[Cfp-interest 1795] WG14 October 2020 meeting CFP paper results
Rajan Bhakta
rbhakta at us.ibm.com
Mon Oct 19 12:36:30 PDT 2020
Summary: All papers passed. There were some issues with some papers that
need follow up however, detailed below with a *.
N2546: Addition of DEC_EVAL_METHOD
Joseph: Also can look at other places with similar issues.
Straw poll: Put N2546 into C23?
20/0/1. Consensus.
N2547: Missing const in decimal payload functions
Goes into C23.
N2548: intmax_t and math functions
*Joseph: Part 3 needs these changes as well.
Straw poll: Put N2548 into C23?
17/0/6. Consensus.
N2564: Range errors and math functions
Accepted as is.
N2552: Editorial cleanup
Accepted as is.
N2558: Annex B with prototype forms
Alex: Love it. Like generic. Want this in more places.
Jens: Agree with liking this, but concern is that Annex B is
automatically extracted from the standard. This change makes it harder for
auto-generation. If you want this, the change should be made to the
standard itself, not just annex B.
*Joseph: Hasn't been updated for the new functions like the max/min, and
other typo's like remquo, nexttoward lists all the types but doesn't
include all the types. A number of other mistakes like that.
Having annotations for the subsets is needed.
Aaron Ballman: Using N and M and MS is awkward. Perhaps use something
different.
We (CFP) discussed this, and wanted consistency with Part 3, but happy
to change if JeanHyde says typeface here doesn't work. - It does work for
JeanHyde.
*Fred: nexttoward does take decimal too. We (CFP) also noticed missing
functions in the existing Annex B.
*David S: This rewrite seems to apply to things like Atomics and other
things too. When I write my paper, I'll send it to CFP for review.
Straw poll: Does the committee want something along the lines of N2558
in Annex B?
11/4/8. Direction to continue along these lines.
Straw poll: Does the committee want something along the lines of N2558
in the main body of the standard (that is reworking the main body to have
this type parameterization)?
11/2/8. Clear direction.
From chat:
Joseph: suggest not using StdType for fadd, dadd etc. (given fadd is
just two functions, dadd just one). It makes sense for fMadd etc,
however.
JeanHyde (private): I love this change and would like to make it
happen, for what it's worth. It might take some effort but I'll be happy
to do my best to try it out in places and demo what it looks like over the
next few months!
Fred: Should CFP also do the complex functions similar to the real
functions for Annex B?
JeanHyde: I think we should to the whole standard, honestly. :D
(But, we can start with Math.h and the FP functions first, see how that
looks and what works best, and then go from there.)
N2559: Update to IEC 60559 2020
*Joseph: Bibliography needs to be updated. Other normative references
like fmax need to be updated. Want to keep the references to the older
versions (like Annex F nextafter in F.3#8) (and add them to the
bibliography). Add the word superseded in the reference to the IEEE 2011
draft.
Straw poll: Do we want N2559 into C23?
15/0/5. Consensus.
N2560: FP hex formatting precision
Straw poll: Do we want N2560 into C23 as is (Footnote form)?
10/1/9. Consensus.
N2561: TS 18661-3 annex update
See N2579 for latest version.
*Joseph: Further updates were in the reflector like things for SNAN,
intmax_t, etc. need to be part of the vote.
What is the intention of the float.h macros in all 3 sets. Not
mentioned in this annex? Ex. hassubnorm, min/max normalized macros. Should
they be added in? New min/max functions should be part of this annex.
Straw poll: Put N2561 into C23 knowing future changes will be needed?
10/3/7. Consensus.
From JeanHyde (private): (For what it's worth, I would have
automatically applied such changes for consistency anyways, just seeing it
as an editorial change to make it consistent by working with Jim Thomas to
get to that point. And without requiring CFP to submit new papers to fix
things that they already sent papers to fix.)
N2570: Feature/want macros for Annex F functions
Straw poll: Does the committee want to address the issue given in N2570?
12/0/6. Consensus.
Straw poll: Does the committee want to keep the WANT macro as described
in N2570?
4/6/8. Not clear direction.
Straw poll: Does the committee want two separate frames as described in
N2570?
4/2/12. No clear direction.
Straw poll: Does the committee want a single frame as described in N2570
example 2?
6/0/11. Clearer direction.
Straw poll: Does the committee want to adopt a single frame as described
in N2570 example 2 into C23?
6/0/12. Clearer direction, but large abstentions.
Reason for abstentions: Lack of domain knowledge, underwhelmed.
Regards,
Rajan Bhakta
z/OS XL C/C++ Compiler Technical Architect
ISO C Standards Representative (Canada, USA), PL22.11 Chair
C/C++ Compiler Development
rbhakta at us.ibm.com
IBM
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.oakapple.net/pipermail/cfp-interest/attachments/20201019/a986e6c0/attachment.htm>
More information about the Cfp-interest
mailing list