877 | 884 | fn boyerMooreHorspoolPreprocess(pattern: []const u8, table: []usize) void { |
| 878 | 885 | for (table) |*c| { |
| 879 | 886 | c.* = pattern.len; |
| ... | ... | @@ -888,22 +895,23 @@ fn boyerMooreHorspoolPreprocess(pattern: []const u8, table: []usize) void { |
| 888 | 895 | /// To start looking at a different index, slice the haystack first. |
| 889 | 896 | // Reverse boyer-moore-horspool algorithm |
| 890 | 897 | pub fn lastIndexOf(comptime T: type, haystack: []const T, needle: []const T) ?usize { |
| 891 | | if (!isValidAlign(T.bit_count) or haystack.len < 32 or needle.len <= 2) |
| 892 | | return lastIndexOfNaive(T, haystack, needle); |
| 893 | | |
| 894 | 898 | if (needle.len > haystack.len or needle.len == 0) return null; |
| 895 | 899 | |
| 896 | | const haystackU8 = sliceAsBytes(haystack); |
| 897 | | const needleU8 = sliceAsBytes(needle); |
| 900 | if (!isValidAlign(T.bit_count) or haystack.len < 32 or needle.len <= 2) |
| 901 | return lastIndexOfLinear(T, haystack, needle); |
| 898 | 902 | |
| 899 | | var table: [256]usize = undefined; |
| 900 | | boyerMooreHorspoolPreprocess(needleU8, table[0..]); |
| 903 | const haystack_bytes = sliceAsBytes(haystack); |
| 904 | const needle_bytes = sliceAsBytes(needle); |
| 901 | 905 | |
| 902 | | var i: usize = 0; |
| 903 | | while (i <= haystackU8.len - needleU8.len) { |
| 904 | | const reverseIndex = haystackU8.len - i - needleU8.len - 1; |
| 905 | | if (mem.eql(u8, haystackU8[reverseIndex .. reverseIndex + needleU8.len], needleU8)) return i; |
| 906 | | i += table[haystackU8[reverseIndex + needleU8.len - 1]]; |
| 906 | var skip_table: [256]usize = undefined; |
| 907 | boyerMooreHorspoolPreprocessReverse(needle_bytes, skip_table[0..]); |
| 908 | |
| 909 | var i: usize = haystack_bytes.len - needle_bytes.len; |
| 910 | while (true) { |
| 911 | if (mem.eql(u8, haystack_bytes[i .. i + needle_bytes.len], needle_bytes)) return i; |
| 912 | const skip = skip_table[haystack_bytes[i]]; |
| 913 | if (skip > i) break; |
| 914 | i -= skip; |
| 907 | 915 | } |
| 908 | 916 | |
| 909 | 917 | return null; |
| ... | ... | @@ -911,27 +919,32 @@ pub fn lastIndexOf(comptime T: type, haystack: []const T, needle: []const T) ?us |
| 911 | 919 | |
| 912 | 920 | // Boyer-moore-horspool algorithm |
| 913 | 921 | pub fn indexOfPos(comptime T: type, haystack: []const T, start_index: usize, needle: []const T) ?usize { |
| 914 | | if (!isValidAlign(T.bit_count) or haystack.len < 32 or needle.len <= 2) |
| 915 | | return indexOfPosNaive(T, haystack, start_index, needle); |
| 916 | | |
| 917 | 922 | if (needle.len > haystack.len or needle.len == 0) return null; |
| 918 | 923 | |
| 919 | | const haystackU8 = sliceAsBytes(haystack); |
| 920 | | const needleU8 = sliceAsBytes(needle); |
| 924 | if (!isValidAlign(T.bit_count) or haystack.len < 32 or needle.len <= 2) |
| 925 | return indexOfPosLinear(T, haystack, start_index, needle); |
| 926 | |
| 927 | const haystack_bytes = sliceAsBytes(haystack); |
| 928 | const needle_bytes = sliceAsBytes(needle); |
| 921 | 929 | |
| 922 | | var table: [256]usize = undefined; |
| 923 | | boyerMooreHorspoolPreprocess(needleU8, table[0..]); |
| 930 | var skip_table: [256]usize = undefined; |
| 931 | boyerMooreHorspoolPreprocess(needle_bytes, skip_table[0..]); |
| 924 | 932 | |
| 925 | 933 | var i: usize = start_index; |
| 926 | | while (i <= haystackU8.len - needleU8.len) { |
| 927 | | if (mem.eql(u8, haystackU8[i .. i + needleU8.len], needleU8)) return i; |
| 928 | | i += table[haystackU8[i + needleU8.len - 1]]; |
| 934 | while (i <= haystack_bytes.len - needle_bytes.len) { |
| 935 | if (mem.eql(u8, haystack_bytes[i .. i + needle_bytes.len], needle_bytes)) return i; |
| 936 | i += skip_table[haystack_bytes[i + needle_bytes.len - 1]]; |
| 929 | 937 | } |
| 930 | 938 | |
| 931 | 939 | return null; |
| 932 | 940 | } |
| 933 | 941 | |
| 934 | 942 | test "mem.indexOf" { |
| 943 | testing.expect(indexOf(u8, "one two three four five six seven eight nine ten", "three four").? == 8); |
| 944 | testing.expect(lastIndexOf(u8, "one two three four five six seven eight nine ten", "three four").? == 8); |
| 945 | testing.expect(indexOf(u8, "one two three four five six seven eight nine ten", "two two") == null); |
| 946 | testing.expect(lastIndexOf(u8, "one two three four five six seven eight nine ten", "two two") == null); |
| 947 | |
| 935 | 948 | testing.expect(indexOf(u8, "one two three four", "four").? == 14); |
| 936 | 949 | testing.expect(lastIndexOf(u8, "one two three two four", "two").? == 14); |
| 937 | 950 | testing.expect(indexOf(u8, "one two three four", "gour") == null); |