In our innovation week I looked at an overlapping mesh of topics that (I think) all dovetailed together nicely. Initially I was interested in checking out Cypress component testing – a new feature released a few versions ago that I had read up on a little but hadn’t had time to play around with much. I also used the week to catch up on work with our more traditional Cypress testing (end to end automation), making some headway into the Funding Adjustment tests. Finally – during all this I was playing around with the new GitHub AI coding assistant (GitHub Copilot).
For Cypress component testing I will break protocol and start with my conclusion: I do not think that we can use component testing on G2, at least not now. It is a form of testing that is for more early in the development process, likely used by devs as a tool in development somewhere around the time unit tests would be created. Using this tool you can “mount” a react component sort of up on blocks, developing quick and cheap tests that would mock out it’s API calls and test it’s behavior. Per some conversations with our devs, the G2 stack doesn’t lend itself very well to this unfortunately. Beyond React there is MUI, Redux and other technologies that would need to be folded in to replicate a true test for a component. Never the less, I think the time was well spent. I learned a lot about React in particular – completing a course on Pluralsite to learn the basics of it. Before innovation week I only had a rudimentary understanding of what a component was and how React functioned and now I think I have a better understanding of it. Any additional frontend knowledge is always a plus, especially for QA purposes.
During development of the E2E Cypress test for Funding Adjustments I used the CoPilot, an AI coding assistant from GitHub. There are a lot of these dotting the landscape this year, and since it looked like CoPilot had gotten the most downloads (on VS Code extension market) I figured I would try it. I did a 1 month free trial and would need to subscribe if I kept using this particular one. What I found was sort of what I expected – it’s not a miracle worker but I think it can be quite handy, particularly for (IMHO) a coder who is at an intermediate level. The code suggestions it offers are pretty cool but not always right – I think a novice would be confused by this, you would need to be able to tell the good suggestions from the bad. But outside of that, it does a nice job of completing repetitive lines of code and noticing patterns in your coding that it can offer to auto-complete. It can save a lot of time and (particularly for intermediate, less experienced people) maybe help you remember all the ingredients of an arrow function that you still have a hard time calling up strictly out of memory.
