[cfe-commits] r46859 - in /cfe/trunk: CodeGen/CGExpr.cpp CodeGen/CodeGenTypes.cpp test/CodeGen/bitfield.c

Chris Lattner clattner at apple.com
Fri Feb 8 14:30:11 PST 2008


On Feb 8, 2008, at 12:02 PM, Lauro Ramos Venancio wrote:

>>
>> Wow, this is going to be seriously painful to deal with for constant
>> initializers.  Any particular suggestions there?  (I suppose it was
>> going to be painful in any case, though.)
>>
>
> I think a mid-term goal should be to handle bitfields in the LLVM
> instead of in the front end.
> Nowadays, I think we should change the placeBitfield function
> (CodeGenType.cpp) to provide more information making easier to
> implement bitfield initializer.

I don't think this will work.  An llvm type like { i5, i11 } doesn't  
pack into 16 bits.

-Chris



More information about the cfe-commits mailing list