| Andrew Tolmach. Combining closure conversion with closure analysis using algebraic types. Presented at ACM SIGPLAN Types in Compilation Workshop, June 8, Amsterdam, January 1997. |
....between flow logics and type systems is an intriguing theoretical question, it has important practical ramifications as well. Flow information is useful for guiding and or enhancing a wide variety of analyses and optimizations, such as closure conversion( SW97, DWM 01] defunctionalization( Tol97, TO98, CJW00] inlining( WJ98] uncurrying ( HH98] eager thunk evaluation ( Fax93] dead code elimination ( WS99] run time check elimination ( WJ98] loop detection[SGL96] and object specialization ( DCG95, PC95] Encoding flow information into type systems enables type directed ....
A. Tolmach. Combining closure conversion with closure analysis using algebraic types. In Proc. First Int'l Workshop on Types in Compilation, June 1997.
....as the source program, we have to worry whether lambda lifting preserves typability. It turns out that none of the published lambda lifters does. This does not come as a complete surprise because closure conversion [21] a close relative to lambda lifting gives rise to typing problems, too [15, 27, 26, 2]. Essentially, all presentations of various aspects of lambda lifting deal with untyped languages [10, 17, 19, 20, 6] This is fine, because they are solving an important problem in an untyped intermediate language after type checking has been performed. In addition, most of them are informal, ....
....The next two approaches differ in that their closures do not contain functions anymore, but rely on separate dispatch functions (apply) that inspect the closure s tag and call the respective function [21] The outcomes of their transformations are first order programs. Oliva and Tolmach [27, 26] employ a solution to closure conversion quite similar to ours [24] It is based on introducing a special closure datatype for each function type that arises the program. They have some refinements that allow them to perform inter module calls and to incorporate the results of a flow analysis. ....
Andrew Tolmach. Combining closure conversion with closure analysis using algebraic types. In Proceedings of the ACM SIGPLAN Workshop on Types in Compilation (TIC'97), Amsterdam, The Netherlands, June 1997.
.... Our algorithm also depends on having the full source program available; this restriction can be lifted if we permit extensible datatype declarations, i.e. datatypes for which the data constructor declarations can be scattered throughout the program, even in separate compilation units #Tolmach, 1997#. Supporting such datatypes requires only a small extension to the type system #Standard ML treats the built in exception type constructor in this way#, but requires a somewhat more expensive implementation of case, and precludes the optimizations discussed in the next section. 10 Optimization ....
Tolmach, A. #1997#. Combining closure conversion with closure analysis using algebraic types. In Workshop on Types in Compilation TIC97. Boston College Computer Science Technical Report BCCS-97-03.
....implemented. Our algorithm also depends on having the full source program available; this restriction can be lifted if we permit extensible datatype declarations, i.e. datatypes for which the data constructor declarations can be scattered throughout the program, even in separate compilation units (Tolmach, 1997). Supporting such datatypes requires only a small extension to the type system (Standard ML treats the built in exception type constructor in this way) but requires a somewhat more expensive implementation of case, and precludes the optimizations discussed in the next section. 10 Optimization ....
Tolmach, A. (1997). Combining closure conversion with closure analysis using algebraic types. In Workshop on Types in Compilation TIC97. Boston College Computer Science Technical Report BCCS-97-03.
....implemented. Our algorithm also depends on having the full source program available; this restriction can be lifted if we permit extensible datatype declarations, i.e. datatypes for which the data constructor declarations can be scattered throughout the program, even in separate compilation units [51]. Supporting such datatypes requires only a small extension to the type system (Standard ML treats the built in exception type constructor in this way) but requires a somewhat more expensive implementation of case, and precludes the optimizations discussed in the next section. 10 Optimization ....
A. Tolmach. Combining closure conversion with closure analysis using algebraic types. In Workshop on Types in Compilation TIC97, June 1997. Boston College Computer Science Technical Report BCCS-97-03.
No context found.
Andrew Tolmach. Combining closure conversion with closure analysis using algebraic types. Presented at ACM SIGPLAN Types in Compilation Workshop, June 8, Amsterdam, January 1997.
No context found.
A. Tolmach. Combining closure conversion with closure analysis using algebraic types. In Proc. First Int'l Workshop on Types in Compilation, June 1997.
No context found.
Andrew Tolmach. Combining closure conversion with closure analysis using algebraic types. In Workshop on Types in Compilation (TIC), June 1997.
No context found.
Andrew Tolmach. Combining closure conversion with closure analysis using algebraic types. Presented at ACM SIGPLAN Types in Compilation Workshop, June 8, Amsterdam, January 1997.
Online articles have much greater impact More about CiteSeer.IST Add search form to your site Submit documents Feedback
CiteSeer.IST - Copyright Penn State and NEC