@lukaseder At least 6 out of 8 strike me as bad ideas. I've written an in-depth response here: gist.github.com/rzwitserloot/7… - perhaps I should post this to jdbc-dev as well? It's not a mailing list I frequent.
#RT@project_lombok: Hey @stripe – when you ask non-US individuals to fill in a W-8/W-9 on your site (e.g. for github sponsors), you can't just assume they know US stuff. The dropdown titled "Tax Classification"? It needs a link to irs.gov/individuals/in… – without this, the qu…
#RT@project_lombok: @lukaseder@mariofusco Ordinarily, I strongly dislike hungarian: If you want to eyeball "this is a field of type 'int'", your IDE can do it automatically. But typevars are fundamentally not replaceable with a reified type, so would be okay for it. Idle t…
#RT@project_lombok: @mariofusco Naming it e.g. `Element` means it's easily confused as an actual typed named `Element`. We really need a third casing, but we don't have that. What other convention? Start with _? Start with $? Some unicode symbol?
#RT@project_lombok: @lukaseder Utility methods should be in the relevant type. J.n.f.Files is a silly kludge; those methods should have been in j.n.f.Path. the Collections class shouldn't exist. Utility classes is a trickier debate. Don't have "com.util.IntList", maybe "com…
#RT@project_lombok: @SunshowerJosiah@javahippie@thokuest In this modern age, anything of any import at all will have some strong detractors. Seeing @thokuest post a vomit gif means they are one. And entitled, rude, and uncivilised. However, folks like that are always arou…
#RT@project_lombok: RT @lukaseder: One of the best design decisions in jOOQ has been to just put everything in a single package and make excessive use of package private visibility everywhere.
#RT@project_lombok: Hey, found a bug. In java.lang.String no less. #L1220" target="_blank" rel="nofollow noopener">github.com/openjdk/jdk/bl… - "shouldn't happen"? the last line in the try block would like to have a word. To trigger: Save 9 asciichars to a file, `Files.readString(that, UTF_16)` and `j.l.Error` is thrown. (Beca…
#RT@project_lombok: @hh_wandsbek@vlad_mihalcea Something like 90% of all setters have nothing to validate / validate elsewhere, and there is nothing meaningful to say that the name doesn't already cover. Why do you think Lombok/records exist and are so popular? The "joke" …
#RT@project_lombok: @lukaseder For a while, `for (int x[] : listOfIntArrays)` wasn't legal syntax (brackets on the varname!) as per JLS, but ecj and javac both accept it. I wonder if instanceof final is in the JLS. Probably.. fun find!
#RT@project_lombok: @motion_data@brunoborges "pollutes your dependencies at runtime"? Not sure what you mean. Lombok is a compile time only dependency. It had literally 0 deps at runtime.
#RT@project_lombok: @tagir_valeev@lukaseder Yeah, that's a good one. I wonder if the JLS can just drop the requirement that `this()/super()` is the first line: The JVM doesn't need or have this rule at all, and its primary aim (avoid confusion with half-initialized objects…
#RT@project_lombok: @tagir_valeev@lukaseder Generally I'd think `private static final Foo TRICKY_CONSTANT = someStaticMethod();` more readable than `private static final Foo TRICKY_CONSTANT = do { .... }`. The -do- construct seems to just inject style debates, not actual n…
#RT@project_lombok: @tagir_valeev@lukaseder Parser-wise, yeah that would work backwards-compatibly. What are the benefits though? Lombok would like it, for being able to 'rewrite' single statements into multi-statements in a 100% always-works fashion, but, other than that....
#RT@project_lombok: @sercastiAWS@vlad_mihalcea Probably no shifting. The 'scalability' argument never really was true (facebook runs on SQL, therefore 'SQL does not scale' is baloney, QED), the _point_ of noSQL is that it is not relational - Hibernate fundamentally is. The…