Both your suggestions look unlikely to collide.
Another possibility is
{-# HOOK ... #-}
which matches the syntax we use for pragmas generally.
Simon
| -----Original Message-----
| From: Manuel M. T. Chakravarty [mailto:[EMAIL PROTECTED]]
| Sent: 05 October 1999 15:16
| To: [EMAIL PROTECTED]
| Subject: Q: Alternative to (# ... #)
|
|
| I currently use (# and #) to identify the binding hooks in
| annotated Haskell modules for the interface generator
| C->Haskell. These lexemes unfortunately collide with GHC's
| syntax for unboxed tuples. So, I was wondering whether to
| maybe use [# ... #] or {# ... #}. C->Haskell removes all of
| these hooks before the Haskell compiler sees the file, but I
| have to use something that is very unlikely to occur as part
| of the `normal' Haskell code in the file; otherwise, the
| tool might misinterpret it as a (malformed) binding hook.
| Any opinions?
|
| Manuel
|
