Changes between Version 5 and Version 6 of Commentary/Compiler/CmmType


Ignore:
Timestamp:
Dec 6, 2006 7:34:26 PM (7 years ago)
Author:
p_tanski
Comment:

--

Legend:

Unmodified
Added
Removed
Modified
  • Commentary/Compiler/CmmType

    v5 v6  
    230230  | F80         -- extended double-precision, used in x86 native codegen only. 
    231231}}} 
    232 There is currently no register for floating point vectors, such as `F128`.  The types of Cmm variables are defined in the Happy parser file [[GhcFile(compiler/cmm/CmmParse.y)]] and the Alex lexer file [[GhcFile(compiler/cmm/CmmLex.x)]].  (Happy and Alex will compile these into `CmmParse.hs` and `CmmLex.hs`, respectively.)   
     232There is currently no register for floating point vectors, such as `F128`.  The types of Cmm variables are defined in the Happy parser file [[GhcFile(compiler/cmm/CmmParse.y)]] and the Alex lexer file [[GhcFile(compiler/cmm/CmmLex.x)]].  (Happy and Alex will compile these into `CmmParse.hs` and `CmmLex.hs`, respectively.)  Cmm recognises the following `C--` types as parseable tokens, listed next to their corresponding {{{define}}}s in [[GhcFile(includes/Cmm.h)]] and their STG types: 
     233|| '''Cmm Token''' || '''Cmm.h #define''' || '''STG type''' || 
     234|| `bits8` || `I8` || `StgChar` or `StgWord8` || 
     235|| `bits16` || `I16` || `StgWord16` || 
     236|| `bits32` || `I32`, `CInt`, `CLong` || `StgWord32`; `StgWord` (depending on architecture) || 
     237|| `bits64` || `I64`, `CInt`, `CLong`, `L_` || `StgWord64`; `StgWord` (depending on architecture) || 
     238|| `float32` || `F_` || `StgFloat` || 
     239|| `float64` || `D_` || `StgDouble` ||