[Cfp-interest 1899] Re: WG14 IEEE 754-C binding meeting minutes 2021/01/13

[Rajan Bhakta]

With regards to the action item below:

Rajan: Test the change in CFP 1891 with IBM's implementation.

The results are a zero quantum exponent so the change is compatible with 
what is done in practice. A couple points however: We do not yet have a 
compoundn or pown implementation provided with our offering so I could not 
check those. Also, hypot gives an INFINTY result (as expected) so the 
addition of that case in CFP 1891 seems to be in error (or I misunderstood 
what was said).

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



From:   "Rajan Bhakta" <rbhakta at us.ibm.com>
To:     cfp-interest at oakapple.net
Date:   01/13/2021 01:06 PM
Subject:        [EXTERNAL] [Cfp-interest 1898] WG14 IEEE 754-C binding 
meeting minutes 2021/01/13
Sent by:        "Cfp-interest" <cfp-interest-bounces at oakapple.net>



Attendees: Rajan, Jim, Fred, Ian, Mike, David O., Damian, David H.

New agenda items:
None

Carry over action items:
None.

Last meeting action items (all done unless specified otherwise below):
Fred: Redo updates to N2546 with the changes in CFP 1859.
Jim: Give a short summary of differences between N2579 and N2601 
(differences between TS part 3 as an Annex updates 2 and 3).
Fred: Update the example in G.5.1 using fmax to use the newer functions as 
a new proposal.
Rajan: Respond to the WG14 reflector message to say CFP wants equivalence 
to strtod and hence we don't want to parse digit separators either.
Fred: Write a WG14 editorial informational paper as per CFP 1821.
Fred: Write a CFP paper for the pow(1,NaN) and the compound(NaN, 0) case 
with respect to quantum exponent of the result.
David H: Look into numerically equivalent vs numerically equal usage in 
the C standard and revisit CFP 1849. - *Not done
Fred/Jim: Have a statement in the main body of the standard saying 
opposite signed zeros compare equal.
Fred: Still want to change negative to say less than zero in certain cases 
in C.
Jim: I will send out something to say negative zero and NaN's with a 
negative sign bit are not negative values in C.
Jim: Need to reword the signbit description in C.

New action items:
Jim/Fred: Go through all the CFP proposals submitted to ensure they were 
put in the C standard draft (N2596) correctly.
Fred: Submit CFP 1869 to WG14 with the typo fix of adding a space to the 
second last change.
Fred: Resend the document relating to updating the example in G.5.1 using 
max to use the newer functions as a new proposal.
Fred: Submit CFP 1870 to WG14.
Fred: Send out the changed text (CFP 1891) to 754.
Rajan: Test the change in CFP 1891 with IBM's implementation.
Fred: Submit the change proposed in CFP 1891 to WG14.
Jim/Fred: Make CFP 1866 into a draft WG14 proposal with the change 
"Positive zeros compare equal to negative zeros."
Fred: Change "negative signed value" to "a value with a negative sign" in 
CFP 1886 and submit it to WG14.
Jim: Make CFP 1881 into a WG14 proposal removing "with a negative sign 
bit".
Jim: Look at other uses of the phrase "sign bit" in the C standard.
Jim: Submit CFP 1879 to WG14 changing "reports" to "determines" in the 
footnote.
Rajan: Look into whether "should" can be used in an ISO standard. (Yes, 
allowed: See https://www.iso.org/foreword-supplementary-information.html)
Jim: Submit CFP 1883 as a WG14 proposal.
Jim: Get permission from WG14 to revise TS 18661-5 along the lines of what 
is proposed for TS 18661-4 in CFP 1856.
Fred: Follow up with the WG14 editor about the changes in CFP 1874.
Jim: Combine CFP 1880 and CFP 1885, and update CFP 1885 to address the 
and/or ambiguity in the first suggested change and put it in the form of a 
proposal.
Jim: Get permission from WG14 to revise TS 18661-5 along the lines of what 
is proposed for TS 18661-4 in CFP 1856.

Next Meeting(s):
Wednesday, February 17th, 2021, 4PM UTC
ISO Zoom teleconference
Please notify the group if this time slot does not work.

WG14 meeting:
Rajan: Went over the WG14 meeting results for our papers. See CFP1868.

C++ liaison:
David O.: Nothing about IEEE, but there is a C/C++ compatibility study 
group starting up (Aaron Ballman).

C23 integration
Latest C2X draft: http://www.open-std.org/jtc1/sc22/wg14/www/docs/n2
{596,573,478}.pdf also as link on CFP wiki
Part 2
Part 3
Part 4ab
Part 5abcd
IEC 60559:2020 support

Action item details
Fred: Redo updates to N2546 with the changes in CFP 1859.
See Cfp-interest 1869: DEC_EVAL_METHOD

Jim: Missing space in second last change (before "and").
Good to submit to WG14.

Jim: Give a short summary of differences between N2579 and N2601 
(differences between TS part 3 as an Annex updates 2 and 3).
See Cfp-interest 1867: AI to list latest changes in TS part 3 as annex

No issues.

Fred: Update the example in G.5.1 using fmax to use the newer functions as 
a new proposal.
*AI*: Fred to resend the document relating to updating the example in 
G.5.1 using max to use the newer functions as a new proposal.

Rajan: Respond to the WG14 reflector message to say CFP wants equivalence 
to strtod and hence we don't want to parse digit separators either.
See WG14 reflector message 18602.

No issues.

Fred: Write a WG14 editorial informational paper as per CFP 1821.
See Cfp-interest 1870: +(x)

No issues (unary + is the copy operation).
*AI*: Fred: Submit CFP 1870 to WG14.

Fred: Write a CFP paper for the pow(1,NaN) and the compound(NaN, 0) case 
with respect to quantum exponent of the result.
See Cfp-interest 1872, 1890, 1891: Quantum exponent of NaN

Jim: We generally don't like getting ahead of IEEE, but this is a case 
where it seems OK.
Mike: Yes, this is a 754 errata that will be updated.
Fred: Sent to 754's list but no response.
Mike: Took the no response as assent.
*AI*: Fred: Send out the changed text (CFP 1891) to 754.
*AI*: Rajan: Test this change (CFP 1891) with IBM's implementation.
*AI*: Fred: Submit the change proposed in CFP 1891 to WG14.

David H: Look into 'numerically equivalent' vs 'numerically equal' usage 
in the C standard and revisit CFP 1849.
Not done.
*Carry over*

Fred/Jim: Have a statement in the main body of the standard saying 
opposite signed zeros compare equal.
See Cfp-interest 1866: AI for unsigned zeros to compare equal

Mike: Should say "Positive zeros compare equal to negative zeros.".
Jim: Used "compare equal" since that is what was used elsewhere.
*AI*: Jim/Fred: Make CFP 1866 into a draft WG14 proposal with the change 
"Positive zeros compare equal to negative zeros."

Fred: Change 'negative' to say 'less than zero' in certain cases in C.
See Cfp-interest 1871, 1886, 1887: Negative

Fred: In the POSIX email list, they said -0 is a negative number. We don't 
mean that so this change. It is fine for most of the math functions, but 
printf has an issue.
David said you could say "a value with a negative sign" vs the "negative 
signed value".
Fred: My understanding is at that level of the hierarchy NaN's don't have 
a sign.
Mike: At the representation level they can have a sign bit, but that's a 
different level.
*AI*: Fred: Change "negative signed value" to "a value with a negative 
sign" in CFP 1886 and submit it to WG14.

Jim: Send out something to say negative zero and NaN's with a negative 
sign bit are not negative values in C.
See Cfp-interest 1881: say negative zero and NaN's with a negative sign 
bit are not negative values

Mike: You could remove the second sentence there. As David said, the fact 
that NaNs can have negative sign bits depends on the encoding.
Jim: I don't want to use the words "negative NaN".
Mike: Translating a NaN to a string can have a negative sign in front, 
even though it doesn't have to have a sign bit. Sign bit only applies when 
talking about an encoding.
Perhaps say "Negative zeros and NaNs are never negative values."
Jim: So remove "with a negative sign bit" from CFP 1881.
*AI*: Jim: Make CFP 1881 into a WG14 proposal removing "with a negative 
sign bit".
*AI*: Jim: Look at other uses of the phrase "sign bit" in the C standard.

Jim: Reword the signbit description in C.
See Cfp-interest 1879: AI to reword signbit description

Mike: Same issue for a negative zero not being a negative value. We should 
add in "or it is a -0" into the returns section.
Jim: None of the cases here describe -0. In the floating point model, all 
finite values have a sign. This includes -0.
Jim: For the footnote, should it be changed so reports->determines?
Seems good.
Jim: F.3 seems an awkward place to place the specification of sigbit. 
Maybe better to put it in F.10?

*AI*: Jim: Submit CFP 1879 to WG14 changing "reports" to "determines" in 
the footnote.

Other issues
fabs and copysign cleanup
See Cfp-interest 1883: fabs and copysign cleanup

Mike: For fabs, why not just say floating point values instead of floating 
point numbers?
Jim: Most descriptions for the functions talk about 'x'.
Mike: The special thing about fabs vs abs is the 'f' meaning floating 
point values.
Mike: Editorial: The change for F.10.8#1 copysign may have an extra space 
before the y in the "copysign(x, y)".
*AI*: Rajan: Look into whether "should" can be used in an ISO standard. 
(Yes, allowed: See 
https://www.iso.org/foreword-supplementary-information.html)
*AI*: Jim: Submit CFP 1883 as a WG14 proposal.

Signaling NaNs
See Cfp-interest 1874, 1875, 1876: SNAN issues

*AI*: Fred: Follow up with the WG14 editor about the changes in CFP 1874.

5.2.4.2.2 clarification
See Cfp-interest 1880, 1882, 1884, 1885: 5.2.4.2.2 clarification

CFP 1880 seems like an improvement.
Rajan: Suggested change 1 has a problem with the and/or binding leaving it 
ambiguous.
Jim: Can be changed with a colon and semicolons? Seems heavyweight.

*AI*: Jim: Combine CFP 1880 and CFP 1885, and update CFP 1885 to address 
the and/or ambiguity in the first suggested change and put it in the form 
of a proposal.

Range errors
See Cfp-interest 1841-1843, 1873: C math errors

Ian: For Issue 5, is that for subnormals?
Jim: It is for ease of use. We thought it may be too difficult to know if 
an operation underflowed or not, so we allowed them to signal underflow.

Jim: At least three separable proposals here: 3,7,4 and 1,8, and the 
remainder. I'll try to break it down for next time.

Parameterization of interfaces TS 18661 revisions
See Cfp-interest 1856: TS 18661-4 revision

Jim: For part 5, it is hard to read now with the integration of parts 1-3 
and parts of 4 into the C standard. It seems to be better as an optional 
extension rather than a stand alone document. Something like an annex.
David H: We need to get buy in to the big idea before we can get it for 
the details.
We should go ahead and try to get permission for that type of update.
*AI*: Jim: Get permission from WG14 to revise TS 18661-5 along the lines 
of what is proposed for TS 18661-4 in CFP 1856.
_______________________________________________
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/20210113/ab8d31c0/attachment-0001.htm>