[Cfp-interest 2508] Re: Fwd: Editorial review

[Fred J. Tydeman]

On Thu, 4 Aug 2022 20:25:55 -0700 Jim Thomas wrote:
>
>As you find errors please email them to cfp-interest at ucbtest.org <mailto:cfp-interest at ucbtest.org>.

N3047_draft: Fixes(?) still to be applied:
References to 'pages' are to 'printed pages'.

Page 25, #21:  "The signaling NaN macros ... each is defined" -> 
       "Each of the signaling NaN macros ... is/are defined"

Page 227: 7.12 <math.h>
The types
 _Decimal32_t, _Decimal64_t
are not listed as being optional 
(depending upon __STDC_IEC_60559_DFP__)
OK on page 484 B.11 <math.h>

Page 227: 7.12 <math.h>
The macros
 HUGE_VAL_D32, HUGE_VAL_D64, HUGE_VAL_D128
are not listed as being optional 
(depending upon __STDC_IEC_60559_DFP__)
OK on page 484 B.11 <math.h>

Page 261, 7.12.11.5 The nextup functions
Change
 If x is the positive number (finite or infinite) or maximum magnitude in
to
 If x is the positive number (finite or infinite) of maximum magnitude in

Page 262, 7.12.11.7 The canonicalize functions
In the Synopsis, three of the 'cx' are missing the '*'.

Page 263: 7.12.12.1 The fdim functions
 A range error may occur.
would be better as:
 A range error occurs for some finite arguments.

Page 467: Bottome of page: I see two lines on top of each other.

Page 604: What are '__suffix__' and '__prefix__' that only appear in
J.6.1 Rule based identifiers and the index?

Pages 671...: There are several suffix terms in the index that are
listed as prefix terms. One example is _DECIMAL_DIG.
Also, these index entries (at least):
 _DECIMAL_DIG identifier prefix, 29, 329, 352,
 _DIG identifier prefix, 550, 601
 _EPSILON identifier prefix, 474, 551, 601
 X_DECIMAL_DIG identifier prefix, 550, 615
 X_DIG identifier prefix, 550, 615
 X_EPSILON identifier prefix, 551, 615
should be changed to suffix.

Page 692: There are 2 index entries for IEC 60559.  I assume that they
should be merged together.

================
Maybe:

Page 261: Add: nextafter: If x!=y, a range error occurs for either
subnormal or zero results.

Page 267: Add: fma:  A domain error occurs for some infinite arguments.
================
Unknown:

Page 270, 7.12.15.3 The quantumdN functions
What do those functions do with a NaN argument?

In many places in <math.h>:
rounded to the type of the function. -> 
rounded to the return type of the function.



---
Fred J. Tydeman        Tydeman Consulting
tydeman at tybor.com      Testing, numerics, programming
+1 (702) 608-6093      Vice-chair of PL22.11 (ANSI "C")
Sample C99+FPCE tests: http://www.tybor.com
Savers sleep well, investors eat well, spenders work forever.