[Cfp-interest] Summary of C FP issues so far

Wed Oct 17 17:47:55 PDT 2018

Hi Jim, others,

Answers below tagged with <Rajan/>.

Also new updates from today:
6.16 pow divide-by-zero case [N 2271]
  Looks good to put in. It will be added to the minutes as an item that 
will go into C2X.
6.17 Min-max functions [N 2273]
  FP study group to split this into two separate proposals. One for parts 
1 and 2 changes based on the TS’s, and the other for part 3 that fits the 
conditionally normative part 3 annex.
6.18 Augmented arithmetic functions [N 2274]
  Not accepted. They said we could bring this forward again as part of a 
future revision of the TS at a later date (i.e. After the IEEE spec is 
published and the C2X integration of parts 1 and 2 are in).
  Nothing to do at this time in WG14, but the floating point group may 
want to bring this forward again for a future revision of the TS.
6.19 C support for IEEE 754-201x [N 2275]
  David: Good job on getting ahead of the IEEE spec and minimizing the 
changes needed.
  Priorities should be do parts 4 and 5 updates to IEEE and then rebase on 
C2X and then bring it forward for a new published revision. 

Regards,

Rajan Bhakta
z/OS XL C/C++ Compiler Technical Architect
ISO C Standards Representative for Canada, PL22.11 Chair (USA)
C Compiler Development
Contact: rbhakta at us.ibm.com, Rajan Bhakta/Houston/IBM

From:   Jim Thomas <jaswthomas at sbcglobal.net>
To:     Rajan Bhakta <rbhakta at us.ibm.com>
Cc:     cfp-interest at oakapple.net
Date:   10/17/2018 08:17 PM
Subject:        Re: [Cfp-interest] Summary of C FP issues so far

Rajan,

Thanks for the summary. I have a few questions below.

Safe travel,

Jim

On Oct 16, 2018, at 5:27 PM, Rajan Bhakta <rbhakta at us.ibm.com> wrote:

Hello,

Here is a summary of all the decisions reached in WG14 regarding the 
floating point study group topics of interest to date. The major items are 
that we need to provide updated Parts 1 and 2 containing all closed 
defects (as of this WG14 meeting, all changed status given below) to the 
WG14 editors, and to create a new version of Part 3 that is a 
conditionally normative annex (if we want Part 3 to be a part of C2X) as a 
separate proposal.

Action items:
  - Give all (CR/DR) changes integrated into the Floating Point TS’s by 
the end of this meeting to the WG14 editors for Parts 1 and 2. Subsequent 
changes will be based on the C2X working draft.

As elaborated below (For the C2X Proposals, Parts 1 and 2 …), right?
<Rajan/>Correct

TS CR's:
    DR17: Move to closed.
    DR18: Move to closed.
    DR19: Move to closed.
    DR13: See N2252. Move to review.
    DR16: Move to review.
    DR20: See N2254. Move to review.
    DR21: See N2283. Move to review.
    DR22: See N2255. Move to review.

For proposed CR/DR's:
P2 CR for llquantep invalid case [N 2262]
  CR23. Move to review.
P1 CR for remainder NaN case [N 2272]
  CR24. Move to review.
P1 CR for totalorder parameters [N 2292]
  CR25. Move to review. The change will be reflected in C2X.
  Note: Jens, Blaine, Aaron, Martin: Prefer Built In Functions. The issue 
is taking the address of the function.

For the C defects:
Move to DR501 into review after incorporating N2253.

For the C2X Proposals:
  Parts 1 and 2 will go in. Note that as per above, the plan is for our 
group to incorporate all changes from closed DR/CR's to the point of the 
end of this WG14 meeting into working drafts of the TS parts 1 and 2 and 
give that to Jens and Larry for incorporation into a working draft of C2X.
    Part 3: New types
      Looking for the floating point study group to turn this into a 
conditionally normative annex (new proposal).

Hmm. The P3 specification includes changes all the way through the 
standard. Did WG14 give any indication of what form an acceptable annex 
might take?
<Rajan/>They indicated it would be a hard problem but they hoped we could 
work through it and leave it as a self contained annex. In a hallway 
conversation with Jens, he seemed ambivalent whether or not we make that 
self contained annex the same as other Annexes or if we basically wrote is 
as changes to C2X like we have now as changes to C11 (but not the changes 
to C11 + part 1 + part 2 obviously).

    Part 4: Mathematical functions:
      No consensus to add this to the working draft of C2X.
    Part 4: Reduction functions:
      No consensus to add this to the working draft of C2X.
    Part 5: Evaluation format pragmas:
      No consensus to add this to the working draft of C2X.
    Part 5: Optimization control pragmas
      No consensus to add this to the working draft of C2X.
    Part 5: Reproducible results
      Not looking for the FP study group to rework this without 
optimization pragma.
      No consensus to add this to the working draft of C2X.

Could you say more about what “no consensus” means?
<Rajan/>Generally it is a nice way of saying "We don't want this"

    Rounding direction macro: Passed (Will be in C2X)
    N2123: Alternate exception handling.
      No consensus to do anything.

Regards,

Rajan Bhakta
z/OS XL C/C++ Compiler Technical Architect
ISO C Standards Representative for Canada, PL22.11 Chair (USA)
C Compiler Development
Contact: rbhakta at us.ibm.com, Rajan Bhakta/Houston/IBM
_______________________________________________
Cfp-interest mailing list
Cfp-interest at oakapple.net
http://mailman.oakapple.net/mailman/listinfo/cfp-interest

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.oakapple.net/pipermail/cfp-interest/attachments/20181017/cc4d3d9c/attachment-0001.html 