Yes, you won't get an argument against doing this from me -- this commit is really work in progress (i.e., a disk flush). The Cambridge lot wanted to revamp their RTS invocation API, so I'm hoping we can come up with something that's close between the two systems whenever that happens. A Diet HEP, if you like. --sigbjorn ----- Original Message ----- From: "Alastair David Reid" <reid@cs.utah.edu> To: "Sigbjorn Finne" <sof@glass.cse.ogi.edu> Cc: <cvs-hugs@haskell.org> Sent: Tuesday, February 26, 2002 08:58 Subject: Re: cvs commit: hugs98/src/bcc32 options.h hugs98/src/djgpp2 options.h hugs98/src/epoc options.h hugs98/src/machugs options.h hugs98/src/msc options.h hugs98/src/winhugs options.h hugs98/src options.h.in server.h server.c
- doIO_Int(int*) -- evaluate (IO Int) action. - doIO_Addr(void**) -- evaluate (IO Addr) action.
Instead of this, how about having doIO leave an HVal on the stack which you can manipulate separately?
doIO_Void, doIO_Int, doIO_Addr, etc. could be provided as utilities on top of this.
-- Alastair Reid reid@cs.utah.edu http://www.cs.utah.edu/~reid/ _______________________________________________ Cvs-hugs mailing list Cvs-hugs@haskell.org http://www.haskell.org/mailman/listinfo/cvs-hugs