CVS difference for ai12s/ai12-0041-1.txt

Differences between 1.3 and version 1.4
Log of other versions for file ai12s/ai12-0041-1.txt

--- ai12s/ai12-0041-1.txt	2014/11/18 06:52:29	1.3
+++ ai12s/ai12-0041-1.txt	2014/11/19 20:58:07	1.4
@@ -1,4 +1,4 @@
-!standard 7.3.2(1/3)                                14-11-17    AI12-0041-1/02
+!standard 7.3.2(1/3)                                14-11-18    AI12-0041-1/03
 !standard 7.3.2(3/3)
 !class binding interpretation 14-10-19
 !status Corrigendum 2015 14-11-17
@@ -49,7 +49,7 @@
 Type_Invariant'Class
 This aspect shall be specified by an expression, called an invariant expression.
 Type_Invariant'Class may be specified on a private_type_declaration{,}[ or] a
-private_extension_declaration{, or an interface_type_declaration}.
+private_extension_declaration{, or a full_type_declaration for an interface type}.
 
 !discussion
 
@@ -106,7 +106,7 @@
 This aspect shall be specified by an @fa<expression>, called an
 @i<invariant expression>. Type_Invariant'Class may be specified on a
 @fa<private_type_declaration>, a @fa<private_extension_declaration>, or
-an @fa<interface_type_declaration>.>
+a @fa<full_type_declaration> for an interface type.>
 
 !ACATS test
 
@@ -218,7 +218,7 @@
 
 ****************************************************************
 
-[Editor's note, November  29, 2012
+Editor's note, November  29, 2012
 
 I put this proposal into the form of an Amendment AI, using the basic
 outline of the original proposal.
@@ -249,5 +249,23 @@
 aspect Type_Interface'Class is allowed. (This surprises the heck out of
 me.) That's small enough that we could consider making this a Binding
 Interpretation.
+
+****************************************************************
+
+Editor's note, November  18, 2014
+
+The wording was:
+
+Type_Invariant'Class
+This aspect shall be specified by an expression, called an invariant expression.
+Type_Invariant'Class may be specified on a private_type_declaration{,}[ or] a
+private_extension_declaration{, or an interface_type_declaration}.
+
+Unfortunately, there is no such thing as an "interface_type_declaration".
+We have to use a longer-winded approach, either
+"full_type_declaration for an interface type" or
+"full_type_declaration that contains an interface_type_definition".
+
+I used the former.
 
 ****************************************************************

Questions? Ask the ACAA Technical Agent