Claus Reinke wrote:
You are right that this doesn't help my performance argument, as performance issues are outside the language definition (not observable in the language definition sense). It was merely an answer to the vehement claims that pattern order is irrelevant.
The order of branches in a case expression *in Core* is irrelevant (except that GHC assumes DEFAULT comes first). The order of pattern matches in a Haskell expression is, of course, semantically significant. Nobody is claiming you can change the order of pattern matches in Haskell in general without it changing the meaning.
|It's not that GHC deliberately re-orders case alternatives, |it's that it doesn't deliberately not do it.
no longer is a safe default strategy (actually, it never was, as the bug shows;-). Neither is sorting patterns by constructor tag, as seems to happen on the way to Core.
I was talking about Core. I thought you were too - sorry. Since what you wanted was the order of tests to somehow remain fixed from Haskell through to assembly code, that would imply not reordering them through Core, which GHC does not guarantee not to do. The order of branches in a case expression has no semantic significance in Core. Cheers, Simon PS. nice bug.