A migration celebration, the wonders of teamwork and sadness of binning work

I'm coming to you this week from my first day of a week of annual leave, which is why this weekly roundup is early. Oh and there won't be one next week! I plan to do very little this next week apart from the odd sewing project and maybe a (local) adventure or two. I can't wait! Let's dive into the weekly roundup.

The migration is done!

Yesterday was my last day before going on holiday and I managed to merge the migration PR my team and I have been working towards for the last month or so. It was a huge anti-climax, which is exactly what you want with these things. The whole team was all on a call in case things went horribly wrong, but we hit merge and... Nothing happened! The pipelines went green, the PR went out and all was good. Of course, it being such a huge project, we found some issues soon after it was merged so we spent the rest of the day fixing forward, but we still had two days of code freeze left because we finished a bit early so there was no real rush. So yes, if you ever have to do a giant migration that touches hundreds of files, pairing with another engineer and getting the team to swarm test the results combined with a soft code freeze seems to be a great approach.

Team work (makes the dream work)

This migration was a real team effort. It was two weeks straight of constant pairing for me and my pairing partner, with daily syncs with the managers and product owner on our team, followed by swarm testing and swarm fixing from everyone. It felt so good to have everyone get stuck in and it really shows the value and support of teamwork. Even though I did a lot of the groundwork for this migration, the rest of the team really took ownership of it and we ended up completing it together. Maybe I'm being sentimental, but I want to say thank you to my team for being such amazing folks! It really is testament to the way we work together that the migration was such a non-event - for a PR that touched many hundreds of files, that is absolutely no mean feat.

Remember those visual regression tests?

You might not, but a few weeks ago, I mentioned that I was going to write some visual regression tests in Playwright and that I was going to use those to check for regressions in the migration work. Well, I ended up writing the tests and then... didn't use them. It just didn't feel like the best use of our time. My pairing partner and I were busy enough implementing all the changes for the migration and then fixing all the broken unit and end to end tests that we didn't really have time before addressing the results of the swarm test. I put in a PR yesterday to delete the visual regression suite from our codebase (it wasn't remotely good enough to be used in our pipelines) and it felt a little sad in one way but such a huge victory in another. We hadn't even needed these tests in the end and that seems to me to be a good thing (see previous note about teamwork). It always feels a bit deflating though when you write some code or implement a feature and it doesn't see the light of day, so I guess I wanted to muse on that. It's important to know when something isn't right and it's best to change tack and I feel like it's best to take those changes in direction with good humour and accept that in tech, this will always happen. It doesn't make the feelings any easier to deal with though, does it... In this particular case, it was a bit of throwaway work mostly written by AI, but when it's a full feature that you've worked on for months, this kind of thing can be a lot tougher to swallow. If you've got any advice or ways you deal with the disappointment of having to bin off a load of work you're proud of, let me know! It's definitely an area I could still grow in.