(j3.2006) Question about generic names, procedure names, and constructors

Van Snyder Van.Snyder
Mon Nov 23 20:48:23 EST 2009

Here's what I have so far in my "last minute stuff" paper.  There's
nothing here that changes technical content of 09-007r3, although the
correction advocated for C845 would change the technical content of the
US NB comment.  Would it be OK for US TAG to caucus and approve that
change at the meeting?

On Mon, 2009-11-23 at 17:36 -0800, Malcolm Cohen wrote:
> Van Snyder wrote:
> > I'm preparing a J3 paper of "Last minute stuff" for 191.
> I suggest you also include the following that Stan reminded me about:
> In 09-007r3 13.7.111 [370:26] please change
> 6 Example. MERGE BITS (13, 18, 22) has the value 20.
> to
> 6 Example. MERGE BITS (13, 18, 22) has the value 4.
> Also, 09-007r3 13.7.150 [388:25] please change
> 7 Example. SHIFTA (IBSET (0, BIT SIZE (0)), 2) is equal to SHIFTL (7, BIT SIZE 
> (0) - 3).
> to
> 7 Example. SHIFTA (IBSET (0, BIT SIZE (0) - 1), 2) is equal to SHIFTL (7, BIT 
> SIZE (0) - 3).
> > Is WG5 allowed to take up such matters if they're not in a NB comment?
> Definitely not if there is any technical effect, unless we want to go back to CD 
> stage.  Even for "bug fixes", our chances of getting away with saying the FCD 
> ballot passed and going to DIS are reduced if we start changing anything not 
> mentioned in a country comment.  Anything else would have to be very purely 
> editorial (like mispelings and missing or extraneous, commas).
> Cheers,
-------------- next part --------------
To: J3                                                       J3/10-xxx
From: Van Snyder
Subject: Last minute stuff
Date: 23 November 2009

[186:5-7 8.1.10 C845]---------------------------------------------------
C845 as revised by J3/09-290r2:

"C845  (R850) An <exit-stmt> shall not appear within a CRITICAL or DO
       CONCURRENT construct if it belongs to an outer construct."

is different from how it was proposed in 09-290, and the result is
different from C845 in 09-007r3.  The revised one allows an EXIT
statement to belong to a CRITICAL or DO CONCURRENT construct, even
though the original didn't.  If we don't want to change its technical
content, the revision proposed in 09-290 should be used.

We certainly don't want to allow EXIT to belong to a DO CONCURRENT
construct.  J3/09-291 advocated to allow EXIT (and branch and return as
well) to exit from a CRITICAL construct, but J3 decided not to advocate
for that change.

Editor: Replace C845:

"C845  (R850) An <exit-stmt> shall not appear within a CRITICAL or DO
       CONCURRENT construct if it belongs to that construct or an outer

Contradicting [281:33], this seems to say that the only time one is
allowed to specify an impure interface body is if the procedure itself
is pure.

Editor: Insert "even" before "if".

[307:35+3-7 NOTE 12.43]-------------------------------------------------
The final sentence of NOTE 12.43 isn't quite right. How can the storage
occupied by a function result be released before the result is used in
the evaluation of an expression?  And the sentence isn't quite right if
the result is a pointer.

Editor: Replace final sentence ("However....") with

"The final value of a function result that is a data object is used
subsequently in the evaluation of the expression that invoked the
function.  If it is not a pointer, an implementation should defer
releasing the storage it occupies until after its value has been used
in expression evaluation."

[370:26 13.7.111]-------------------------------------------------------
Editor: Replace
"6 Example. MERGE BITS (13, 18, 22) has the value 20."
"6 Example. MERGE BITS (13, 18, 22) has the value 4."

[388:25 13.7.150]-------------------------------------------------------
Editor: Replace
"7 Example. SHIFTA (IBSET (0, BIT_SIZE (0)), 2) is equal to
   SHIFTL (7, BIT_SIZE (0) - 3)."
"7 Example. SHIFTA (IBSET (0, BIT_SIZE (0) - 1), 2) is equal to
   SHIFTL (7, BIT_SIZE (0) - 3)."

[442:19 16.3.1p3]------------------------------------------------------- appears not to be germane. appears to be the correct
cross reference.

Editor: Replace "" by "".

More information about the J3 mailing list