Version 1.5 of ai12s/ai12-0176-1.txt

Unformatted version of ai12s/ai12-0176-1.txt version 1.5
Other versions for file ai12s/ai12-0176-1.txt

!standard 6.1.1(26.4/4)          16-01-28 AI12-0176-1/02
!class confirmation 16-01-28
!status ARG Approved 10-0-0 (By Letter Ballot) 16-02-26
!status work item 16-01-28
!status ARG Approved 7-0-1 15-10-16
!status work item 15-10-08
!status received 15-09-25
!priority Low
!difficulty Easy
!subject 6.1.1(26.4/4) only applies to tagged types
6.1.1(26.4/4) only applies to tagged types.
6.1.1(26.4/4) starts "If X is of a specific type T then" but the following wording uses T'Class. Clearly, this only applies to tagged types and the wording should say so, right? (Yes.)
This wording originated in AI12-0032-1. In that AI, and in the Corrigendum documents, the wording is:
* If X is of a specific tagged type T then
One presumes the questioner was reading one of the versions of the consolidated RM, where indeed the wording is missing the word "tagged". But the consolidated RM is unofficial; the wording in the Corrigendum document is used if there is a difference. Thus there is nothing to fix here.
No ASIS effect.
!ACATS test
This should be covered by tests for X'Old.

From: Tucker Taft
Sent: Friday, September 25, 2015  9:20 AM

Ultra-minor editorial issue:

I happened to be reading the section on Preconditions and Postconditions
(6.1.1) and ran across this in 26.4/4:

    If X is of a specific type T then
       ... T'Class(X)

It seems we should say "If X is of a specific *tagged* type T then..."

Perhaps "specific" implies tagged, but I am not sure of that, and adding "tagged"
would help the reader in any case, I believe.


From: Randy Brukardt
Sent: Thursday, October 8, 2015  9:52 PM

3.4.1(3/2) says:
Every type is either a specific type, a class-wide type, or a universal type.
A specific type is one defined by a type_declaration, a
formal_type_declaration, or a full type definition embedded in another
construct. Class-wide and universal types are implicitly defined...

Ergo, this bullet applies to almost all types as written, and then it is of
course nonsense for untagged types. So it's not quite as minor an issue as you
seem to imply. (The intent is fairly obvious anyway, so we don't need to rush
on this one but it should get fixed.)


From: Randy Brukardt
Sent: Thursday, January 28, 2016  7:28 PM

[The important part of a longer, unrelated message. - Editor]

P.S. In looking at AI12-0032-1, I see that AI12-0176-1 is unnecessary, as
AI12-0032-1 clearly says "specific tagged type" in 6.1.1(26.4/4), both in the
wording and the !corrigendum (and in the actual Corrigendum document).
Apparently, the editor botched the consolidated RM again. Self-directed grumble.


From: Randy Brukardt
Sent: Friday, February 26, 2016  8:00 PM

Voting results for Letter Ballot on AI12-0176-1/02. This AI is approved by a
10-0-0 vote.

Approve as is: Randy Brukardt, Steve Baird, Bob Duff, Tucker Taft, Jean-Pierre
Rosen, Jeff Cousins, Ed Schonberg, Tullio Vardanega, Gary Dismukes, and Brad

Not voting: John Barnes, Alan Burns, Pascal Leroy, Erhard Ploedereder, and
Florian Schanda.


Questions? Ask the ACAA Technical Agent