[cfp-interest 3942] JTC1/SC22/WG14/CFP 2026/05/13 Meeting notes

Mon May 18 10:37:14 PDT 2026

CFP Meeting

2026/05/13: 8:00 AM PDT/3:00 PM UTC

[Please submit proposed changes to these minutes to Jerome or to the group.
Revision changes appear at the bottom.]

Zoom

Join from PC, Mac, Linux, iOS or Android:
https://iso.zoom.us/j/9495503012?pwd=vHaKRZLdPqqyOkfDlPVblVgXbIbzeY.1

Attendees and introductions

     Rajan Bhakta, Jerome Coonen, Damian McGuckin, Joshua Cranmer, David
Hough, Ariel Burton, Guy Davidson

Approval of agenda

  https://cfp-wiki.esi.com.au/pub/CFP/WebHome/CFP_agenda_20260513_V1.pdf
<https://cfp-wiki.esi.com.au/pub/CFP/WebHome/CFP%20meeting%20agenda-20260415.pdf?_t=1776288481>

Previous meeting notes

  https://www.open-std.org/jtc1/sc22/wg14/www/docs/n3878.pdf - 2026/04/15
Minutes

Study group logistics

    Next meeting: 24 June 2026, 8:00 AM PDT/3:00 PM UTC

    ISO Zoom teleconference

    Group: Clang meets the 1st and 3rd Wed each month and 754 meets every
other Wed, so there is plenty of potential for conflict. For now, we stick
with the 3rd Wed in June, with possible updates to follow.

    David: Mailman is undergoing revisions to old systems, but improving.
Everything and everyone in this system is aging.

C documents

    The latest C2Y draft is N3854 Mar. 2026
https://www.open-std.org/jtc1/sc22/wg14/www/docs/n3854.pdf

    C23 has been published ISO/IEC 9899, available for purchase.
https://www.iso.org/standard/82075.html

IEEE 754 update

    Damian: New routines nextXXXX() are related to roundToOdd
functionality. Wording changes related to rounding directions are underway,
similar to what’s happening here in C. They are considering removing
“direction” from “rounding direction attributes”. There is much new
specification about the scope of error flags. A proposal to eliminate flags
in favor of reporting via NaNs is under discussion. That proposal has led
to further discussion about NaN payloads and propagation, especially
narrowing and widening. There is some consideration of “decimal
re-encoding”.

    Damian: Elaborating on the role of the WG14 liaison to P754R, any
formal report to WG14 about P75, must be approved by the standards
association. The relationship is not symmetric, that is, WG14 allows all
good-faith contact with P754R about C and does not require quite so formal
an arrangement.

WG14 update

    Rajan: A few emails received this month are included in the discussion.

C++ liaison

    Joshua: Week of 8 June at Mendel University in Brno, n Czechia? Guy is
working on a paper of interest to this group.

    Guy: There is interest in improving floating point in C++, which is
good news.

TS-4 and TS-5 revisions

    None

News

    None

Carryover action items from last meeting

    None

Action items from last meeting

  Jerome: Submit the fromfp paper.
Done

  Rajan & Jerome: Write a proposal to fix complex range bounds.
Done

  Damian: Update Annex F proposal for discussion at next meeting.
Done

Discussion of issues

   1.

   WG14: nexttoward has no exception to conversion
   Rajan: Hubert sent email about nexttoward vs. nextafter, referring to
   the conversion of the 2nd parameter to the type of the return value of the
   function.
   Damian: These functions are destined for deprecation; they are
   C-specific as an adjunct to the nextafter family, which has been deprecated
   in IEEE 754 since 2008. Nextafter, the original, keeps compatibility with
   754-1985. (And subsequent to the meeting, Rajan observes the matter of
   language here. It is arguments that are converted to the type of the
   parameter, which simply persists in its defined type.)
   Actions

   2.

   WG14: Annex H and non-arithmetic interchange formats
   Rajan: This pertains to H.11.4.1. The language suggests that if an
   implementation does not  have nonarithmetic interchange types, then it
   doesn’t have to implement any of the functions in H.11.4. That is not true,
   because H.11.4 includes strictly arithmetic conversions, too.
   Group: Tries several proposed rewordings on the fly, before the clear
   direction is to assign an Action.

   3.

   Annex F Special Cases
   Damian: Discusses changes suggested by Jim, based on recent C2Y changes.
   Notes changes to Table F2 related to fromfp. Much work on the paragraphs on
   infinities and NaNs, where changes had been fragmented. F.3 consolidates
   much
   Challenge question for the group: Name a function that, given a quiet
   NaN argument,  yields an exception.
   Action

   4.

   Editorial issues in Table F.2
   Damian: This cleans up some language, using the general word
   “facilities” to encompass functions and other items. “Operations” is a
   related word, also used in its general sense.
   Group: Note that FLT_RADIX is a macro with no argument, leading to some
   existential discussion about its status relative to the other “actions” in
   Table F.2.
   Multiple actions, including to clean up F.3#1.

   5.

   convertFrom and SNaN potential work item
   Rajan & Damian: This item pertains to conversions of signaling NaNs and
   subtleties of literal “0.0”, which may take an algebraic sign. Responded to
   Jan, who’s happy. Closed.

   6.

   Editorial issues in 6.3.2.5
   Damian: This issue pertains to fine points of language in conversions
   between different floating types.
   Rajan: This is not a good case for the WG14 English language study
   group, which is focused on the wording of incoming proposals..

   7.

   Usage of Floating Point
   Damian: This item is presented for general interest. There is no action.
   Closed.

   8.

   Hexadecimal literal enhancements
   Damian: Currently, a hex floating point literal must have a trailing
   exponent field. But the string conversion functions don’t require one. We
   expect 754 to change to make the exponent optional, leading to a change in
   C. Will act after 754 takes action.

   9.

   Sign of 0 for fromfp functions
   Jerome: This is submitted, thanks to Joseph who brought it up and to
   Rajan, who commented.

   10.

   Range bounds for math functions
   Rajan & Jerome: Have responded to helpful comments from Paul.
   Damian: Be careful of the typography in the use of i.
   Action

   11.

   Totalorder changes
   Damian: Reviews improved language for Annex F change for this operation.
   Rajan: Can we remove the conjunction “and”, which has proved
   problematical in other contexts?
   Action

   12.

   Fromfp quantum exponent
   Carry over

   13.

   Floating direction in the standard
   Damian: Setting the rounding direction is subject to the optimizer
   reordering code and misplacing the intended manipulation of rounding. We
   can’t guarantee that an optimizer won’t rearrange fesetround and
   fegetround.
   Rajan: This is a WG14 issue, not a CFP action item.

   14.

   Editorial issues F.2.2
   Damian: This item pertains to functions like hypot, which can swallow
   NaNs. Fixed.

   15.

   Editorial issues F.3, etc
   Damian: This pertains to issues already covered. Done

   16.

   Editorial issues F.10.1
   Damian: This is the other of the two docs up for review.

Other issues

    None

Adjournment

    10:00 AM PDT

-----------------------------------------------

Action items to be carried over

None

New action items

Jerome: Sort out the alleged issues with nexttoward and nextafter. Report
to CFP with CC to originator of the topic.

Fred: Please provide Jerome and Damian a brief background on nexttoward. In
paper N321 from 2018, Fred refers to the example Hubert quoted in his
email. Here is N2321:

https://www.open-std.org/jtc1/sc22/wg14/www/docs/n2321.htm

Rajan & Jerome: Propose improved wording in H.11.4.1

Group: For the June meeting, make a CONCERTED EFFORT to read and ponder
Damian’s reports on Annex F.
https://cfp-wiki.esi.com.au/pub/CFP/WebHome/C2Y-F-Introduction-20260401.pdf
https://cfp-wiki.esi.com.au/pub/CFP/WebHome/C2Y-F-SpecialCases-20260401.pdf

Damian: Suggest modification F.3#1 to tighten the language of “operations”
and related terms.

Ariel: Find a footnote or other language re. FLT_RADIX not being an
operation.

Damian: Incorporate Ariel’s language about FLT_RADIX (in the action above)
into the Annex F introduction.

Fred: Add discussion of conversion of floating types in 6.3.2.5, as an open
issue.

https://mailman.oakapple.net/pipermail/cfp-interest/2026-April/003915.html

Rajan: Mention study in email 3903 to WG14.

https://mailman.oakapple.net/pipermail/cfp-interest/2026-April/003919.html

Rajan & Jerome: Review the range bounds proposal with attention to the
typography of the imaginary unit i, and submit the proposal to WG14.

Damian: Propagate Rajan’s “and” suggestion through the parallel cases to
the total order case in the discussion. Also, use the 754 wording “total
ordering”.

Carryover discussion items

None

Signoff

Respectfully submitted.

-Jerome Coonen
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.oakapple.net/pipermail/cfp-interest/attachments/20260518/4d9ee530/attachment-0001.htm>