Unexpected productIterator order

[Simon Adameit]

Hello,

we noticed some strange behaviour of productIterator in Scala 2.11.5:

case class Data(a: String, b: String, `a-some`: String)

val m = Data("1", "2", "3")

assert(m.productIterator.mkString(", ") == "1, 2, 3")

Result: "1, [3, 2]" did not equal "1, [2, 3]"

"a-some" can be any mangled symbol with an "a" prefix.

Is the order of productIterator on case classes guaranteed to be in parameter order? If so, this seems to be a bug. Can anyone reproduce it?

Regards,

Simon Adameit

[Simon Schäfer]

Product.productElement (http://scala-lang.org/api/current/index.html#scala.Product@productElement%28n:Int%29:Any) says:

The n^th element of this product, 0-based. In other words, for a product A(x1, ..., xk), returns x(n+1) where 0 < n < k.

Therefore, it is a bug in productElement. However, productIterator doesn't specify ordering...

Regards,

Simon Adameit

--
You received this message because you are subscribed to the Google Groups "scala-language" group.
To unsubscribe from this group and stop receiving emails from it, send an email to scala-languag...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Jason Zaugg]

This appears to have the same root cause as SI-8831. I have added your example to that bug, I'll include it as a test case when we fix this.

For now the workaround is to avoid using backticked case class param names that start with the name of another parameter.

-jason