• Update to faststat 0.0.7 #4347.

• Update dev dependencies #4345.

• Let analysisstorer know about webpagereplay so that correct information is displayed #4349.

• Pass on webpagereplay flag to browsertime #4350.

• Update to Browsertime 23.5.0 #4351

• Added --userTimingBlockList #2227.

Fixed

• Updated to webdriver 4.28 #2226.

• Updated dev dependencies #2228.

• Fix so the check for gnirehtet always use --android.gnirehtet to fix problens setting it in sitespeed.io #2229.

• Update to execa 9.5.2 #2231.

• Update to adbkit 3.3.8 #2232.

• Fix so you cannot use gnirehtet with webpagereplay #2230.

• Updated to Browsertime 23.4.5 that have a fix for visual metrics running videos from mobile #4343.

• Change bottom margin from 10 to 14 pixel when recoring a video on mobile, se bug #2224 and PR #2225.

• Add validation for dataset lengths for the Wilcoxon test. Thank you Pavel Bairov for PR 4341.

• Updated to Browsertime 23.4.4 that has a more robust check for gnirehtet configuration #4342.

• A more robust gnirehtet check when you run your Android test #2223.

• Fix killing WebPageReplay correctly and correct path to WebPageReplay files #108.

• Fix broken path to the WebPageReplay script when running as npm package and fix so you also can run desktop test using WebPageReplay when you use npm #107.

• Added extra logs if the compare plugin fails when using Python so we easily can see the input parameters.

• Updated to Browsertime 23.4.3 #4336.

• Make sure Firefox displays power consumption in mWH #4334.

• Update to USB power profiling 1.5.0 #2221.

• Improve how we compute the power consumption numbers from the Firefox Profiler output PR by Nazım Can Altınova #2220

• Fixed the release flow so the Docker containers get the correct tag.

• A new Docker setup #95

• Fixed the release flow so the Docker containers get the correct tag.

• A new Docker setup #95

• Firefox 133 in the Docker container #4327.

Fixed

• Axe Core 4.10.2 #3328

• Fix security warnings in the Docker container.

• Make sure the compare plugin can handle if your run Firefox with CPU options #4324.

• Updated to Firefox 132 and Edge 131 in the Docketr container. The performance regression in Firefox 132 #2211 has been fixed.

• Browsertime has been updated to 23.4.1. There's a couple of notable additions:

• Cleanup of the extra timings that is added to the timeline for Chrome #2218

• Collect more information about the largest contentful paint #2206

• Add extra timings and new custom timeline to the Chrome timeline #2217. Inspired by [

Fixed

• More configuration tuning for Firefox #2216.

• Disable search suggestion requests for Firefox #2215

• Fix for including HTTP requests in the Firefox profile #2214.

• Fix when running --enableProfileRun and get visual metrics. It used to generate an error log, this fixes that #2213.

• Disable some calls back to home from Firefox #2212.

• Disable security warning in Firefox 132 #2208 that fixes #2207.

• Upgraded Chrome/Chromedriver to 131 (with a new Browsertime) #4316.

• It seems to be a problem with Firefox 132 inside of Docker. Lets return to 131 for some time #4313.

• Try/catch broken dates #130.

• Removed dayjs dependency #128.

• Removed uuid dependency #127.

• Updated to tough cookie 5.0.0 #126.

• Bull 4.16.3 #87

• Express 5.0.0 #86.

• Update body parser #88.

• Bull 4.16.3 #87

• Device id was broken in the front end #80 and #81.

• The internal configuration was broken in the way that removing/adding testrunners failed removing correct configurations #84. This fix also needs #83 in the testrunners.

• Smarter ifb0 handling on Linux #93.

• Fix bug so you can choose emulated mobile in GUI #76.

• Correctly set the hostname in the configuration file #83.

• Fix a bug setting the baseline for compare plugin #71.

• Correct version in package.json

• When running compare plugin tests using Docker, set the baseline directory automatically #70

• • Make it possible to choose container that runs the test, using the API #69

• Merged slug/label column in search result #68

• Make it possible to choose container that runs the test, using the API #69

• Log testrunner version on startup #67.

• Upgrade to Bull 4.15.1 #57.

• Default to max 50 finished jobs in the queue #58.

• Update dev dependencies like ESLint 9 #62.

• Update DayJS 1.11.11 #63

• Make sure all uncaught errors are caught #64

• Add a keydb configuration file #56

• Always take care of Network.responseReceivedExtraInfo even if its fired early, thank you ifduyue for PR #120.

• Include early hints requests, thank you Ben Schwarz for PR #118.

• If the queue system is down when you add a new test, make sure the status of the test in the database is failed #50.

• If the queue is not up when you access the start page, make sure that at least after X seconds you get a error page #53.

• search: Make label configurable in search result #46.

• search: Make run date readable #45.

• Show version number in footer and API #44.

• Search: Fix for searching for test type #42.

• Added a safer header check #125.

• If the meta section from a HAR file from Browsertime was missing, we didn't catch that. Fixed in #124.

• Update to latest minimist #122.

• Another update to PageXray with a safer HTTP header check.

• Updated PageXray.

• Expose third party web and wappalyzer versions #111.

Fixed

• Updated to Wappalyzer 6.10.66 #110.

• Updated third party web 0.24.0 #108.

• Fix correct LCP attribute #107.

Changed

• Remove advice about disableFLoCHeader #106.

• Updated tough-cookie, thank you lallenlowe for PR #107.

• Catch if local storage or session storage isn't accessible.

• Upgraded to third party web 0.23.0. #98.

• Upgraded to latest wappalyzer #100.

Fixed

• Make sure to skip advice if Chrome is missing FCP #95.

• Update dev dependencies

• Updated dependencies to latest versions #105.

• Remove adding headers from extra parameters. Using the CDP Fetch API manipulating responses, the extra response info holds the old un-manipulated headers, so when copying the ones that do not exist, this breaks. Fixed in #100.

• Fixed exporting to work with ESM modules too.

• Dropped support for NodeJS 12 since it's unmaintaned.

• Fixed --version so it actually shows the version again #75.

• The E6 Module convertion in 4.0 broke version check (--version) making installs failed. PR #74 is a temporaty fix for that.

• Removed the dependecy of route. Using ip route instead and we already have ip as a requirement.

• Added missing SSL timings to the time per domain #117

• Upgraded to PageXray 4.4.2 and Third Party Web 0.17.1.

• Updated minimist dependency.

• Updated dev dependencies to latest versions #68

Changed

• Converting the project to ES6 module. If you run Throttle by command line it will work exactly as before except that we dropped support for NodeJS 10 #67. You can read Sindre Sorhus ESM package guide.

• Upgraded to PageXray 4.4.1 and Third Party Web 0.15.0.

• Bumped minimist to 1.2.6

• Upgraded to PageXray 4.0.0.

• Upgraded to third party web 0.13.0.

• Add _isLinkPreload on the request object if it is isLinkPreload #94.

• Fallback if mime types aren't mapped and use file endings. At the moment we fallback for some of the font types #109.

• Add support for setting packet loss. Add --packetLoss and set the loss in percentage.

• Updated third-party-web to 0.12.6.

• Use Chrome(ium) render blocking information to know if a request is render blocking or not #66.

• Report offending JavaScript assets if the JavaScript max limits kicks in #70.

• New largest contentful paint advice #76.

• New first contentful paint advice #77.

• Added TBT in the CPU longtask advice #80.

• Report content and transfer size for offending URLs #81.

• Report offending assets with transfer/content size for page size limit #82.

Fixed

• Fix cases when JQuery is undefined. Thank you shubham jajodia for PR #64.

• A better way to find offending layout shifters. Thank you shubham jajodia for PR #65.

• Removed mentions aboout server push #69

• Added more information on how to debug CPU advice #71.

Changed

• Moved AMP advice to best practice instead of privacy #67.

• Increased favicon max size advice from 5 to 10 kb #68

• Renamed the fastRender advice to avoidRenderBlocking #73

• Remove the third party async advice #74

• Updated the layout shift advice to use cumulative layout shift #75

• Changed id of the Google Tag Manager advice #79

• Updated to use latest Coach core and Browertime.

• Updated to use latest Browsertime with latest Chromedriver

• Updated Docker container to use latest Chrome/Firefox

• Use Browsertime 11.0.0

• Use Coach-code 6.0

• Use latest Firefox and Chrome

• When you run on Mac OS we changed so localhost traffic is not throttled by default. To throttle on localhost use --localhost. This makes sence if you use WebDriver to drive your browser on your local machine. And Mac OS now works the same as Linux.

• Fix breakage when multiple default routes are declared , thank you Andy Richardson for PR #62.

• Added missing in_body_parser_blocking render blocking info for Chrome.

• Include request method per asset #105.

• Attach info about render blocking assets coming/working in Chrome 92 #104.

• Attach Google Web Vitals data from Browsertime #101.

• The new header structure actually sucked. Instead follow other tools and flatten headers returning an array #98.

• Removed test directory from the release, thank you Yury Michurin for the PR #75.

• We moved out the core functionality to coach-core. This makes it easy to use the Coach in your own code.

• Upgraded to Browsertime 8.0.1

• Updated throttle frontend to latest version where you can see throttling profiles set by using the API and with custom throttling support.

• Added support for using Github gists see #34.

• Added support for using a config file to configure which HARs to test.

• Added support for copy/paste directly you HAR, configuration file or a gist URL or id.

• Refactor the CSS into Sass, development build process now uses BrowserSync.

• Service worker added for offline support. Workbox is used to achieve this.

• Added support for showing changes in request/responses if you compare the same page #63.

Changed

• New internal structure and use the new configuration object all the way.

• Sort CPU events by name to make it easier to spot regressions

• Upgraded to latest release of template7

• Added more Visual Metrics.

• Upgrad to PageXray 2.4.0

• Layout -> Rendering in latest Browsertime.

• Upgraded to PerfCascade 2.5.2 fixiing Edge tab bug

• CPU metrics are now showed automatically.

• Styled the PageXray info table to make it easier to read.

• Major code refactoring to make it easier to do changes in the future.

• Uploading files failed on Firefox.

• Upgraded to PerfCascade 2.4.1

• You can change/upload HAR files when you already compare two HARs.

Fixed

• The layout in the PageXray table was dependent of the length of the URL. That could make some pages look really bad.

• Automatically load the HAR files if both files are given in the URL . Thanks Ivru for the PR #15.

• You can automatically load one HAR file by adding ?har1=URL&compare=1 as the full URL.

• Updated to PageXray 2.2.1

• You can now drag and drop one HAR file with multiple pages and compare the pages with each others #16.

• Upgraded to PerfCascade 2.2.2 that makes HAR files from WebPageTest Linux render

• Log errors to the console

• Upgraded to PageXray 2.1.0 and added more CPU metrics