David M. Lloyd
David M. Lloyd
It would be a subtly incompatible change in spec. And some sources might not have a concept of "order" at all.
I'm not sure there's a better solution than "don't rely on ordering".
Lexicographical sorting of keys could be an option. Not the same thing but at least would yield predictability.
With #209, this is closer to reality. The CDI integration uses a generic converter resolution strategy that could be adapted here. But how the API is exposed is a tricky...
This may actually be a bug in [`smallrye/smallrye-common`'s expression module string parser](https://github.com/smallrye/smallrye-common/blob/main/expression/src/main/java/io/smallrye/common/expression/Expression.java#L245).
~~The expression `${\}` should also work to produce a literal backslash.~~ No, that's incorrect; sorry. I'm reading the code and I was in the `ESCAPES` section when I thought I...
In this regard (as in many others) I think we can see that MP Config is fundamentally broken. The caution/lesson here (which, as you know, I've stated repeatedly to the...
In my opinion I think we're better off having a consistent syntax that we can support, and just documenting the differences with MP Config and giving up on spec compliance...
I'd have to review the syntax rules again before having an informed opinion, but I think it's a reasonable idea. My feeling is that `\$` is likely to still be...
Just adding a note here that JPMS modules are not necessary in Java 9+ generally speaking (i.e. the classpath still works fine), but JavaFX (OpenJFX) does use them.