3 min read
Nobody merged my first five pull requests
My first five pull requests are still open five years later. This is the story of what happened after: building my own gem, committing to Selenium, and ending up a beginner again in ruby/ruby, on purpose.
The first pull request I ever opened was on the train from Aarhus to Horsens. September 2021, laptop on my knees, adding a visible method to an element class in a gem called tap_watir. It was a really small change, but for me it felt huge.
I think already, at that point in time the original maintainers of the project were gone, so nobody saw the 5 PRs I opened, which by the way, are still open up to this date.
So the next thing I tried was Ruby Raider, and I started it five months after the tap_watir PRs went quiet. The first three months were a bit chaotic, after all it is my first open source project. But I still remember the rush when I saw people using my project, there is nothing like building something and getting feedback from real people.
By 2023 Ruby Raider depended on Selenium for basically everything, and every time I hit a rough edge in the Ruby binding I thought, someone should fix that.
In September that year I decided the someone could probably be me. I didn't go for anything cool. I went for RBS files. My theory was that if I picked the chore nobody wanted, the bar for getting it in would be lower.
Well, it took two months, but I felt the rush of contributing to a project that I essentially built my career on, I was absolutely over the moon.
After that PR my excitement just kept going, and nowadays I think I have around 60+ PRs merged in Selenium.
But to be honest, having commit rights didn't mean my code was always right.
In 2025 I had a Firefox fix that never went in, a BiDi command closed, a timeout change closed, and a beta Chrome PR that got merged, reverted by me a week later cause it broke the build, and then merged again. Funny enough, half the time a PR gets closed I agree with the reason by the time I finish reading it.
In 2025 I had a Firefox fix that never went in, a BiDi command closed, a timeout change closed, and a beta Chrome PR that got merged, reverted by me a week later cause it broke the build, and then merged again. Funny enough, half the time a PR gets closed I agree with the reason by the time I finish reading it.
By the end of 2025 I took another big jump and opened my first PR to ruby/ruby.
A typo fix in some comments, nothing fancy, but it got merged and it felt really close to that first visible method in 2021. Since then I have around a dozen PRs there and in ruby/openssl, a few merged, most closed or still waiting. Four years later I'm a beginner again, and it still feels exciting.
A typo fix in some comments, nothing fancy, but it got merged and it felt really close to that first visible method in 2021. Since then I have around a dozen PRs there and in ruby/openssl, a few merged, most closed or still waiting. Four years later I'm a beginner again, and it still feels exciting.
I still have a PR open from October, that I check now and then. Open source got me comfortable with being wrong in public, with waiting, and with people I have never met telling me no. Which are skills I use every day.
I'm still learning and I'm still committing, sometimes from a train. So, If you have five PRs sitting open somewhere, open a sixth.