Add bit_array.split, rework tests for split_once

Author: ackurat

CHANGELOG.md

@@ -2,6 +2,7 @@
 
 ## Unreleased
 
+- The `bit_array` module gains the `split` and `split_once` functions.
 - The deprecated `drop_left`, `drop_right`, `pad_left`, `pad_right`,
   `trim_left`, and `trim_right` functions have been removed.
 - Fixed a bug that would result in `list.unique` having quadratic runtime.

src/gleam/bit_array.gleam

@@ -64,11 +64,16 @@ pub fn slice(
 /// The result will not include the pattern, and returns an error if the
 /// pattern is not found.
 /// 
+/// This function runs in linear time.
+///
 /// ## Examples
 /// 
 /// ```gleam
 /// split_once(from: <<1, 2, 3>>, on: <<2>>)
 /// // -> Ok(#(<<1>>, <<3>>))
+/// 
+/// split_once(from: <<0>>, on: <<1>>)
+/// // -> Error(Nil)
 /// ```
 @external(erlang, "gleam_stdlib", "bit_array_split_once")
 @external(javascript, "../gleam_stdlib.mjs", "bit_array_split_once")
@@ -77,6 +82,29 @@ pub fn split_once(
   on pattern: BitArray,
 ) -> Result(#(BitArray, BitArray), Nil)
 
+/// Splits a bit array into parts at the locations of the pattern.
+///
+/// The result will not include the pattern, and returns an empty
+/// list if the pattern is not found.
+///
+/// This function runs in linear time.
+///
+/// ## Examples
+///
+/// ```gleam
+/// split(from: <<0, 1, 0, 2, 0, 3>>, on: <<0>>)
+/// // -> Ok([<<1>>, <<2>>, <<3>>])
+/// 
+/// split(from: <<0>>, on: <<1>>)
+/// // -> Ok([])
+/// ```
+@external(erlang, "gleam_stdlib", "bit_array_split")
+@external(javascript, "../gleam_stdlib.mjs", "bit_array_split")
+pub fn split(
+  from bits: BitArray,
+  on pattern: BitArray,
+) -> Result(List(BitArray), Nil)
+
 /// Tests to see whether a bit array is valid UTF-8.
 ///
 pub fn is_utf8(bits: BitArray) -> Bool {

src/gleam_stdlib.erl

@@ -14,7 +14,8 @@
     inspect/1, float_to_string/1, int_from_base_string/2,
     utf_codepoint_list_to_string/1, contains_string/2, crop_string/2,
     base16_encode/1, base16_decode/1, string_replace/3, slice/3,
-    bit_array_to_int_and_size/1, bit_array_pad_to_bytes/1, bit_array_split_once/2
+    bit_array_to_int_and_size/1, bit_array_pad_to_bytes/1, bit_array_split_once/2,
+    bit_array_split/2
 ]).
 
 %% Taken from OTP's uri_string module
@@ -235,12 +236,17 @@ bit_array_split_once(Bin, Sub) ->
     try
         case binary:split(Bin, [Sub]) of
             [<<>>, <<>>] -> {error, nil};
-            [Part1, Part2] -> {ok, {Part1, Part2}};
+            [A, B] -> {ok, {A, B}};
             _ -> {error, nil}
         end
     catch error:badarg -> {error, nil}
     end.
 
+bit_array_split(Bin, Sub) ->
+    try {ok, binary:split(Bin, [Sub], [global, trim_all])}
+    catch error:badarg -> {error, nil}
+    end.
+
 base_decode64(S) ->
     try {ok, base64:decode(S)}
     catch error:_ -> {error, nil}

src/gleam_stdlib.mjs

@@ -427,16 +427,19 @@ export function bit_array_slice(bits, position, length) {
 
 export function bit_array_split_once(bits, pattern) {
   try {
-    if (!(bits instanceof BitArray) || !(pattern instanceof BitArray) || pattern.buffer.length < 1 || pattern.buffer.length >= bits.buffer.length) {
+    if (!(bits instanceof BitArray)
+      || !(pattern instanceof BitArray)
+      || pattern.buffer.length < 1
+      || pattern.buffer.length >= bits.buffer.length) {
       return new Error(Nil);
     }
 
-    let i = 0;
     const n = bits.buffer.length - pattern.buffer.length + 1;
-
-    find: for (; i < n; i++) {
+    find: for (let i = 0; i < n; i++) {
       for (let j = 0; j < pattern.buffer.length; j++) {
-        if (bits.buffer[i + j] !== pattern.buffer[j]) continue find;
+        if (bits.buffer[i + j] !== pattern.buffer[j]) {
+          continue find;
+        }
       }
       const before = bits.buffer.slice(0, i);
       const after = bits.buffer.slice(i + pattern.buffer.length);
@@ -449,6 +452,58 @@ export function bit_array_split_once(bits, pattern) {
   }
 }
 
+export function bit_array_split(bits, pattern) {
+  try {
+    if (!(bits instanceof BitArray)
+      || !(pattern instanceof BitArray)
+      || pattern.buffer.length < 1) {
+      return new Error(Nil);
+    }
+
+    if (bits.buffer.length === 0) {
+      return new Ok(List.fromArray([]));
+    }
+
+    if (bits.buffer.length === pattern.buffer.length) {
+      let isEqual = true;
+      for (let i = 0; i < bits.buffer.length; i++) {
+        if (bits.buffer[i] !== pattern.buffer[i]) {
+          isEqual = false;
+          break;
+        }
+      }
+      if (isEqual) {
+        return new Ok(List.fromArray([]));
+      }
+    }
+
+    const results = [];
+    let lastIndex = 0;
+    const n = bits.buffer.length - pattern.buffer.length + 1;
+
+    find: for (let i = 0; i < n; i++) {
+      for (let j = 0; j < pattern.buffer.length; j++) {
+        if (bits.buffer[i + j] !== pattern.buffer[j]) {
+          continue find;
+        }
+      }
+      if (i > lastIndex) {
+        results.push(new BitArray(bits.buffer.slice(lastIndex, i)));
+      }
+      lastIndex = i + pattern.buffer.length;
+      i = lastIndex - 1;
+    }
+
+    if (lastIndex < bits.buffer.length) {
+      results.push(new BitArray(bits.buffer.slice(lastIndex)));
+    }
+
+    return new Ok(List.fromArray(results.length ? results : [bits]));
+  } catch (e) {
+    return new Error(Nil);
+  }
+}
+
 export function codepoint(int) {
   return new UtfCodepoint(int);
 }

test/gleam/bit_array_test.gleam

@@ -215,16 +215,24 @@ pub fn split_once_test() {
   |> bit_array.split_once(<<"h":utf8>>)
   |> should.equal(Ok(#(<<>>, <<"ello":utf8>>)))
 
-  <<"hello":utf8>>
+  <<0, 1, 0, 2, 0, 3>>
+  |> bit_array.split_once(<<0, 2>>)
+  |> should.equal(Ok(#(<<0, 1>>, <<0, 3>>)))
+
+  <<0, 1, 2, 0, 3, 4, 5>>
+  |> bit_array.split_once(<<>>)
+  |> should.equal(Error(Nil))
+
+  <<>>
   |> bit_array.split_once(<<1>>)
   |> should.equal(Error(Nil))
 
-  <<"hello":utf8>>
-  |> bit_array.split_once(<<"":utf8>>)
+  <<1>>
+  |> bit_array.split_once(<<1>>)
   |> should.equal(Error(Nil))
 
-  <<"hello":utf8>>
-  |> bit_array.split_once(<<"hello":utf8>>)
+  <<0>>
+  |> bit_array.split_once(<<1>>)
   |> should.equal(Error(Nil))
 }
 
@@ -237,6 +245,49 @@ pub fn split_once_erlang_only_test() {
   |> should.equal(Error(Nil))
 }
 
+pub fn split_test() {
+  <<"hello":utf8>>
+  |> bit_array.split(<<"l":utf8>>)
+  |> should.equal(Ok([<<"he":utf8>>, <<"o":utf8>>]))
+
+  <<0, 1, 0, 2, 0, 3>>
+  |> bit_array.split(<<0>>)
+  |> should.equal(Ok([<<1>>, <<2>>, <<3>>]))
+
+  <<1, 0>>
+  |> bit_array.split(<<0>>)
+  |> should.equal(Ok([<<1>>]))
+
+  <<0, 1, 0, 2, 0, 3>>
+  |> bit_array.split(<<0, 2>>)
+  |> should.equal(Ok([<<0, 1>>, <<0, 3>>]))
+
+  <<1>>
+  |> bit_array.split(<<0>>)
+  |> should.equal(Ok([<<1>>]))
+
+  <<1>>
+  |> bit_array.split(<<1>>)
+  |> should.equal(Ok([]))
+
+  <<>>
+  |> bit_array.split(<<1>>)
+  |> should.equal(Ok([]))
+
+  <<0, 1, 2, 0, 3, 4, 5>>
+  |> bit_array.split(<<>>)
+  |> should.equal(Error(Nil))
+}
+
+// This test is target specific since it's using non byte-aligned BitArrays
+// and those are not supported on the JavaScript target.
+@target(erlang)
+pub fn split_erlang_only_test() {
+  <<0, 1, 2:7>>
+  |> bit_array.split(<<1>>)
+  |> should.equal(Error(Nil))
+}
+
 pub fn to_string_test() {
   <<>>
   |> bit_array.to_string