JTC1/SC22/WG14
N885
Draft Minutes for Jun 21-25 1999
Meeting of ISO/JTC1/SC22/WG14 and NCITS/J11
WG14/N885 J11/99-020
21 Jun 98 09:00-12:00 13:00-17:00
22 Jun 98 09:00-12:00 13:00-17:00
23 Jun 98 09:00-12:00 13:00-17:00
24 Jun 98 09:00-12:00 13:00-15:30
BSI
389 Chiswick High Road
London W4 4AL
England
Derek Jones
derek@knosof.co.uk
Voice: +44 252 520667
------------------------------------------------------------
Monday Feb 1st
1. Opening activities
1.1 Opening Comments
1.2 Introduction of Participants
Benito, John Perennial WG14, Convener, Chair
Farance, Frank Farance,Inc J11
Feather, Clive BSI WG14, UK HOD
Finch, Tony Demon Internet WG14, UK
Fukutomi, Hiroshi ADACS WG14, Japan HOD
Jaeschke, Rex Self J11
Jones, Derek Knowledge Software WG14, UK, Secretary
Jones, Larry SDRC J11, Project Editor
Keaton, David Self J11
Leca, Antoine Self WG14, France HOD
MacDonald, Tom SGI J11
Maclaren, Nick U of Cambridge WG14, UK
Mak, Raymond IBM J11, WG14, Canada
Muller, Jeff Oracle J11
Parks, John Compqa/Digital J11
Schwab, David Oracle J11
Seymour, Bill Self J11, Rationale Editor
Simonsen, Keld RAP J11 guest
Tydeman, Fred Tydeman Consulting J11,
Wakker, Willem ACE WG14, Netherlands HOD
Walls, Douglas Sun J11, USA HOD
1.3 Host Facilities/local information
Geoff Williams introduced BSI and the facilities.
1.4 Procedures for this Meeting
1.5 Approval of Previous Minutes [N859]
N882 - Final Minutes Portland (Walls)
Approved No changes.
1.6 Review of Action Items and Resolutions
(Open) Meyers: Write rationale on 64-bit with ideas on how to
reduce code size if not used.
(Closed) Keld Simonsen will coordinate with POSIX, WG20, Single
Unix on localeconv.
N884
(Closed) Tom MacDonald: Rationale for this change to FOR (and
related).
(Open) Seebach craft words for rationale why we did not adopt
conformance requirements in to C9X.
(Closed) Keaton+Larry to come up with words for why
Number of arguments no change needed.
(Open) Meyers: Produce rationale for macro replacement
(Open) Seebach: Provide strsep() rationale.
(Open) Jones - Provide strftime() rationale.
(Open) Meyers - %m rationale.
(Closed) Clive, Provide London (6/99) meeting info.
(Closed) Rex, Needs to provide responses to individuals
submitting U.S. comments on the FCD
(Closed) All, Resolutions of all comments needed by end
this meeting in electronic form.
(Closed) Walls, Request of the SC22 tag how the US position
on ISO/14652, the Standard on cultural conventions is being formed.
J11 would like to express an opinion. WG14 has expressed to SC22
they use this work.
(Close) MacDonald rationale side-effects in VLA
(Close) MacDonald rationale for static added to array parameter syntax
1.7 Approval of Agenda [N864]
Agenda modified and agreed.
1.8 Distribution of New Documents
No new documents
1.9 Information on Next Meeting
Next meeting: Kona, Hawaii
ACTION Plum: Forward meeting location information into post London mailing
ACTION Benito: Liaison to decide length of Hawaii meeting
1.10 Identification of National Bodies/J11 voting members
WG14 National bodies: U.K. U.S., Canada, Japan, Netherlands, France
J11 present 12 members, 10 eligible to vote out of 16 with voting privileges.
2. Project Editor's Report [N???] (Jones)
Forwarded document to SC22. Will have to send a final text once voting
is complete. ITTF has not yet started the FDIS ballot.
ACTION Jones(L): Call for editorial reviewers, post ITTF ballot
3. Rational Editor's Report (Seymour)
A new document, N881, has been posted to the WG14 site. There are
some missing items that will be addressed at this meeting.
5a. FCD Status Report (Benito)
Document was forwarded to ITTF on April 27. We are waiting for
the start of the 2 month, JTC level, ballot.
ACTION Farance: Try to find out start date of FDIS ballot and post
it to the WG14 reflector.
Work items relating to the FDIS
Identified National body comments that were causing them to vote No.
Such comments had been posted to the WG14 reflector by Japan, Canada
and the UK.
UK comments, email ref ?
size_t to be represented in a type no larger than unsigned long.
Could the issues be addressed via an Amendment?
Leaves open a window of opportunity between publishing
the standard and the amendment for implementors to make use
of freedom available under the current wording.
Canada supports this issue as a recommended practice.
fenv issues. There was general agreement that the changes
suggested by the comment were an improvement on the current
specification.
Canadian comments, email ref 7280
Conversion of signed integer types. General agreement on
the substance of the comments. No hard proposals on how
to progress at this time.
long long issues. There was general discussion.
[static]. The Canadians felt that there had not been sufficient
time to digest the new syntax/semantics added in Portland.
They also thought that there were better, more general
ways of specifying these requirements. There was no support for
an alternative proposal.
Also there is no support to remove the current syntax.
4. Liaison Activities
4.1 J11 + ANSI (C)
No news from J11.
4.2 WG14 + ISO/SC22 (C)
There is a SC22 plenary September in Berlin. Benito will be representing
WG14.
4.3 J16/WG21 (C++)
WG21 met in Dublin in March. Started work on Technical Report to address
performance issues. Want to make C++ more effective in the embedded
systems market.
ACTION Benito: Post high performance C++ Reflector address to WG14
reflector.
4.4 WG15 (POSIX)
The WG15 liaison was not present to give a report.
4.5 WG20 (I18N)
No news from WG20.
4.6 Other Liaison Activities
(Wakker) WG11 has issued an elementary functions CD. Plan to work
on a CD covering complex functions.
Japanese comments
ACTION Maclaren: Write rationale for the isblank function.
ACTION Jones(L): To correct the names of the div family of functions
in the synopsis.
tmpnam. It was agreed that the second possibility, given in the comment,
expressed the intent. This comment will be used as the basis of text
that will be added to the Rationale.
Rationale 64k wording. Rationale editor agreed that such wording had been
agreed and that he would add it.
Rationale for long long. Agreed to craft some words for review
tomorrow.
Wakker presented DSP C (N854). He would like fixed pointed data types
to be added to the next revision of the C standard.
Work items for the rationale
ACTION Maclaren,Tydeman,MacDonald: Provide some rationale for math error
handling.
ACTION Feather: Write rationale 6.2.6, 6.2.5 wording for representation
of values.
ACTION MacDonald: Write rationale VLA <-> longjmp interaction.
Rex is retiring as convenor of J11. He was presented with a token of our
respect.
Tuesday 22 June
Morning
Review of Rationale
The committee reviewed the rationale on a page by page basis up to
page 38.
ACTION Thomas: 6.3.1.6, 6.3.1.7 write some Rationale.
ACTION Meyers: Rationale for rvalue arrays in 6.3.2.1.
Afternoon
Broke out into small groups to draft wording for various outstanding points.
ACTION Muller: Produce rationale wording to clarify paragraph
starting at page 47, line 36.
The committee reviewed the rationale on a page by page basis from
page 39 to 67.
Wednesday 23 June
Morning
A number of delegates whose first language was not English were leaving
today. To make use of their experience it was decided to review the
Multibyte Support Extensions Rationale today.
The committee reviewed the rationale on a page by page basis from
page 155 to 172.
Another pass was made over the National body comments.
Japan
long long. There are cases where the long long type may appear in an
existing C90 program that is compiled using a C9X compiler.,
The decimal constant literal 3000000000 has type unsigned long in C90,
it has type long long in C9X.
It was agreed that wording would be added to the Rationale listing this
case as a quiet change.
ACTION Jones(L): Draft words to describe the tmpnam issue to add to
the Rationale.
UK
long long. Could this issue been be handled in a TC? The UK were not
certain that publication would be sufficiently close to the publication
of the IS to prevent implementors making use of the flexibility offered
by C9X. WG14 could not offer any assurance that a TC would be
published to a given schedule or even gather enough support to be
published.
fenv.
Afternoon
Straw vote of NBs on what their likely support would be if the
given item were submitted via the DR machinery for publication
as a Technical Report.
UK issues
long long: YES 2, NO 1, ABSTAIN 2
fenv: YES 4, NO 0, ABSTAIN 1
Canadian issues
Conformance: YES 1, NO 1, ABSTAIN 3
long long: YES 1, NO 1, ABSTAIN 3
[static]: YES 1, NO 1, ABSTAIN 3
The committee reviewed the Rationale on a page by page basis from
page 68 to 91.
16. Future Work Items
Items identified:
o Architecture Support Extensions Group: DSP-C, etc
o Conformance
o Sequence Points
o time
A schedule for the work items and a charter will be considered at the
next meeting of WG14.
Additional proposals must be sponsored by a National Body.
Broke into subgroups to craft words for the Rationale.
Thursday 24 June
Morning
The committee reviewed the rationale on a page by page basis from
page 92 to 128.
Afternoon
The committee reviewed the rationale on a page by page basis from
page 129 to 154. This completed the pass over the Rationale by
the committee of the whole.
Keld presented N884 for discussion.
ACTION Keld: Edit the abbreviated date/time given in N884 to be
compatible with C9X.
ACTION Keld: To report to WG15 that "WG14 wishes to have their C
locale registered by the appropriate registration authority. It
notes that there are differences between the C and POSIX locale."
ACTION MacDonald: Extract Rationale for signaling NANs in NCEG TR.
ACTION MacDonald: Rationale for C9X tag compatibility change.
24. Administration
24.1 Future Meetings
24.1.1 Future Meeting Schedule
Oct 18-22, 1999 Hawaii Plum Hall
Apr ??-??, 2000 Japan
Oct ??-??, 2000 ??????
24.1.2 Future Agenda Items
DR processing. The convenor requested that NB's filter DRs before
submission.
Technical Report, see agenda item 16.
24.1.3 Future Mailings (sponsor for WG14 mailings)
post London deadline Jul-10, 1999
pre Hawaii deadline Sep-18, 1999
Submit input for rationale revision by Aug-21, 1999
24.2 Resolutions
None
24.2.1 Review of Decisions Reached
None
24.2.2 Formal Vote on Resolutions
None
24.2.3 Review of Action Items
ACTION Plum: Forward meeting location information into post London mailing
ACTION Benito: Liaison to decide length of Hawaii meeting
ACTION Jones(L): Call for editorial reviewers, post ITTF ballot
ACTION Farance: Try to find out start date of FDIS ballot and post
it to the WG14 reflector.
ACTION Benito: Post high performance C++ Reflector address to WG14
reflector.
ACTION Maclaren: Write rationale for the isblank function.
ACTION Jones(L): To correct the names of the div family of functions
in the synopsis.
ACTION Maclaren,Tydeman,MacDonald: Provide some rationale for math error
handling.
ACTION Feather: Write rationale 6.2.6, 6.2.5 wording for representation
of values.
ACTION MacDonald: Write rationale VLA <-> longjmp interaction.
ACTION Thomas: 6.3.1.6, 6.3.1.7 write some Rationale.
ACTION Meyers: Rationale for rvalue arrays in 6.3.2.1.
ACTION Keld: Edit the abbreviated date/time given in N884 to be
compatible with C9X.
ACTION Keld: To report to WG15 that "WG14 wishes to have their C
locale registered by the appropriate registration authority. It
notes that there are differences between the C and POSIX locale."
ACTION MacDonald: Extract Rationale for signaling NANs in NCEG TR.
ACTION MacDonald: Rationale for C9X tag compatibility change.
ACTION Items for the Rationale wording
----------------------------------------------------------------------
Feather :
pp.41 para at line 31, re concatenation of wide and narrow strings.
pp. 67, add words re inline.
pp. 100, 7.4.1 add table of relationships w.r.t is* functions.
explain/introduce new terms we added in Portland (indeterminate, etc.)
----------------------------------------------------------------------
MacDonald:
pp. 64 6.7.3.1, add (old) example re restrict per Raymond Mak's request.
pp. 98, line 98, C9X reverts to C89 behavior w.r.t assert's accepting any
scalar type.
pp. 116, line 35-37, improve wording.
pp. 151, lines 9-10, imaginary types as args, is this correct?
----------------------------------------------------------------------
Jones, Larry:
Quiet change re the type of sizeof.
pp. 68-69, replacement sentence for struct hack.
pp. 94, why const was put in library prototypes in C89.
pp. 116, line 44.
pp. 136, realloc of zero no longer necessarily frees memory.
----------------------------------------------------------------------
Benito:
pp. 98, line 19-20, 6-character limit increase C95, C9X.
write up intimax, atoll, and others.
----------------------------------------------------------------------
Tydeman/Thomas:
pp. 111, various new math functions, e.g., log1p, log2.
pp. 110-111, tan, change from C89, existing rationale disagrees with
standard.
----------------------------------------------------------------------
Muller, Jeff:
pp. 135, line 18-22, improve wording.
----------------------------------------------------------------------
Farance:
new functions in inttypes.h.
----------------------------------------------------------------------
Unassigned:
pp. 129, 7.19.6.5, snprintf family, explain why there are differences
between the narrow and wide char versions (buffering?)
----------------------------------------------------------------------
24.2.4 Thanks to Host
BSI were thanked for hosting the meeting.
24.2.5 Other Business
25 Adjournment
The meeting was adjourned at 15:30
J11 Tag meeting
Thursday 24 June 1999, 14:00-14:20
Appoint delegation and HOD for future WG14 meetings
Delegation for October 1999 meeting
Farance
Jones
Seymour
Tydeman
Walls, HOD
Officer vacancies
Chair. The current chair, Jaeschke, has decided not to stand
again. MacDonald has let his name go forward for this post.
International Representative, IR. The current representative,
Walls, has submitted his name to continue in the post.
AOB
The outgoing Chair was universally acclaimed.