| author |  Andrew Kelley <andrew@ziglang.org> 2026-06-25 16:28:27-07:00 |
| committer | Andrew Kelley <andrew@ziglang.org> 2026-06-25 16:28:27-07:00 |
| log | 4c5394b367562ccca37f66ef64cb2b3978c0c283 |
| tree | decce3d323bdab00a8ee196554ade52c79571851 |
| parent | 7db2ef6104921c42ce42e8edda848160eda8973d |
langref: make section for slice by length
3 files changed, 31 insertions(+), 10 deletions(-)
doc/langref.html.in+19-1
| ... | ... | @@ -1955,7 +1955,7 @@ or |
| 1955 | 1955 | TODO talk about C ABI interop<br> |
| 1956 | 1956 | TODO consider suggesting std.MultiArrayList |
| 1957 | 1957 | </p> |
| 1958 | {#see_also|@splat|@shuffle|@select|@reduce#} | |
| 1958 | {#see_also|@splat|@shuffle|@select|@reduce|Slicing by Length#} | |
| 1959 | 1959 | |
| 1960 | 1960 | {#header_open|Relationship with Arrays#} |
| 1961 | 1961 | <p>Vectors and {#link|Arrays#} each have a well-defined <strong>bit layout</strong> |
| ... | ... | @@ -2150,6 +2150,24 @@ or |
| 2150 | 2150 | |
| 2151 | 2151 | {#see_also|Pointers|for|Arrays#} |
| 2152 | 2152 | |
| 2153 | {#header_open|Slicing by Length#} | |
| 2154 | <p>Even though Zig only has syntax for slicing based on start and end indices, by slicing twice, | |
| 2155 | one can express a <strong>slice by length</strong> operation.</p> | |
| 2156 | <p>The pattern {#syntax#}[a .. a + b]{#endsyntax#} is always better expressed | |
| 2157 | {#syntax#}[a..][0..b]{#endsyntax#} because:</p> | |
| 2158 | <ul> | |
| 2159 | <li>Slices are represented in memory as a pointer and length. Despite | |
| 2160 | syntactically appearing as twice the work, it is actually one less | |
| 2161 | subtraction in machine code.</li> | |
| 2162 | <li>If {#syntax#}a{#endsyntax#} is known at runtime and | |
| 2163 | {#syntax#}b{#endsyntax#} is known at {#link|comptime#}, the former | |
| 2164 | results in a slice but the latter results in a single-item | |
| 2165 | {#link|pointer|Pointers#} to an {#link|array|Arrays#}, a generally more | |
| 2166 | safe type because the length is compile-time known.</li> | |
| 2167 | </ul> | |
| 2168 | {#code|slicing_by_length.zig#} | |
| 2169 | {#header_close#} | |
| 2170 | ||
| 2153 | 2171 | {#header_open|Sentinel-Terminated Slices#} |
| 2154 | 2172 | <p> |
| 2155 | 2173 | The syntax {#syntax#}[:x]T{#endsyntax#} is a slice which has a runtime-known length |
doc/langref/slicing_by_length.zig created+12
| ... | ... | @@ -0,0 +1,12 @@ |
| 1 | const expectEqual = @import("std").testing.expectEqual; | |
| 2 | ||
| 3 | test "example" { | |
| 4 | var array = [_]i32{ 1, 2, 3, 4 }; | |
| 5 | var runtime_start: usize = 1; | |
| 6 | _ = &runtime_start; | |
| 7 | const length = 2; | |
| 8 | const array_ptr_len = array[runtime_start..][0..length]; | |
| 9 | try expectEqual(*[length]i32, @TypeOf(array_ptr_len)); | |
| 10 | } | |
| 11 | ||
| 12 | // test |
doc/langref/test_basic_slices.zig-9
| ... | ... | @@ -21,15 +21,6 @@ test "basic slices" { |
| 21 | 21 | const array_ptr = array[0..array.len]; |
| 22 | 22 | try expectEqual(*[array.len]i32, @TypeOf(array_ptr)); |
| 23 | 23 | |
| 24 | // You can perform a slice-by-length by slicing twice. This allows the compiler | |
| 25 | // to perform some optimisations like recognising a comptime-known length when | |
| 26 | // the start position is only known at runtime. | |
| 27 | var runtime_start: usize = 1; | |
| 28 | _ = &runtime_start; | |
| 29 | const length = 2; | |
| 30 | const array_ptr_len = array[runtime_start..][0..length]; | |
| 31 | try expectEqual(*[length]i32, @TypeOf(array_ptr_len)); | |
| 32 | ||
| 33 | 24 | // Using the address-of operator on a slice gives a single-item pointer. |
| 34 | 25 | try expectEqual(*i32, @TypeOf(&slice[0])); |
| 35 | 26 | // Using the `ptr` field gives a many-item pointer. |