From LPMeissner@msn.com Mon Sep 30 21:00:15 1996 Received: from ratatosk.DK.net (root@ratatosk.DK.net [193.88.44.22]) by dkuug.dk (8.6.12/8.6.12) with ESMTP id UAA22535 for ; Mon, 30 Sep 1996 20:58:59 +0100 Received: from upsmot02.msn.com (upsmot02.msn.com [204.95.110.79]) by ratatosk.DK.net (8.6.12/8.6.12) with ESMTP id UAA24545 for ; Mon, 30 Sep 1996 20:58:56 +0100 Received: from upmajb06 ([204.95.110.89]) by upsmot02.msn.com (8.6.8.1/Configuration 4) with SMTP id MAA14768; Mon, 30 Sep 1996 12:01:29 -0700 Date: Mon, 30 Sep 96 19:54:45 UT From: "Loren Meissner" Message-Id: To: keith.bierman@sun.com Cc: d.muxworthy@ed.ac.uk, jkr@letterbox.rl.ac.uk, sc22wg5@dkuug.dk Subject: Request for clarification I am trying to convert the Preliminary WG5 minutes that I distributed in early August to a final version (subject to approval by WG5, of course). I see that there has been some email discussion of the intent of one of the straw votes on floating point exception handling. In my notebook at the meeting I wrote: ================================== In case of Integer Overflow during conversion from decimal string, Processor must signal "IEEE INVALID" or may generate a processor- dependent signal different from other exceptions - must signal IEEE INVALID yes 12; no 6 (Doc N1195R1 p. 15 "IEEE INVALID") ================================== In some email around 9 August, John Reid asks if it should be "during conversion to an integer". Keith Bierman says yes, "the topic in question is the specific conversion of floating point to integer conversions". David Muxworthy says (apparently in response to John, not Keith) it was about "overflow at R to I conversions (not the other possibilities mentioned under INVALID on page 15 of N1195r1). The vote was about whether "may" or "must" IEEE_INVALID be signalled on R to I overflow . . ." My problem now is whether we are talking about a DECIMAL STRING as I have it in my original notes - - I think that would be referring to conversion DURING INPUT from a string of decimal digits in an external field to an internal integer - - or is it talking about converting an internal real number to an internal integer as David seems to suggest? - Loren Meissner