Playwright has around 4500 unique tests as of today. We run all our tests across all operating systems (win/ubuntu20/22/amd/arm64/macos11/macos12, arm etc) and browsers (wk/ff/cr and channels like Google Chrome or Edge) which we support. Some with video/tracing enabled. All our tests are inside the “tests” directory inside the project repository. We have the golden rule of always adding a test for every specific small feature or bug, so it does not regress, hope that helps!
01:08 📙 Debian 12 is now fully supported across all browsers in Playwright 1.37. 04:11 🛠 UI mode updates in Playwright 1.37 allow for project dependency support and provide better debugging capabilities with console logs. 06:29 🖥 The console and network tabs in UI mode now show all the information across the test, making it easier to view logs and network requests. 12:58 📊 The new CLI tool called merge reports in Playwright 1.37 allows for merging multiple blob reports into a single HTML report, making it easier to view the results of parallel test runs. 18:41 🖥 The merge reports tool can be used in any CI environment and provides a convenient way to analyze test results and failures.
@@nirmalchakraborty8608 From what I see, these improvements are mostly for the NodeJS ecosystem (except the Debian 12 support maybe, and the trace viewer improvements). It's kind of logical, since NodeJS is where the team developed their own feature-rich test runner/framework. Most of the focus went to that. The other ecosystems still have the library (which is pretty awesome by itself) and some basic integration with the de-facto test runners for those ecosystems.
Really nice. But, sharding only works good if all test are taking approximately same time. IF 10 test are taking longer then another then overall build time will be a bit higher and most worker will be ideal. If we can add one more feature like AI-Sharding which can give pending test from one shard to another shard’s worker only if worker is ideal?
Has anyone noticed that when a test case has completed execution on Windows, it takes nearly 30 secs before the process exits? Maybe it's just me. But same thing is not noticed when executing on Linux platform. This has been happening since long time and not on any new versions
If you have a reproduction for it, please file an issue on GitHub. Usually this is because it’s saving the trace after the test has been finished. For that change trace from on to on-first-retry. Hope that helps, thanks!
I love the approach to creating videos alongside the release notes to demo changes, it adds a lot of helpful context, thank you!
merge report is a great feature! and 25000 tests.. wow! Could you do a vlog on the various patterns used to develop these tests, please?
💯 +1
Playwright has around 4500 unique tests as of today. We run all our tests across all operating systems (win/ubuntu20/22/amd/arm64/macos11/macos12, arm etc) and browsers (wk/ff/cr and channels like Google Chrome or Edge) which we support. Some with video/tracing enabled. All our tests are inside the “tests” directory inside the project repository. We have the golden rule of always adding a test for every specific small feature or bug, so it does not regress, hope that helps!
You two are a great duo.
01:08 📙 Debian 12 is now fully supported across all browsers in Playwright 1.37.
04:11 🛠 UI mode updates in Playwright 1.37 allow for project dependency support and provide better debugging capabilities with console logs.
06:29 🖥 The console and network tabs in UI mode now show all the information across the test, making it easier to view logs and network requests.
12:58 📊 The new CLI tool called merge reports in Playwright 1.37 allows for merging multiple blob reports into a single HTML report, making it easier to view the results of parallel test runs.
18:41 🖥 The merge reports tool can be used in any CI environment and provides a convenient way to analyze test results and failures.
All these features are available in other programming language bindings or it’s specifically for js
@@nirmalchakraborty8608 From what I see, these improvements are mostly for the NodeJS ecosystem (except the Debian 12 support maybe, and the trace viewer improvements). It's kind of logical, since NodeJS is where the team developed their own feature-rich test runner/framework. Most of the focus went to that. The other ecosystems still have the library (which is pretty awesome by itself) and some basic integration with the de-facto test runners for those ecosystems.
I think just for js
@@nirmalchakraborty8608
Awesome changes! waited for that UI mode dependencies support, now I can share it with the organization!
Nice updates. Thank you for sharing the demo.
This is great work, makes reading of test report easy.
thanks guys a lot!
thank debs thank aslushnikov much.
Looks great!
really awesome!!
Really nice. But, sharding only works good if all test are taking approximately same time. IF 10 test are taking longer then another then overall build time will be a bit higher and most worker will be ideal. If we can add one more feature like AI-Sharding which can give pending test from one shard to another shard’s worker only if worker is ideal?
👏
I wanna introduce this new guy on my current job, any advice to add in my poc ?
BTW this new features are good
Has anyone noticed that when a test case has completed execution on Windows, it takes nearly 30 secs before the process exits? Maybe it's just me. But same thing is not noticed when executing on Linux platform. This has been happening since long time and not on any new versions
If you have a reproduction for it, please file an issue on GitHub.
Usually this is because it’s saving the trace after the test has been finished. For that change trace from on to on-first-retry.
Hope that helps, thanks!
So in Merge Reports do you have also integrated, that the date is written in the header?
Please file a feature request for it, thanks!
Cool! But I got one question.
What about traces and screenshots that are stored in the test-results folder? How do we collect them across all shards?
They are inside the blob and will be part of the final HTML report.
Hi! How to update for v.1.37?
-> npm install -D @playwright/test@latest
playwright.dev/docs/best-practices#keep-your-playwright-dependency-up-to-date
any tips work with firebase admin please ?