Add support for imports of UTF-16 encoded text w/ Byte Order Mark (BOM) (#314)

Before this PR, we blindly assumed that all text imported into speedscope was UTF-8 encoded. This, unsurprisingly, is not always true. After this PR, we support text that's UTF-16 encoded, with either the little-endian or big-endian byte-order-mark.

Fixed #291
This commit is contained in:
Jamie Wong
2020-09-29 15:40:49 -07:00
committed by GitHub
parent f3a1c09c9b
commit f758130455
7 changed files with 146 additions and 9 deletions
Binary file not shown.
Binary file not shown.
@@ -104,6 +104,114 @@ exports[`importFromBGFlameGraph with CRLF: indexToView 1`] = `0`;
exports[`importFromBGFlameGraph with CRLF: profileGroup.name 1`] = `"simple-crlf.txt"`;
exports[`importFromBGFlameGraph with UTF-16, Big Endian 1`] = `
Object {
"frames": Array [
Frame {
"col": undefined,
"file": undefined,
"key": "a",
"line": undefined,
"name": "a",
"selfWeight": 0,
"totalWeight": 14,
},
Frame {
"col": undefined,
"file": undefined,
"key": "b",
"line": undefined,
"name": "b",
"selfWeight": 5,
"totalWeight": 14,
},
Frame {
"col": undefined,
"file": undefined,
"key": "c",
"line": undefined,
"name": "c",
"selfWeight": 5,
"totalWeight": 5,
},
Frame {
"col": undefined,
"file": undefined,
"key": "d",
"line": undefined,
"name": "d",
"selfWeight": 4,
"totalWeight": 4,
},
],
"name": "simple-utf16-be.txt",
"stacks": Array [
"a;b;c 2",
"a;b;d 4",
"a;b;c 3",
"a;b 5",
],
}
`;
exports[`importFromBGFlameGraph with UTF-16, Big Endian: indexToView 1`] = `0`;
exports[`importFromBGFlameGraph with UTF-16, Big Endian: profileGroup.name 1`] = `"simple-utf16-be.txt"`;
exports[`importFromBGFlameGraph with UTF-16, Little Endian 1`] = `
Object {
"frames": Array [
Frame {
"col": undefined,
"file": undefined,
"key": "a",
"line": undefined,
"name": "a",
"selfWeight": 0,
"totalWeight": 14,
},
Frame {
"col": undefined,
"file": undefined,
"key": "b",
"line": undefined,
"name": "b",
"selfWeight": 5,
"totalWeight": 14,
},
Frame {
"col": undefined,
"file": undefined,
"key": "c",
"line": undefined,
"name": "c",
"selfWeight": 5,
"totalWeight": 5,
},
Frame {
"col": undefined,
"file": undefined,
"key": "d",
"line": undefined,
"name": "d",
"selfWeight": 4,
"totalWeight": 4,
},
],
"name": "simple-utf16-le.txt",
"stacks": Array [
"a;b;c 2",
"a;b;d 4",
"a;b;c 3",
"a;b 5",
],
}
`;
exports[`importFromBGFlameGraph with UTF-16, Little Endian: indexToView 1`] = `0`;
exports[`importFromBGFlameGraph with UTF-16, Little Endian: profileGroup.name 1`] = `"simple-utf16-le.txt"`;
exports[`importFromBGFlameGraph: indexToView 1`] = `0`;
exports[`importFromBGFlameGraph: profileGroup.name 1`] = `"simple.txt"`;
+2 -2
View File
@@ -76,7 +76,7 @@ Object {
"totalWeight": 0.0005174240213818848,
},
],
"name": "Grabación de Control temporal 1",
"name": "Grabación de Control temporal 1",
"stacks": Array [
"injectModule;(anonymous) 1.00ms",
" 39.93ms",
@@ -98,4 +98,4 @@ Object {
exports[`importFromSafari: indexToView 1`] = `0`;
exports[`importFromSafari: profileGroup.name 1`] = `"Grabación de Control temporal 1"`;
exports[`importFromSafari: profileGroup.name 1`] = `"Grabación de Control temporal 1"`;
+8
View File
@@ -7,3 +7,11 @@ test('importFromBGFlameGraph', async () => {
test('importFromBGFlameGraph with CRLF', async () => {
await checkProfileSnapshot('./sample/profiles/stackcollapse/simple-crlf.txt')
})
test('importFromBGFlameGraph with UTF-16, Little Endian', async () => {
await checkProfileSnapshot('./sample/profiles/stackcollapse/simple-utf16-le.txt')
})
test('importFromBGFlameGraph with UTF-16, Big Endian', async () => {
await checkProfileSnapshot('./sample/profiles/stackcollapse/simple-utf16-be.txt')
})
+20 -5
View File
@@ -51,15 +51,30 @@ export class MaybeCompressedDataReader implements ProfileDataSource {
async readAsText(): Promise<string> {
const buffer = await this.readAsArrayBuffer()
let ret: string = ''
// By default, we assume the file is utf-8 encoded.
let encoding = 'utf-8'
const array = new Uint8Array(buffer)
if (array.length > 2) {
if (array[0] === 0xff && array[1] === 0xfe) {
// UTF-16, Little Endian encoding
encoding = 'utf-16le'
} else if (array[0] === 0xfe && array[1] === 0xff) {
// UTF-16, Big Endian encoding
encoding = 'utf-16be'
}
}
if (typeof TextDecoder !== 'undefined') {
const decoder = new TextDecoder()
const decoder = new TextDecoder(encoding)
return decoder.decode(buffer)
} else {
// JavaScript strings are UTF-16 encoded, but we're reading data
// from disk that we're going to asusme is UTF-8 encoded.
const array = new Uint8Array(buffer)
// JavaScript strings are UTF-16 encoded, but we're reading data from disk
// that we're going to blindly assume it's ASCII encoded. This codepath
// only exists for older browser support.
console.warn('This browser does not support TextDecoder. Decoding text as ASCII.')
let ret: string = ''
for (let i = 0; i < array.length; i++) {
ret += String.fromCharCode(array[i])
}
+8 -2
View File
@@ -16,7 +16,7 @@
// Because we're not going to use this in our actual build, it's okay for this
// to be inefficient.
(function () {
;(function () {
const nodeVersion = process.versions.node
const versionParts = nodeVersion.split('.')
const majorVersion = parseInt(versionParts[0], 10)
@@ -48,4 +48,10 @@
this.splice(0, this.length, ...arrayWithIndices.map(x => x[0]))
return this
}
})()
})()
;(function () {
// TextDecoder is a global API in browsers, but an imported API in node.
//
// Let's emulate it being a global API during tests.
global.TextDecoder = require('util').TextDecoder
})()