Skip to main content
9 October 2023

2 different ways to use Baseline in your visual regression testing process

While you can directly compare two websites with VRT there is a notion of a Baseline set that you always compare against. But why would you need one?

We see two main use cases:

  1. monitoring
  2. save time in pull request testing

 

Monitoring

Let's say you set up daily monitoring. This is when you compare your screenshots always with "yesterday's" version.

Imagine there was a bug introduced on a certain day. This means that only the next day's monitoring run will trigger the error. All subsequent ones won't.

Because of the nature of VRT you do get a tendency to get false positives here and there. So you can easily accidentally skip that day with an actual error and see that tests got green back again the next day.

baseline daily monitoring use case

This is where the Baseline would help.

Instead of comparing with yesterday's version, you could compare against the Baseline. So instead of just next-day error you would see all subsequent tests turn red and it will force you to check them out and approve the change (set a new baseline) or fix the bug.

 

Pull requests testing

Once you set up testing pull requests you would like to compare their screenshots with some particular environment (for example staging).

You create screenshots from staging every time you do screenshots from your pull request environment, but instead, you could reuse screenshots from the staging environment multiple times.

You can take them only when you actually deploy changes to the staging environment and assign them as your baseline. In your pull request pipeline you will be comparing your pull request screenshots with baseline instead.

 

How to set baseline screenshots in Diffy

You can set either whole set of screenshots or individual screenshots using drop down menu.

How to set baseline in Diffy

 

 

Dealing with ShadowDOM with jQuery

Recently we started testing https://urban.org and we wanted to take screenshots of the dropdown menu and search box, but they were implemented as components with shadowDOM.
2 mins read read

Regression Testing: Your Step-by-Step Action Plan

Regression testing can be an overwhelming task. You can try to do too much and fail because of a lack of resources. In this article we will talk about how you can start with small practical steps to get results fast. Once you set up basic tests like visual regression testing you can start with more complex implementations including automation.
7 min read

25 Web Dev & QA Testing Statistics Every Developer Should Know [2022]

Trends move quickly in the web development and QA testing space. These trends can help predict where the QA and testing industry is headed. Development teams can adapt emerging best practices and methodologies while gaining a better understanding of the challenges they face throughout the development cycle.

 

7 min read

Diffy helps your QA team

to ensure that websites don't get visual bugs