Excerpts from Simon Marlow's message of Wed Sep 08 04:34:05 -0400 2010: > The patch doesn't compile for me: > > compiler/hsSyn/Convert.lhs:373:22: > Not in scope: data constructor `Interruptible'
This is because I have some unrecorded changes in template-haskell library. I've sent the patch. > One idea that occurred to me for bound threads: instead of > pthread_cancel(), we could try pthread_kill() with a suitable signal, > which should be enough to kick the thread out of a blocking system call > and might return it to Haskell land where we can raise the exception. Yeah, cancellation and killing are two different things, but both will basically get you what you want in Linux. However, the state of POSIX style signals is even worse on Windows, so that doesn't really help the porting story. On Vista we can use CancelSynchronousIo in combination with an event of some sort to possibly recover the thread, but I haven't thought this through enough. Edward _______________________________________________ Cvs-ghc mailing list [email protected] http://www.haskell.org/mailman/listinfo/cvs-ghc
