Ah I see my mistake , your just using the regions in the compiler not at run time but knowing an object or static root is immutable is also very useful to a GC or run time clone / threading routine. It is probably worth using a bit in the type field for this for runtime objects ( provided your ok with mutable xyz and immutable xyz being different types or masking it for type comparisons) . Though for a GC immutable is immutable and all its descendents.
Ben >-----Original Message----- >From: William Leslie [mailto:[email protected]] >Sent: Friday, November 05, 2010 9:02 AM >To: [email protected]; Discussions about the BitC language >Subject: Re: [bitc-dev] Vectors and mutability and bears (oh my) > >On 5 November 2010 11:19, Ben Kloosterman <[email protected]> wrote: >> Arent regions good for heap but problematic for static data ( and >maybe >> stack) ? tagging all static data would be expensive ( though static >> immutable data is often const or readonly so maybe there is some way) > >Your region system can capture as much or as little information as you >want it to. The region system for the compiler I'm working on at the >moment has little interest in distinguishing different indexes of the >same region with array type, for example. When it does, you can >annotate the region of interest as needed, I use this for global, >static lists of functions, for example. > >As far as dependent types go, it would be nice to specify *which* >dependent types imply region types. > >-- >William Leslie >No virus found in this incoming message. >Checked by AVG - www.avg.com >Version: 9.0.864 / Virus Database: 271.1.1/3230 - Release Date: 11/03/10 >16:36:00 _______________________________________________ bitc-dev mailing list [email protected] http://www.coyotos.org/mailman/listinfo/bitc-dev
