CVS difference for ai05s/ai05-0001-1.txt
--- ai05s/ai05-0001-1.txt 2008/11/14 02:39:45 1.4
+++ ai05s/ai05-0001-1.txt 2008/12/02 06:01:19 1.5
@@ -3595,3 +3595,35 @@
****************************************************************
+From: Randy Brukardt
+Sent: Monday, November 10, 2008 9:11 PM
+
+I'm working on the minutes of the Portland meeting, and I have:
+
+"These containers cannot be controlled, because Ada.Finalization is Preelaborated
+and these packages are Pure. There probably should be an AARM implementation note
+to that effect."
+
+But when looking at this, I remembered that the container types are defined
+to "need finalization" (that is, are potentially controlled). We surely don't
+want the bounded forms to have this property, so we need explicit wording to
+say that the type Vector in Bounded_Vectors does not "need finalization". (That
+is, to repeal that part of A.18.2(84/2), as we're planning to inherit everything
+except a specific list of changes from the unbounded forms.)
+
+Similar wording is needed for all of the other bounded form container types.
+Otherwise, Adam will make us fix it in future. ;-)
+
+I preliminarily put a mention of this into the minutes, but since we didn't talk
+about it explicitly in Portland, I wanted to mention it here as well.
+
+****************************************************************
+
+From: Tucker Taft
+Sent: Monday, November 10, 2008 9:22 PM
+
+Good catch. Yes, we clearly don't want the bounded container types to
+"need finalization."
+
+****************************************************************
+
Questions? Ask the ACAA Technical Agent