CVS difference for ai05s/ai05-0212-1.txt

Differences between 1.22 and version 1.23
Log of other versions for file ai05s/ai05-0212-1.txt

--- ai05s/ai05-0212-1.txt	2012/05/18 06:45:20	1.22
+++ ai05s/ai05-0212-1.txt	2012/12/05 04:20:23	1.23
@@ -6819,7 +6819,8 @@
 > where someone wants to ensure that a package can be used as a
 > signature -- because adding anything that would cause trouble will
 > automatically be detected (and not be breaking some client's
-> instantiations!). So I don't see any harm in adding that feature now to solve the immediate problem cleanly.
+> instantiations!). So I don't see any harm in adding that feature now to solve
+> the immediate problem cleanly.
 
 I'm not sure I agree that the formal incomplete proposal won't be regretted
 later if/when a more general solution is found.  My understanding is that there
@@ -6882,16 +6883,21 @@
 > this problem is
 > possible.)
 
-Is the incomplete-private proposal better than the signature-generic proposal in that it would allow container packages (with bodies) to be instantiated with incomplete private types?  If so, that would be a nice advantage, but I thought the proposal was 
aimed primarily at the iterators package.
+Is the incomplete-private proposal better than the signature-generic proposal in that
+it would allow container packages (with bodies) to be instantiated with incomplete
+private types?  If so, that would be a nice advantage, but I thought the proposal
+was aimed primarily at the iterators package.
 
-
 > As I noted previously, the proposal that Tucker is suggesting applies
 > to all generics, and the freezing depends only on the contents. I view
 > this as dangerous for maintenance, as a clueless maintainer (me,
 > perhaps :-)
 
 I think this depends on the details of the signature-package proposal.
-I agree that if the freezing depends on the contents of the generic spec, then there might be somewhat of a maintenance hazard.  But if the generic spec is marked as freezable, then the hazard goes away, with less wasted effort than marking each formal pr
ivate type.
+I agree that if the freezing depends on the contents of the generic spec, then there
+might be somewhat of a maintenance hazard.  But if the generic spec is marked as
+freezable, then the hazard goes away, with less wasted effort than marking each
+formal private type.
 
 ****************************************************************
 

Questions? Ask the ACAA Technical Agent