[Cfp-interest] C Standard meeting 18661 DR's result

[Rajan Bhakta]

TS 18661-5 Update [N 2004] and Slide Deck [N 2005]
  Can go directly to DTS instead of PDTS.
  No objection to going to DTS or to ballot.


  Adding Fundamental Type for short float [N 2016]
    Use Part 3 type with the encoding/conversion functions to get any 
other type.
    Not restricting sizes makes it useless for actual computations.
    5.3) Without restricting sizes, you can't be sure.
    5.5) Can get different results by conversion to float for some 
functions vs keeping short float.
 
    Requiring conversion from short float to float does not make sense 
since there could be some values that aren't representable as floats (more 
accurate as short float).
    Jens:
    - Don't like the syntax (multiple keywords for a type).
    - Are these required by C or optional?
    - Implications with complex?
    - What about tgmath?
    Response: Easy to implement a wrapper.
    Derek: Adding information in float.h for the limits would make it 
useful.
    _Float16 is not defined as per float and double. This proposal allows 
this.
    Jens: Have extended floating point types.
    Andrew: Since everything uses the 'half' keyword, why use 'short 
float'?
    Was _Float8x considered? Not IEEE so no.
    float is not good enough since 5.2.4.2.2 requires at least 6 digits.
    Willem: How about making it a TS?
    Straw poll: Should WG14 continue along the lines of N2016? 18/1/5. 
Consensus to move forward with this.
    Paper solicited from the authors.


Regards,

Rajan Bhakta
z/OS XL C/C++ Compiler Technical Architect
ISO C Standards Representative for Canada
C Compiler Development
Contact: rbhakta at us.ibm.com, Rajan Bhakta/Houston/IBM



From:   Rajan Bhakta/Houston/IBM at IBMUS
To:     cfp-interest at oakapple.net
Date:   04/11/2016 03:01 PM
Subject:        [Cfp-interest] C Standard meeting 18661 DR's result
Sent by:        cfp-interest-bounces at oakapple.net



Hi,

In the C meeting, the review of the proposed defects and resolutions for 
the IEEE 754 C binding Technical Specification came up with the following:

  TS 18661 DR's: [N 2029]
    TS18661-1:
    1) Marked as DR1-1
    2) Marked as DR2-1
    3) Editorial. Can be published with a TC.
    4) Marked as DR3-1
 
    TS18661-2:
    1) Editorial.

    TS18661-3:
    1) Editorial
    2) Editorial
    3) Marked as DR1-3
 
Result: Accepting all Suggested Technical Corrigendum's as Proposed 
Technical Corrigendum's.

I will be presenting Part 5 right after this current discussion of N2016 
(the half float type proposal) and will let you know how it goes.

Regards,

Rajan Bhakta
z/OS XL C/C++ Compiler Technical Architect
ISO C Standards Representative for Canada
C Compiler Development
Contact: rbhakta at us.ibm.com, Rajan Bhakta/Houston/IBM
_______________________________________________
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/20160411/c7414e86/attachment.html