Moving to Java 17 or how I learned to stop worrying and love NoSuchMethodError
iconsolutions.com
iconsolutions.com
"Barely" is an understatement. You just shouldn't do this as part of a big version migration. It's noise, and introduces more changes that could cause bugs, and then you end up wondering if the bugs were caused by the big version migration, or by all the little changes that were supposedly cosmetic.
Arguably you shouldn't make those types of changes on their own at all. If you're later refactoring the code for a "real" purpose, or making other type of changes in that area, sure, go ahead, the code that you're working on to their more modern forms. But otherwise they're akin to rewrites: the existing code works, and has been tested and debugged, and changing it just so it "looks more modern" is pointless.
I had to sit them all down and explain a variant of your point and the one another poster made about second system syndrome.
This seems to be a application of the second system effect noticed by Fred Brooks in the 1960s (that the second version of a system is extraordinarily dangerous because of all the "while we're in here why don't we..." opportunities)
Moving away from traditional JNI (especially with critical JNI natives) also complicates what used to quick and easy work. But at least Panama clearly comes with interop and compiler-level improvements.
Jigsaw though... I can't even.
Thank you Spring Boot, and long may it continue.
X, or how I learned to stop worrying and love Y
X for fun and profit
A tale of two X
X considered harmful
That just sounds like my life.
X considered harmful considered harmful
X is all you need
I find it especially annoying as it's never even true.
I've been building on 21/22 recently and the amount of boilerplate you need for some basic stuff like serving routes flask-style has dropped dramatically.
Edit: It's possible though that it relates to the CentOS 7 EOL.