Changes between Version 14 and Version 15 of Status/Nov07
- Timestamp:
- 12/10/07 00:55:27 (6 years ago)
Legend:
- Unmodified
- Added
- Removed
- Modified
-
Status/Nov07
v14 v15 22 22 == Syntactic and front-end enhancements == 23 23 24 Several people have developed superficial but perhaps-very-useful 25 syntactic innovations, which are (or will shortly be) in the HEAD: 26 27 * Three improvements to records[[BR]][[BR]] 28 * '''Wild-card patterns for records'''. If you have 24 Several people have developed syntactic innovations, which are (or will shortly be) in the HEAD: 25 26 * '''Three improvements to records'''[[BR]][[BR]] 27 * ''Wild-card patterns for records''. If you have 29 28 {{{ 30 29 data T = MkT {x,y::Int, z::Bool} … … 43 42 record; while in a record construction it uses variables with 44 43 those names to initialise the record fields. Here's the [http://www.haskell.org/ghc/dist/current/docs/users_guide/syntax-extns.html#record-wildcards user manual entry][[BR]][[BR]] 45 * '' 'Record puns''' is a slightly less abbreviated approach. You can write 'f' like this:44 * ''Record puns'' is a slightly less abbreviated approach. You can write 'f' like this: 46 45 {{{ 47 46 f (MkT {x,y}) = x+y … … 49 48 whereas Haskell 98 requires you to write "x=x,y=y" in the pattern. Similarly 50 49 in record construction.[[BR]][[BR]] 51 * '' 'Record field disambiguation''' is useful when there are several types in50 * ''Record field disambiguation'' is useful when there are several types in 52 51 scope, all with the same field name. For example, suppose another data type S 53 52 had an 'x' field. Then if you write … … 158 157 159 158 * Michael Adams came for an internship and built a CPS converter 160 for GHC's internal C-- data type. He had barely left when Norman 161 Ramsey arrived for a short sabbatical. Based on his experience of 162 building back ends for the Quick C-- compiler, he worked on a new 163 zipper-based data structure to represent C-- code, and a sophisticated 164 dataflow framework so that you can write new dataflow analyses in 165 30 mins. 159 for GHC's internal C-- data type. 160 161 * He had barely left when Norman Ramsey arrived for a short sabbatical. Based on his experience of building back ends for the Quick C-- compiler, he worked on a new zipper-based data structure to represent C-- code, and a sophisticated dataflow framework so that you can write new dataflow analyses in 30 mins. 166 162 167 163 * Ben Lippmeir spent his internship building a graph-colouring,
