[Cfp-interest 1944] Fwd: (SC22WG14.19047) Preferred quantum exponent for default-initialized 0

[Fred J. Tydeman]

..................Begin Forwarded Message..................

From: Joseph Myers <joseph at codesourcery.com>
To: <sc22wg14 at open-std.org>
Date: Wed, 10 Mar 2021 22:17:15 +0000
Subject: (SC22WG14.19047) Preferred quantum exponent for default-initialized 0

This is an issue that I've previously mentioned, but in the middle of 
threads on other subjects so it may have been missed.  Could the 
floating-point group look at it and come up with a proposal for C23?

The rule (6.7.9#10) that "If an object that has static or thread storage 
duration is not initialized explicitly, then ... if it has arithmetic 
type, it is initialized to (positive or unsigned) zero" does not specify 
the quantum exponent of the result in the decimal floating-point case.

In practice, the default initialization is to representations with all 
bits 0, and in the IEEE decimal floating-point interchange encodings (both 
binary and decimal) that's a zero with the least possible quantum 
exponent.  So I think it would be appropriate for the table of preferred 
quantum exponents in 5.2.4.2.3 to have an entry added for default 
initialization, saying the preferred quantum exponent is least possible.

-- 
Joseph S. Myers
joseph at codesourcery.com

...................End Forwarded Message..................


---
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.