Add string support to Buffer.fill

This commit is contained in:
Jarred Sumner
2022-08-18 21:09:26 -07:00
parent d150a2f4dd
commit 0e4db0b28f
7 changed files with 119 additions and 77 deletions

View File

@@ -20,7 +20,7 @@ extern "C" JSC::EncodedJSValue Bun__encoding__toStringLatin1(const uint8_t* inpu
extern "C" JSC::EncodedJSValue Bun__encoding__toStringHex(const uint8_t* input, size_t len, JSC::JSGlobalObject* globalObject);
extern "C" JSC::EncodedJSValue Bun__encoding__toStringBase64(const uint8_t* input, size_t len, JSC::JSGlobalObject* globalObject);
extern "C" JSC::EncodedJSValue Bun__encoding__toStringURLSafeBase64(const uint8_t* input, size_t len, JSC::JSGlobalObject* globalObject);
extern "C" void Bun__Buffer_fill(JSC::JSGlobalObject*, Bun__ArrayBuffer*, ZigString*, uint32_t, uint32_t, WebCore::BufferEncodingType);
extern "C" void Bun__Buffer_fill(ZigString*, void*, size_t, WebCore::BufferEncodingType);
namespace WebCore {

View File

@@ -986,17 +986,15 @@ static inline JSC::EncodedJSValue jsBufferPrototypeFunction_fillBody(JSC::JSGlob
}
auto startPtr = castedThis->typedVector() + start;
auto str_ = value.toWTFString(lexicalGlobalObject);
ZigString str = Zig::toZigString(str_);
// ZigString str = Zig::toString(value.toString(lexicalGlobalObject));
// Bun__ArrayBuffer buf;
// JSC__JSValue__asArrayBuffer_(JSC::JSValue::encode(castedThis), lexicalGlobalObject,
// &buf);
// Bun__Buffer_fill(lexicalGlobalObject, &buf, &str, start, end, encoding);
Bun__Buffer_fill(&str, startPtr, end - start, encoding);
RELEASE_AND_RETURN(throwScope, JSValue::encode(castedThis));
}
}
static inline JSC::EncodedJSValue jsBufferPrototypeFunction_includesBody(JSC::JSGlobalObject* lexicalGlobalObject, JSC::CallFrame* callFrame, typename IDLOperation<JSBuffer>::ClassParameter castedThis)
{
auto& vm = JSC::getVM(lexicalGlobalObject);
@@ -1004,7 +1002,7 @@ static inline JSC::EncodedJSValue jsBufferPrototypeFunction_includesBody(JSC::JS
UNUSED_PARAM(throwScope);
UNUSED_PARAM(callFrame);
return JSC::JSValue::encode(JSC::JSValue(reinterpret_cast<uint8_t*>(castedThis->vector())[0]));
return jsBufferPrototypeFunction_indexOfBody(lexicalGlobalObject, callFrame, castedThis)
}
static inline JSC::EncodedJSValue jsBufferPrototypeFunction_indexOfBody(JSC::JSGlobalObject* lexicalGlobalObject, JSC::CallFrame* callFrame, typename IDLOperation<JSBuffer>::ClassParameter castedThis)
{
@@ -1031,6 +1029,7 @@ static inline JSC::EncodedJSValue jsBufferPrototypeFunction_swap64Body(JSC::JSGl
auto& vm = JSC::getVM(lexicalGlobalObject);
return JSC::JSValue::encode(jsUndefined());
}
static inline JSC::EncodedJSValue jsBufferPrototypeFunction_toStringBody(JSC::JSGlobalObject* lexicalGlobalObject, JSC::CallFrame* callFrame, typename IDLOperation<JSBuffer>::ClassParameter castedThis)
{
auto& vm = JSC::getVM(lexicalGlobalObject);

View File

@@ -2431,7 +2431,6 @@ pub const JSValue = enum(JSValueReprInt) {
.JSImmutableButterfly,
.JSSourceCode,
.JSScriptFetcher,
.InternalFunction,
.JSScriptFetchParameters,
=> true,
else => false,

View File

@@ -204,17 +204,6 @@ function writeBigUInt64BE(value, offset) {
return offset + 8;
}
function slice(start, end) {
"use strict";
if (start === undefined && end === undefined) {
return this;
}
Buffer[Symbol.species] ||= Buffer;
return new Buffer(this.buffer, this.byteOffset + (start || 0), (end || this.byteLength) - (start || 0));
}
function utf8Write(text, offset, length) {
"use strict";
return this.write(text, offset, length, "utf8");

View File

@@ -17,78 +17,85 @@ pub const BufferVectorized = struct {
extern fn memset_pattern16(b: *anyopaque, pattern16: *const anyopaque, len: usize) void;
pub fn fill(
globalObject: *JSGlobalObject,
this: *JSC.ArrayBuffer,
str: *JSC.ZigString,
start: u32,
end: u32,
buf_ptr: [*]u8,
fill_length: usize,
encoding: JSC.Node.Encoding,
) callconv(.C) void {
const allocator = JSC.VirtualMachine.vm.allocator;
var stack_fallback = std.heap.stackFallback(512, allocator);
var stack_fallback_allocator = stack_fallback.get();
var input_string = str.toSlice(stack_fallback_allocator);
if (input_string.len == 0) return;
if (str.len == 0) return;
defer input_string.deinit();
var buf = buf_ptr[0..fill_length];
var buf = this.slice()[start..end];
const written = switch (encoding) {
JSC.Node.Encoding.utf8 => if (str.is16Bit())
JSC.WebCore.Encoder.writeU16(str.utf16SliceAligned().ptr, str.utf16SliceAligned().len, buf.ptr, buf.len, JSC.Node.Encoding.utf8)
else
JSC.WebCore.Encoder.writeU8(str.slice().ptr, str.slice().len, buf.ptr, buf.len, JSC.Node.Encoding.utf8),
JSC.Node.Encoding.ascii => if (str.is16Bit())
JSC.WebCore.Encoder.writeU16(str.utf16SliceAligned().ptr, str.utf16SliceAligned().len, buf.ptr, buf.len, JSC.Node.Encoding.ascii)
else
JSC.WebCore.Encoder.writeU8(str.slice().ptr, str.slice().len, buf.ptr, buf.len, JSC.Node.Encoding.ascii),
JSC.Node.Encoding.latin1 => if (str.is16Bit())
JSC.WebCore.Encoder.writeU16(str.utf16SliceAligned().ptr, str.utf16SliceAligned().len, buf.ptr, buf.len, JSC.Node.Encoding.latin1)
else
JSC.WebCore.Encoder.writeU8(str.slice().ptr, str.slice().len, buf.ptr, buf.len, JSC.Node.Encoding.latin1),
JSC.Node.Encoding.buffer => if (str.is16Bit())
JSC.WebCore.Encoder.writeU16(str.utf16SliceAligned().ptr, str.utf16SliceAligned().len, buf.ptr, buf.len, JSC.Node.Encoding.buffer)
else
JSC.WebCore.Encoder.writeU8(str.slice().ptr, str.slice().len, buf.ptr, buf.len, JSC.Node.Encoding.buffer),
JSC.Node.Encoding.utf16le,
JSC.Node.Encoding.ucs2,
=> if (str.is16Bit())
JSC.WebCore.Encoder.writeU16(str.utf16SliceAligned().ptr, str.utf16SliceAligned().len, buf.ptr, buf.len, JSC.Node.Encoding.utf16le)
else
JSC.WebCore.Encoder.writeU8(str.slice().ptr, str.slice().len, buf.ptr, buf.len, JSC.Node.Encoding.utf16le),
JSC.Node.Encoding.base64 => if (str.is16Bit())
JSC.WebCore.Encoder.writeU16(str.utf16SliceAligned().ptr, str.utf16SliceAligned().len, buf.ptr, buf.len, JSC.Node.Encoding.base64)
else
JSC.WebCore.Encoder.writeU8(str.slice().ptr, str.slice().len, buf.ptr, buf.len, JSC.Node.Encoding.base64),
JSC.Node.Encoding.base64url => if (str.is16Bit())
JSC.WebCore.Encoder.writeU16(str.utf16SliceAligned().ptr, str.utf16SliceAligned().len, buf.ptr, buf.len, JSC.Node.Encoding.base64url)
else
JSC.WebCore.Encoder.writeU8(str.slice().ptr, str.slice().len, buf.ptr, buf.len, JSC.Node.Encoding.base64url),
JSC.Node.Encoding.hex => if (str.is16Bit())
JSC.WebCore.Encoder.writeU16(str.utf16SliceAligned().ptr, str.utf16SliceAligned().len, buf.ptr, buf.len, JSC.Node.Encoding.hex)
else
JSC.WebCore.Encoder.writeU8(str.slice().ptr, str.slice().len, buf.ptr, buf.len, JSC.Node.Encoding.hex),
};
var slice = input_string.slice();
switch (encoding) {
JSC.Node.Encoding.utf8,
JSC.Node.Encoding.ascii,
JSC.Node.Encoding.latin1,
JSC.Node.Encoding.buffer,
=> {
switch (slice.len) {
0 => unreachable,
1 => {
@memset(buf.ptr, slice[0], 1);
return;
},
2...16 => {
if (comptime Environment.isMac) {
var pattern: [16]u8 = undefined;
var remain: []u8 = pattern[0..];
if (written <= 0) {
return;
}
while (remain.len > 0) {
for (slice[0..]) |a| {
remain[0] = a;
remain = remain[1..];
}
}
var contents = buf[0..@intCast(usize, written)];
buf = buf[@intCast(usize, written)..];
memset_pattern16(buf.ptr, &pattern, buf.len);
return;
}
},
else => {},
}
if (contents.len == 1) {
@memset(buf.ptr, contents[0], buf.len);
return;
}
var in_there = @minimum(slice.len, buf.len);
@memcpy(buf.ptr, slice.ptr, in_there);
if (in_there < slice.len) {
return;
}
const minimum_contents = contents;
while (buf.len >= contents.len) {
const min_len = @minimum(contents.len, buf.len);
std.mem.copy(u8, buf[0..min_len], contents[0..min_len]);
if (buf.len <= contents.len) {
break;
}
buf = buf[min_len..];
contents.len *= 2;
}
// var ptr = buf.ptr + @as(usize, start) + slice.len;
// const fill_length = @as(usize, end) - @as(usize, start);
// // while (in_there < fill_length - in_there) {
// // std.mem.copy(ptr)
// // ptr += in_there;
// // in_there *= 2;
// // }
},
else => {},
while (buf.len > 0) {
const to_fill = @minimum(minimum_contents.len, buf.len);
std.mem.copy(u8, buf[0..to_fill], minimum_contents[0..to_fill]);
buf = buf[to_fill..];
}
}
};
comptime {
if (!JSC.is_bindgen) {
@export(BufferVectorized, .{ .name = "Bun__Buffer__fill" });
@export(BufferVectorized.fill, .{ .name = "Bun__Buffer_fill" });
}
}

View File

@@ -1462,6 +1462,8 @@ pub fn fixDeadCodeElimination() void {
std.mem.doNotOptimizeAway(&napi_ref_threadsafe_function);
std.mem.doNotOptimizeAway(&napi_add_async_cleanup_hook);
std.mem.doNotOptimizeAway(&napi_remove_async_cleanup_hook);
std.mem.doNotOptimizeAway(&@import("../bun.js/node/buffer.zig").BufferVectorized.fill);
}
comptime {
@@ -1552,5 +1554,6 @@ comptime {
_ = napi_ref_threadsafe_function;
_ = napi_add_async_cleanup_hook;
_ = napi_remove_async_cleanup_hook;
_ = @import("../bun.js/node/buffer.zig").BufferVectorized.fill;
}
}

View File

@@ -211,6 +211,51 @@ it("Buffer.copy", () => {
expect(array1.join("")).toBe(array2.join(""));
});
export function fillRepeating(dstBuffer, start, end) {
let len = dstBuffer.length, // important: use indices length, not byte-length
sLen = end - start,
p = sLen; // set initial position = source sequence length
// step 2: copy existing data doubling segment length per iteration
while (p < len) {
if (p + sLen > len) sLen = len - p; // if not power of 2, truncate last segment
dstBuffer.copyWithin(p, start, sLen); // internal copy
p += sLen; // add current length to offset
sLen <<= 1; // double length for next segment
}
}
describe("Buffer.fill string", () => {
for (let text of [
"hello world",
"1234567890",
"\uD83D\uDE00",
"😀😃😄😁😆😅😂🤣☺️😊😊😇",
]) {
it(text, () => {
var input = new Buffer(1024);
input.fill(text);
var demo = new Uint8Array(1024);
var encoded = new TextEncoder().encode(text);
demo.set(encoded);
fillRepeating(demo, 0, encoded.length);
expect(input.join("")).toBe(demo.join(""));
});
}
});
it("Buffer.fill 1 char string", () => {
var input = new Buffer(1024);
input.fill("h");
var demo = new Uint8Array(1024);
var encoded = new TextEncoder().encode("h");
demo.set(encoded);
fillRepeating(demo, 0, encoded.length);
expect(input.join("")).toBe(demo.join(""));
});
it("Buffer.concat", () => {
var array1 = new Uint8Array(128);
array1.fill(100);