Thinking of Rector's focus - from Nette and smaller projects to Symfony

[TomasVotruba]

Hello Rector community,

I have a dilema.
We now support many frameworks/packages in Rector: https://github.com/rectorphp/rector/tree/master/config/level

There are many configs with few rules and just few with bigger coverage - mostly Symfony.

We also try to cover Nette. And that's my dilema.
Nette is not developed for last 5 years and it doesn't seem it's gonna change. It's really dead project and it makes no sense to upgrade, bease last version 2.4 was release 3 years ago, so there is nowhere to upgrade to 😞

I'm considering dropping Nette and other small support, because it's not usable in practise and could cause people much frustration and wanting them "better XY cases coverage".
That could improve our focus on other projects, like Symfony, Laravel and Doctrine, that already evolve

The main purpose of Rector is to help with BC break upgrades, which is currently most useful in Symfony. Nette didn't have a BC break for ~3 years.

What do you think?

[TomasVotruba]

From in person feedback:

• "it makes sense to focus on narrow code, working great for few packagse, rather than poorly for many"
• also "not evolving packages without BC breaks are easier to upgrade manually, Rector helps much more with packages with many BC breaks"

we'll try to focus on few evoling packages more, rather than many and barely.

Main focus should go on Symfony, Sylius - since Symfony makes BC breaks every 2 years, according to release plan, possibly Doctrine.

Packages with almost no BC breaks in last 2 years will be dropped.

[TomasVotruba]

List of most used packages: https://packagist.org/explore/popular

From community there are requests for https://github.com/guzzle/guzzle, that is also in top 10

---

[TomasVotruba]

https://www.jetbrains.com/research/devecosystem-2018/php/