A 3-part series.
Upgrading Doctrine Fixtures can be challenging work. It requires the coordination of four different package groups on Github. One cannot be upgraded without the other.
We first prepare for the jump, then take the leap with confidence. I'm writing this post as I handle the upgrade, so my thoughts are fresh, and the process is ongoing. It may evolve based on new information. Let's see where the rabbit hole goes.
In the first part, we've kicked off the plan to upgrade all these packages to their latest version, like a blind map into unknown territory. Since then, we've put in a couple of months of hard work and climbed the terrain.
Today, we look at the practical steps we've taken and the new challenges we discovered after the first hill.
In previous parts (part 1 and part 2) we looked at a situation where we are stuck with Alice, Nelmio, Hautelook, Faker legacy mess and need a way out to modern maintainable Doctrine.
This is the final part. The upgrade is finished and closed for the time being. In an unexpected way, it was fast, easy to migrate to, and will be even easier to maintain. It took me a couple of months of deep thinking and experiments, because there was a trap solution that would take us 6-8 months of hard work and would create even more legacy code. Lesson learned: don't fall for marketing and GPT answers.
Let's look into it.
