cvs commit: fptools/ghc/compiler/cmm PprC.hs
Donald Bruce Stewart
dons at cse.unsw.edu.au
Tue Feb 10 20:39:32 EST 2004
> | So at the moment I'm assuming the C backend isn't going to generate
> any void*
> | casts. But I'm not certain this is the case.
> I'll talk to Simon about this.
> | I was walking through PprAbsC, looking for things that could be ported
> | across to Cmm directly. cCheckMacroText and friends looked like
> | candidates, until I checked that we don't have those macros in Cmm.
> | cCheckMacroText HP_CHK_NP = SLIT("HP_CHK_NP")
> | So HP_CHK_NP isn't defined in Cmm. I don't fully understand how to
> | convert this code to the Cmm situation, at the moment. But I've
> | allocated the next few weeks to backend-hacking.
> This one is easy: just throw away all that stuff. The code generator
> directly generates the *output* of the current macro. Instead of
> generating a CMacro call, it generates "if .. then jump c23" etc.
> That's part of the point -- it has to happen for the native code gen and
> for C--, so it makes it easier to do it for C too.
That's what I assumed, hence the comment about it not living here
More information about the Cvs-ghc