Add benchmark script (#7869)
The script `benchmark.js` will collect page load metrics from the
extension, and print them to a file or the console. A method for
collecting metrics was added to the web driver to help with this.
This script will calculate the min, max, average, and standard
deviation for four metrics: 'firstPaint', 'domContentLoaded', 'load',
and 'domInteractive'. The variation between samples is sometimes high,
with the results varying between samples if only 3 were taken. However,
all tests I've done locally with 5 samples have produced results within
one standard deviation of each other. The default number of samples has
been set to 10, which should be more than enough to produce consistent
results.
The benchmark can be run with the npm script `benchmark:chrome` or
`benchmark:firefox`, e.g. `yarn benchmark:chrome`.
2020-01-21 17:02:45 +01:00
|
|
|
#!/usr/bin/env node
|
|
|
|
|
2021-02-04 19:15:23 +01:00
|
|
|
const path = require('path');
|
|
|
|
const { promises: fs, constants: fsConstants } = require('fs');
|
2021-06-17 23:06:53 +02:00
|
|
|
const yargs = require('yargs/yargs');
|
|
|
|
const { hideBin } = require('yargs/helpers');
|
2021-02-04 19:15:23 +01:00
|
|
|
const ttest = require('ttest');
|
2021-06-21 17:16:18 +02:00
|
|
|
const { retry } = require('../../development/lib/retry');
|
|
|
|
const { exitWithError } = require('../../development/lib/exit-with-error');
|
2021-02-04 19:15:23 +01:00
|
|
|
const { withFixtures } = require('./helpers');
|
|
|
|
const { PAGES } = require('./webdriver/driver');
|
Add benchmark script (#7869)
The script `benchmark.js` will collect page load metrics from the
extension, and print them to a file or the console. A method for
collecting metrics was added to the web driver to help with this.
This script will calculate the min, max, average, and standard
deviation for four metrics: 'firstPaint', 'domContentLoaded', 'load',
and 'domInteractive'. The variation between samples is sometimes high,
with the results varying between samples if only 3 were taken. However,
all tests I've done locally with 5 samples have produced results within
one standard deviation of each other. The default number of samples has
been set to 10, which should be more than enough to produce consistent
results.
The benchmark can be run with the npm script `benchmark:chrome` or
`benchmark:firefox`, e.g. `yarn benchmark:chrome`.
2020-01-21 17:02:45 +01:00
|
|
|
|
2021-02-04 19:15:23 +01:00
|
|
|
const DEFAULT_NUM_SAMPLES = 20;
|
|
|
|
const ALL_PAGES = Object.values(PAGES);
|
Add benchmark script (#7869)
The script `benchmark.js` will collect page load metrics from the
extension, and print them to a file or the console. A method for
collecting metrics was added to the web driver to help with this.
This script will calculate the min, max, average, and standard
deviation for four metrics: 'firstPaint', 'domContentLoaded', 'load',
and 'domInteractive'. The variation between samples is sometimes high,
with the results varying between samples if only 3 were taken. However,
all tests I've done locally with 5 samples have produced results within
one standard deviation of each other. The default number of samples has
been set to 10, which should be more than enough to produce consistent
results.
The benchmark can be run with the npm script `benchmark:chrome` or
`benchmark:firefox`, e.g. `yarn benchmark:chrome`.
2020-01-21 17:02:45 +01:00
|
|
|
|
2020-11-03 00:41:28 +01:00
|
|
|
async function measurePage(pageName) {
|
2021-02-04 19:15:23 +01:00
|
|
|
let metrics;
|
Add benchmark script (#7869)
The script `benchmark.js` will collect page load metrics from the
extension, and print them to a file or the console. A method for
collecting metrics was added to the web driver to help with this.
This script will calculate the min, max, average, and standard
deviation for four metrics: 'firstPaint', 'domContentLoaded', 'load',
and 'domInteractive'. The variation between samples is sometimes high,
with the results varying between samples if only 3 were taken. However,
all tests I've done locally with 5 samples have produced results within
one standard deviation of each other. The default number of samples has
been set to 10, which should be more than enough to produce consistent
results.
The benchmark can be run with the npm script `benchmark:chrome` or
`benchmark:firefox`, e.g. `yarn benchmark:chrome`.
2020-01-21 17:02:45 +01:00
|
|
|
await withFixtures({ fixtures: 'imported-account' }, async ({ driver }) => {
|
2021-02-04 19:15:23 +01:00
|
|
|
await driver.navigate();
|
2021-04-12 17:32:38 +02:00
|
|
|
await driver.fill('#password', 'correct horse battery staple');
|
|
|
|
await driver.press('#password', driver.Key.ENTER);
|
2021-04-07 16:57:40 +02:00
|
|
|
await driver.findElement('.selected-account__name');
|
2021-02-04 19:15:23 +01:00
|
|
|
await driver.navigate(pageName);
|
|
|
|
await driver.delay(1000);
|
|
|
|
metrics = await driver.collectMetrics();
|
|
|
|
});
|
|
|
|
return metrics;
|
Add benchmark script (#7869)
The script `benchmark.js` will collect page load metrics from the
extension, and print them to a file or the console. A method for
collecting metrics was added to the web driver to help with this.
This script will calculate the min, max, average, and standard
deviation for four metrics: 'firstPaint', 'domContentLoaded', 'load',
and 'domInteractive'. The variation between samples is sometimes high,
with the results varying between samples if only 3 were taken. However,
all tests I've done locally with 5 samples have produced results within
one standard deviation of each other. The default number of samples has
been set to 10, which should be more than enough to produce consistent
results.
The benchmark can be run with the npm script `benchmark:chrome` or
`benchmark:firefox`, e.g. `yarn benchmark:chrome`.
2020-01-21 17:02:45 +01:00
|
|
|
}
|
|
|
|
|
2020-11-03 00:41:28 +01:00
|
|
|
function calculateResult(calc) {
|
Add benchmark script (#7869)
The script `benchmark.js` will collect page load metrics from the
extension, and print them to a file or the console. A method for
collecting metrics was added to the web driver to help with this.
This script will calculate the min, max, average, and standard
deviation for four metrics: 'firstPaint', 'domContentLoaded', 'load',
and 'domInteractive'. The variation between samples is sometimes high,
with the results varying between samples if only 3 were taken. However,
all tests I've done locally with 5 samples have produced results within
one standard deviation of each other. The default number of samples has
been set to 10, which should be more than enough to produce consistent
results.
The benchmark can be run with the npm script `benchmark:chrome` or
`benchmark:firefox`, e.g. `yarn benchmark:chrome`.
2020-01-21 17:02:45 +01:00
|
|
|
return (result) => {
|
2021-02-04 19:15:23 +01:00
|
|
|
const calculatedResult = {};
|
Add benchmark script (#7869)
The script `benchmark.js` will collect page load metrics from the
extension, and print them to a file or the console. A method for
collecting metrics was added to the web driver to help with this.
This script will calculate the min, max, average, and standard
deviation for four metrics: 'firstPaint', 'domContentLoaded', 'load',
and 'domInteractive'. The variation between samples is sometimes high,
with the results varying between samples if only 3 were taken. However,
all tests I've done locally with 5 samples have produced results within
one standard deviation of each other. The default number of samples has
been set to 10, which should be more than enough to produce consistent
results.
The benchmark can be run with the npm script `benchmark:chrome` or
`benchmark:firefox`, e.g. `yarn benchmark:chrome`.
2020-01-21 17:02:45 +01:00
|
|
|
for (const key of Object.keys(result)) {
|
2021-02-04 19:15:23 +01:00
|
|
|
calculatedResult[key] = calc(result[key]);
|
Add benchmark script (#7869)
The script `benchmark.js` will collect page load metrics from the
extension, and print them to a file or the console. A method for
collecting metrics was added to the web driver to help with this.
This script will calculate the min, max, average, and standard
deviation for four metrics: 'firstPaint', 'domContentLoaded', 'load',
and 'domInteractive'. The variation between samples is sometimes high,
with the results varying between samples if only 3 were taken. However,
all tests I've done locally with 5 samples have produced results within
one standard deviation of each other. The default number of samples has
been set to 10, which should be more than enough to produce consistent
results.
The benchmark can be run with the npm script `benchmark:chrome` or
`benchmark:firefox`, e.g. `yarn benchmark:chrome`.
2020-01-21 17:02:45 +01:00
|
|
|
}
|
2021-02-04 19:15:23 +01:00
|
|
|
return calculatedResult;
|
|
|
|
};
|
Add benchmark script (#7869)
The script `benchmark.js` will collect page load metrics from the
extension, and print them to a file or the console. A method for
collecting metrics was added to the web driver to help with this.
This script will calculate the min, max, average, and standard
deviation for four metrics: 'firstPaint', 'domContentLoaded', 'load',
and 'domInteractive'. The variation between samples is sometimes high,
with the results varying between samples if only 3 were taken. However,
all tests I've done locally with 5 samples have produced results within
one standard deviation of each other. The default number of samples has
been set to 10, which should be more than enough to produce consistent
results.
The benchmark can be run with the npm script `benchmark:chrome` or
`benchmark:firefox`, e.g. `yarn benchmark:chrome`.
2020-01-21 17:02:45 +01:00
|
|
|
}
|
2021-02-04 19:15:23 +01:00
|
|
|
const calculateSum = (array) => array.reduce((sum, val) => sum + val);
|
|
|
|
const calculateAverage = (array) => calculateSum(array) / array.length;
|
|
|
|
const minResult = calculateResult((array) => Math.min(...array));
|
|
|
|
const maxResult = calculateResult((array) => Math.max(...array));
|
|
|
|
const averageResult = calculateResult((array) => calculateAverage(array));
|
Add benchmark script (#7869)
The script `benchmark.js` will collect page load metrics from the
extension, and print them to a file or the console. A method for
collecting metrics was added to the web driver to help with this.
This script will calculate the min, max, average, and standard
deviation for four metrics: 'firstPaint', 'domContentLoaded', 'load',
and 'domInteractive'. The variation between samples is sometimes high,
with the results varying between samples if only 3 were taken. However,
all tests I've done locally with 5 samples have produced results within
one standard deviation of each other. The default number of samples has
been set to 10, which should be more than enough to produce consistent
results.
The benchmark can be run with the npm script `benchmark:chrome` or
`benchmark:firefox`, e.g. `yarn benchmark:chrome`.
2020-01-21 17:02:45 +01:00
|
|
|
const standardDeviationResult = calculateResult((array) => {
|
2021-06-16 17:23:51 +02:00
|
|
|
if (array.length === 1) {
|
|
|
|
return 0;
|
|
|
|
}
|
2021-02-04 19:15:23 +01:00
|
|
|
const average = calculateAverage(array);
|
|
|
|
const squareDiffs = array.map((value) => Math.pow(value - average, 2));
|
|
|
|
return Math.sqrt(calculateAverage(squareDiffs));
|
|
|
|
});
|
2020-07-20 19:02:49 +02:00
|
|
|
// 95% margin of error calculated using Student's t-distribution
|
2020-11-03 00:41:28 +01:00
|
|
|
const calculateMarginOfError = (array) =>
|
2021-02-04 19:15:23 +01:00
|
|
|
ttest(array).confidence()[1] - calculateAverage(array);
|
2020-11-03 00:41:28 +01:00
|
|
|
const marginOfErrorResult = calculateResult((array) =>
|
2021-06-16 17:23:51 +02:00
|
|
|
array.length === 1 ? 0 : calculateMarginOfError(array),
|
2021-02-04 19:15:23 +01:00
|
|
|
);
|
Add benchmark script (#7869)
The script `benchmark.js` will collect page load metrics from the
extension, and print them to a file or the console. A method for
collecting metrics was added to the web driver to help with this.
This script will calculate the min, max, average, and standard
deviation for four metrics: 'firstPaint', 'domContentLoaded', 'load',
and 'domInteractive'. The variation between samples is sometimes high,
with the results varying between samples if only 3 were taken. However,
all tests I've done locally with 5 samples have produced results within
one standard deviation of each other. The default number of samples has
been set to 10, which should be more than enough to produce consistent
results.
The benchmark can be run with the npm script `benchmark:chrome` or
`benchmark:firefox`, e.g. `yarn benchmark:chrome`.
2020-01-21 17:02:45 +01:00
|
|
|
|
2021-06-21 17:16:18 +02:00
|
|
|
async function profilePageLoad(pages, numSamples, retries) {
|
2021-02-04 19:15:23 +01:00
|
|
|
const results = {};
|
Add benchmark script (#7869)
The script `benchmark.js` will collect page load metrics from the
extension, and print them to a file or the console. A method for
collecting metrics was added to the web driver to help with this.
This script will calculate the min, max, average, and standard
deviation for four metrics: 'firstPaint', 'domContentLoaded', 'load',
and 'domInteractive'. The variation between samples is sometimes high,
with the results varying between samples if only 3 were taken. However,
all tests I've done locally with 5 samples have produced results within
one standard deviation of each other. The default number of samples has
been set to 10, which should be more than enough to produce consistent
results.
The benchmark can be run with the npm script `benchmark:chrome` or
`benchmark:firefox`, e.g. `yarn benchmark:chrome`.
2020-01-21 17:02:45 +01:00
|
|
|
for (const pageName of pages) {
|
2021-02-04 19:15:23 +01:00
|
|
|
const runResults = [];
|
Add benchmark script (#7869)
The script `benchmark.js` will collect page load metrics from the
extension, and print them to a file or the console. A method for
collecting metrics was added to the web driver to help with this.
This script will calculate the min, max, average, and standard
deviation for four metrics: 'firstPaint', 'domContentLoaded', 'load',
and 'domInteractive'. The variation between samples is sometimes high,
with the results varying between samples if only 3 were taken. However,
all tests I've done locally with 5 samples have produced results within
one standard deviation of each other. The default number of samples has
been set to 10, which should be more than enough to produce consistent
results.
The benchmark can be run with the npm script `benchmark:chrome` or
`benchmark:firefox`, e.g. `yarn benchmark:chrome`.
2020-01-21 17:02:45 +01:00
|
|
|
for (let i = 0; i < numSamples; i += 1) {
|
2021-06-21 17:16:18 +02:00
|
|
|
let result;
|
|
|
|
await retry(retries, async () => {
|
|
|
|
result = await measurePage(pageName);
|
|
|
|
});
|
|
|
|
runResults.push(result);
|
Add benchmark script (#7869)
The script `benchmark.js` will collect page load metrics from the
extension, and print them to a file or the console. A method for
collecting metrics was added to the web driver to help with this.
This script will calculate the min, max, average, and standard
deviation for four metrics: 'firstPaint', 'domContentLoaded', 'load',
and 'domInteractive'. The variation between samples is sometimes high,
with the results varying between samples if only 3 were taken. However,
all tests I've done locally with 5 samples have produced results within
one standard deviation of each other. The default number of samples has
been set to 10, which should be more than enough to produce consistent
results.
The benchmark can be run with the npm script `benchmark:chrome` or
`benchmark:firefox`, e.g. `yarn benchmark:chrome`.
2020-01-21 17:02:45 +01:00
|
|
|
}
|
|
|
|
|
2020-02-15 21:34:12 +01:00
|
|
|
if (runResults.some((result) => result.navigation.lenth > 1)) {
|
2021-02-04 19:15:23 +01:00
|
|
|
throw new Error(`Multiple navigations not supported`);
|
2020-11-03 00:41:28 +01:00
|
|
|
} else if (
|
|
|
|
runResults.some((result) => result.navigation[0].type !== 'navigate')
|
|
|
|
) {
|
|
|
|
throw new Error(
|
|
|
|
`Navigation type ${
|
|
|
|
runResults.find((result) => result.navigation[0].type !== 'navigate')
|
|
|
|
.navigation[0].type
|
|
|
|
} not supported`,
|
2021-02-04 19:15:23 +01:00
|
|
|
);
|
Add benchmark script (#7869)
The script `benchmark.js` will collect page load metrics from the
extension, and print them to a file or the console. A method for
collecting metrics was added to the web driver to help with this.
This script will calculate the min, max, average, and standard
deviation for four metrics: 'firstPaint', 'domContentLoaded', 'load',
and 'domInteractive'. The variation between samples is sometimes high,
with the results varying between samples if only 3 were taken. However,
all tests I've done locally with 5 samples have produced results within
one standard deviation of each other. The default number of samples has
been set to 10, which should be more than enough to produce consistent
results.
The benchmark can be run with the npm script `benchmark:chrome` or
`benchmark:firefox`, e.g. `yarn benchmark:chrome`.
2020-01-21 17:02:45 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
const result = {
|
2020-08-18 18:36:45 +02:00
|
|
|
firstPaint: runResults.map((metrics) => metrics.paint['first-paint']),
|
2020-11-03 00:41:28 +01:00
|
|
|
domContentLoaded: runResults.map(
|
|
|
|
(metrics) =>
|
|
|
|
metrics.navigation[0] && metrics.navigation[0].domContentLoaded,
|
|
|
|
),
|
|
|
|
load: runResults.map(
|
|
|
|
(metrics) => metrics.navigation[0] && metrics.navigation[0].load,
|
|
|
|
),
|
|
|
|
domInteractive: runResults.map(
|
|
|
|
(metrics) =>
|
|
|
|
metrics.navigation[0] && metrics.navigation[0].domInteractive,
|
|
|
|
),
|
2021-02-04 19:15:23 +01:00
|
|
|
};
|
Add benchmark script (#7869)
The script `benchmark.js` will collect page load metrics from the
extension, and print them to a file or the console. A method for
collecting metrics was added to the web driver to help with this.
This script will calculate the min, max, average, and standard
deviation for four metrics: 'firstPaint', 'domContentLoaded', 'load',
and 'domInteractive'. The variation between samples is sometimes high,
with the results varying between samples if only 3 were taken. However,
all tests I've done locally with 5 samples have produced results within
one standard deviation of each other. The default number of samples has
been set to 10, which should be more than enough to produce consistent
results.
The benchmark can be run with the npm script `benchmark:chrome` or
`benchmark:firefox`, e.g. `yarn benchmark:chrome`.
2020-01-21 17:02:45 +01:00
|
|
|
|
|
|
|
results[pageName] = {
|
|
|
|
min: minResult(result),
|
|
|
|
max: maxResult(result),
|
|
|
|
average: averageResult(result),
|
|
|
|
standardDeviation: standardDeviationResult(result),
|
2020-01-21 22:12:40 +01:00
|
|
|
marginOfError: marginOfErrorResult(result),
|
2021-02-04 19:15:23 +01:00
|
|
|
};
|
Add benchmark script (#7869)
The script `benchmark.js` will collect page load metrics from the
extension, and print them to a file or the console. A method for
collecting metrics was added to the web driver to help with this.
This script will calculate the min, max, average, and standard
deviation for four metrics: 'firstPaint', 'domContentLoaded', 'load',
and 'domInteractive'. The variation between samples is sometimes high,
with the results varying between samples if only 3 were taken. However,
all tests I've done locally with 5 samples have produced results within
one standard deviation of each other. The default number of samples has
been set to 10, which should be more than enough to produce consistent
results.
The benchmark can be run with the npm script `benchmark:chrome` or
`benchmark:firefox`, e.g. `yarn benchmark:chrome`.
2020-01-21 17:02:45 +01:00
|
|
|
}
|
2021-02-04 19:15:23 +01:00
|
|
|
return results;
|
Add benchmark script (#7869)
The script `benchmark.js` will collect page load metrics from the
extension, and print them to a file or the console. A method for
collecting metrics was added to the web driver to help with this.
This script will calculate the min, max, average, and standard
deviation for four metrics: 'firstPaint', 'domContentLoaded', 'load',
and 'domInteractive'. The variation between samples is sometimes high,
with the results varying between samples if only 3 were taken. However,
all tests I've done locally with 5 samples have produced results within
one standard deviation of each other. The default number of samples has
been set to 10, which should be more than enough to produce consistent
results.
The benchmark can be run with the npm script `benchmark:chrome` or
`benchmark:firefox`, e.g. `yarn benchmark:chrome`.
2020-01-21 17:02:45 +01:00
|
|
|
}
|
|
|
|
|
2020-11-03 00:41:28 +01:00
|
|
|
async function isWritable(directory) {
|
Add benchmark script (#7869)
The script `benchmark.js` will collect page load metrics from the
extension, and print them to a file or the console. A method for
collecting metrics was added to the web driver to help with this.
This script will calculate the min, max, average, and standard
deviation for four metrics: 'firstPaint', 'domContentLoaded', 'load',
and 'domInteractive'. The variation between samples is sometimes high,
with the results varying between samples if only 3 were taken. However,
all tests I've done locally with 5 samples have produced results within
one standard deviation of each other. The default number of samples has
been set to 10, which should be more than enough to produce consistent
results.
The benchmark can be run with the npm script `benchmark:chrome` or
`benchmark:firefox`, e.g. `yarn benchmark:chrome`.
2020-01-21 17:02:45 +01:00
|
|
|
try {
|
2021-02-04 19:15:23 +01:00
|
|
|
await fs.access(directory, fsConstants.W_OK);
|
|
|
|
return true;
|
Add benchmark script (#7869)
The script `benchmark.js` will collect page load metrics from the
extension, and print them to a file or the console. A method for
collecting metrics was added to the web driver to help with this.
This script will calculate the min, max, average, and standard
deviation for four metrics: 'firstPaint', 'domContentLoaded', 'load',
and 'domInteractive'. The variation between samples is sometimes high,
with the results varying between samples if only 3 were taken. However,
all tests I've done locally with 5 samples have produced results within
one standard deviation of each other. The default number of samples has
been set to 10, which should be more than enough to produce consistent
results.
The benchmark can be run with the npm script `benchmark:chrome` or
`benchmark:firefox`, e.g. `yarn benchmark:chrome`.
2020-01-21 17:02:45 +01:00
|
|
|
} catch (error) {
|
|
|
|
if (error.code !== 'EACCES') {
|
2021-02-04 19:15:23 +01:00
|
|
|
throw error;
|
Add benchmark script (#7869)
The script `benchmark.js` will collect page load metrics from the
extension, and print them to a file or the console. A method for
collecting metrics was added to the web driver to help with this.
This script will calculate the min, max, average, and standard
deviation for four metrics: 'firstPaint', 'domContentLoaded', 'load',
and 'domInteractive'. The variation between samples is sometimes high,
with the results varying between samples if only 3 were taken. However,
all tests I've done locally with 5 samples have produced results within
one standard deviation of each other. The default number of samples has
been set to 10, which should be more than enough to produce consistent
results.
The benchmark can be run with the npm script `benchmark:chrome` or
`benchmark:firefox`, e.g. `yarn benchmark:chrome`.
2020-01-21 17:02:45 +01:00
|
|
|
}
|
2021-02-04 19:15:23 +01:00
|
|
|
return false;
|
Add benchmark script (#7869)
The script `benchmark.js` will collect page load metrics from the
extension, and print them to a file or the console. A method for
collecting metrics was added to the web driver to help with this.
This script will calculate the min, max, average, and standard
deviation for four metrics: 'firstPaint', 'domContentLoaded', 'load',
and 'domInteractive'. The variation between samples is sometimes high,
with the results varying between samples if only 3 were taken. However,
all tests I've done locally with 5 samples have produced results within
one standard deviation of each other. The default number of samples has
been set to 10, which should be more than enough to produce consistent
results.
The benchmark can be run with the npm script `benchmark:chrome` or
`benchmark:firefox`, e.g. `yarn benchmark:chrome`.
2020-01-21 17:02:45 +01:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2020-11-03 00:41:28 +01:00
|
|
|
async function getFirstParentDirectoryThatExists(directory) {
|
2021-02-04 19:15:23 +01:00
|
|
|
let nextDirectory = directory;
|
2020-08-13 02:39:17 +02:00
|
|
|
for (;;) {
|
Add benchmark script (#7869)
The script `benchmark.js` will collect page load metrics from the
extension, and print them to a file or the console. A method for
collecting metrics was added to the web driver to help with this.
This script will calculate the min, max, average, and standard
deviation for four metrics: 'firstPaint', 'domContentLoaded', 'load',
and 'domInteractive'. The variation between samples is sometimes high,
with the results varying between samples if only 3 were taken. However,
all tests I've done locally with 5 samples have produced results within
one standard deviation of each other. The default number of samples has
been set to 10, which should be more than enough to produce consistent
results.
The benchmark can be run with the npm script `benchmark:chrome` or
`benchmark:firefox`, e.g. `yarn benchmark:chrome`.
2020-01-21 17:02:45 +01:00
|
|
|
try {
|
2021-02-04 19:15:23 +01:00
|
|
|
await fs.access(nextDirectory, fsConstants.F_OK);
|
|
|
|
return nextDirectory;
|
Add benchmark script (#7869)
The script `benchmark.js` will collect page load metrics from the
extension, and print them to a file or the console. A method for
collecting metrics was added to the web driver to help with this.
This script will calculate the min, max, average, and standard
deviation for four metrics: 'firstPaint', 'domContentLoaded', 'load',
and 'domInteractive'. The variation between samples is sometimes high,
with the results varying between samples if only 3 were taken. However,
all tests I've done locally with 5 samples have produced results within
one standard deviation of each other. The default number of samples has
been set to 10, which should be more than enough to produce consistent
results.
The benchmark can be run with the npm script `benchmark:chrome` or
`benchmark:firefox`, e.g. `yarn benchmark:chrome`.
2020-01-21 17:02:45 +01:00
|
|
|
} catch (error) {
|
|
|
|
if (error.code !== 'ENOENT') {
|
2021-02-04 19:15:23 +01:00
|
|
|
throw error;
|
2020-08-15 13:58:11 +02:00
|
|
|
} else if (nextDirectory === path.dirname(nextDirectory)) {
|
2021-02-04 19:15:23 +01:00
|
|
|
throw new Error('Failed to find parent directory that exists');
|
Add benchmark script (#7869)
The script `benchmark.js` will collect page load metrics from the
extension, and print them to a file or the console. A method for
collecting metrics was added to the web driver to help with this.
This script will calculate the min, max, average, and standard
deviation for four metrics: 'firstPaint', 'domContentLoaded', 'load',
and 'domInteractive'. The variation between samples is sometimes high,
with the results varying between samples if only 3 were taken. However,
all tests I've done locally with 5 samples have produced results within
one standard deviation of each other. The default number of samples has
been set to 10, which should be more than enough to produce consistent
results.
The benchmark can be run with the npm script `benchmark:chrome` or
`benchmark:firefox`, e.g. `yarn benchmark:chrome`.
2020-01-21 17:02:45 +01:00
|
|
|
}
|
2021-02-04 19:15:23 +01:00
|
|
|
nextDirectory = path.dirname(nextDirectory);
|
Add benchmark script (#7869)
The script `benchmark.js` will collect page load metrics from the
extension, and print them to a file or the console. A method for
collecting metrics was added to the web driver to help with this.
This script will calculate the min, max, average, and standard
deviation for four metrics: 'firstPaint', 'domContentLoaded', 'load',
and 'domInteractive'. The variation between samples is sometimes high,
with the results varying between samples if only 3 were taken. However,
all tests I've done locally with 5 samples have produced results within
one standard deviation of each other. The default number of samples has
been set to 10, which should be more than enough to produce consistent
results.
The benchmark can be run with the npm script `benchmark:chrome` or
`benchmark:firefox`, e.g. `yarn benchmark:chrome`.
2020-01-21 17:02:45 +01:00
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2020-11-03 00:41:28 +01:00
|
|
|
async function main() {
|
2021-06-17 23:06:53 +02:00
|
|
|
const { argv } = yargs(hideBin(process.argv)).usage(
|
|
|
|
'$0 [options]',
|
|
|
|
'Run a page load benchmark',
|
|
|
|
(_yargs) =>
|
|
|
|
_yargs
|
|
|
|
.option('pages', {
|
|
|
|
array: true,
|
|
|
|
default: ['home'],
|
|
|
|
description:
|
|
|
|
'Set the page(s) to be benchmarked. This flag can accept multiple values (space-separated).',
|
|
|
|
choices: ALL_PAGES,
|
|
|
|
})
|
|
|
|
.option('samples', {
|
|
|
|
default: DEFAULT_NUM_SAMPLES,
|
|
|
|
description: 'The number of times the benchmark should be run.',
|
|
|
|
type: 'number',
|
|
|
|
})
|
|
|
|
.option('out', {
|
|
|
|
description:
|
|
|
|
'Output filename. Output printed to STDOUT of this is omitted.',
|
|
|
|
type: 'string',
|
|
|
|
normalize: true,
|
2021-06-21 17:16:18 +02:00
|
|
|
})
|
|
|
|
.option('retries', {
|
|
|
|
default: 0,
|
|
|
|
description:
|
|
|
|
'Set how many times each benchmark sample should be retried upon failure.',
|
|
|
|
type: 'number',
|
2021-06-17 23:06:53 +02:00
|
|
|
}),
|
|
|
|
);
|
|
|
|
|
2021-06-21 17:16:18 +02:00
|
|
|
const { pages, samples, out, retries } = argv;
|
Add benchmark script (#7869)
The script `benchmark.js` will collect page load metrics from the
extension, and print them to a file or the console. A method for
collecting metrics was added to the web driver to help with this.
This script will calculate the min, max, average, and standard
deviation for four metrics: 'firstPaint', 'domContentLoaded', 'load',
and 'domInteractive'. The variation between samples is sometimes high,
with the results varying between samples if only 3 were taken. However,
all tests I've done locally with 5 samples have produced results within
one standard deviation of each other. The default number of samples has
been set to 10, which should be more than enough to produce consistent
results.
The benchmark can be run with the npm script `benchmark:chrome` or
`benchmark:firefox`, e.g. `yarn benchmark:chrome`.
2020-01-21 17:02:45 +01:00
|
|
|
|
2021-02-04 19:15:23 +01:00
|
|
|
let outputDirectory;
|
|
|
|
let existingParentDirectory;
|
2021-06-17 23:06:53 +02:00
|
|
|
if (out) {
|
|
|
|
outputDirectory = path.dirname(out);
|
|
|
|
existingParentDirectory = await getFirstParentDirectoryThatExists(
|
|
|
|
outputDirectory,
|
|
|
|
);
|
|
|
|
if (!(await isWritable(existingParentDirectory))) {
|
|
|
|
throw new Error('Specified output file directory is not writable');
|
Add benchmark script (#7869)
The script `benchmark.js` will collect page load metrics from the
extension, and print them to a file or the console. A method for
collecting metrics was added to the web driver to help with this.
This script will calculate the min, max, average, and standard
deviation for four metrics: 'firstPaint', 'domContentLoaded', 'load',
and 'domInteractive'. The variation between samples is sometimes high,
with the results varying between samples if only 3 were taken. However,
all tests I've done locally with 5 samples have produced results within
one standard deviation of each other. The default number of samples has
been set to 10, which should be more than enough to produce consistent
results.
The benchmark can be run with the npm script `benchmark:chrome` or
`benchmark:firefox`, e.g. `yarn benchmark:chrome`.
2020-01-21 17:02:45 +01:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2021-06-21 17:16:18 +02:00
|
|
|
const results = await profilePageLoad(pages, samples, retries);
|
Add benchmark script (#7869)
The script `benchmark.js` will collect page load metrics from the
extension, and print them to a file or the console. A method for
collecting metrics was added to the web driver to help with this.
This script will calculate the min, max, average, and standard
deviation for four metrics: 'firstPaint', 'domContentLoaded', 'load',
and 'domInteractive'. The variation between samples is sometimes high,
with the results varying between samples if only 3 were taken. However,
all tests I've done locally with 5 samples have produced results within
one standard deviation of each other. The default number of samples has
been set to 10, which should be more than enough to produce consistent
results.
The benchmark can be run with the npm script `benchmark:chrome` or
`benchmark:firefox`, e.g. `yarn benchmark:chrome`.
2020-01-21 17:02:45 +01:00
|
|
|
|
2021-06-17 23:06:53 +02:00
|
|
|
if (out) {
|
Add benchmark script (#7869)
The script `benchmark.js` will collect page load metrics from the
extension, and print them to a file or the console. A method for
collecting metrics was added to the web driver to help with this.
This script will calculate the min, max, average, and standard
deviation for four metrics: 'firstPaint', 'domContentLoaded', 'load',
and 'domInteractive'. The variation between samples is sometimes high,
with the results varying between samples if only 3 were taken. However,
all tests I've done locally with 5 samples have produced results within
one standard deviation of each other. The default number of samples has
been set to 10, which should be more than enough to produce consistent
results.
The benchmark can be run with the npm script `benchmark:chrome` or
`benchmark:firefox`, e.g. `yarn benchmark:chrome`.
2020-01-21 17:02:45 +01:00
|
|
|
if (outputDirectory !== existingParentDirectory) {
|
2021-02-04 19:15:23 +01:00
|
|
|
await fs.mkdir(outputDirectory, { recursive: true });
|
Add benchmark script (#7869)
The script `benchmark.js` will collect page load metrics from the
extension, and print them to a file or the console. A method for
collecting metrics was added to the web driver to help with this.
This script will calculate the min, max, average, and standard
deviation for four metrics: 'firstPaint', 'domContentLoaded', 'load',
and 'domInteractive'. The variation between samples is sometimes high,
with the results varying between samples if only 3 were taken. However,
all tests I've done locally with 5 samples have produced results within
one standard deviation of each other. The default number of samples has
been set to 10, which should be more than enough to produce consistent
results.
The benchmark can be run with the npm script `benchmark:chrome` or
`benchmark:firefox`, e.g. `yarn benchmark:chrome`.
2020-01-21 17:02:45 +01:00
|
|
|
}
|
2021-06-17 23:06:53 +02:00
|
|
|
await fs.writeFile(out, JSON.stringify(results, null, 2));
|
Add benchmark script (#7869)
The script `benchmark.js` will collect page load metrics from the
extension, and print them to a file or the console. A method for
collecting metrics was added to the web driver to help with this.
This script will calculate the min, max, average, and standard
deviation for four metrics: 'firstPaint', 'domContentLoaded', 'load',
and 'domInteractive'. The variation between samples is sometimes high,
with the results varying between samples if only 3 were taken. However,
all tests I've done locally with 5 samples have produced results within
one standard deviation of each other. The default number of samples has
been set to 10, which should be more than enough to produce consistent
results.
The benchmark can be run with the npm script `benchmark:chrome` or
`benchmark:firefox`, e.g. `yarn benchmark:chrome`.
2020-01-21 17:02:45 +01:00
|
|
|
} else {
|
2021-02-04 19:15:23 +01:00
|
|
|
console.log(JSON.stringify(results, null, 2));
|
Add benchmark script (#7869)
The script `benchmark.js` will collect page load metrics from the
extension, and print them to a file or the console. A method for
collecting metrics was added to the web driver to help with this.
This script will calculate the min, max, average, and standard
deviation for four metrics: 'firstPaint', 'domContentLoaded', 'load',
and 'domInteractive'. The variation between samples is sometimes high,
with the results varying between samples if only 3 were taken. However,
all tests I've done locally with 5 samples have produced results within
one standard deviation of each other. The default number of samples has
been set to 10, which should be more than enough to produce consistent
results.
The benchmark can be run with the npm script `benchmark:chrome` or
`benchmark:firefox`, e.g. `yarn benchmark:chrome`.
2020-01-21 17:02:45 +01:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2021-06-21 17:16:18 +02:00
|
|
|
main().catch((error) => {
|
|
|
|
exitWithError(error);
|
2021-02-04 19:15:23 +01:00
|
|
|
});
|