[GHC] #1085: Redesign of the scheme for getting the datacon symbol name of a closure in the RTS

GHC trac at galois.com
Fri Jan 5 10:48:18 EST 2007


#1085: Redesign of the scheme for getting the datacon symbol name of a closure in
the RTS
----------------------------+-----------------------------------------------
 Reporter:  mnislaih        |          Owner:                  
     Type:  task            |         Status:  new             
 Priority:  normal          |      Milestone:  6.8             
Component:  Runtime System  |        Version:  6.6             
 Severity:  normal          |     Resolution:                  
 Keywords:                  |     Difficulty:  Moderate (1 day)
 Testcase:                  |   Architecture:  Unknown         
       Os:  Unknown         |  
----------------------------+-----------------------------------------------
Comment (by simonmar):

 No, we definitely don't want to build all the libs twice!

 This only affects constructors, so the binary size increase should be
 minimal.  We'll just live with it.  Besides, we already derive `Data` for
 lots of types, and that instance already contains these strings - perhaps
 we could arrange to share them?

 The right way is to make a new kind of info table for constructors, just
 as we currently have `StgFunInfoTable` for functions, and
 `StgRetInfoTable` for return addresses.  The constructor info table would
 have a single extra field, containing a pointer to the name of the
 constructor (in UTF-8, of course).

-- 
Ticket URL: <http://hackage.haskell.org/trac/ghc/ticket/1085>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler


More information about the Glasgow-haskell-bugs mailing list