<html><head><meta http-equiv="Content-Type" content="text/html; charset=utf-8"></head><body style="word-wrap: break-word; -webkit-nbsp-mode: space; line-break: after-white-space;" class=""><br class=""><div><br class=""><blockquote type="cite" class=""><div class="">On Aug 25, 2019, at 6:11 PM, Fred J. Tydeman <<a href="mailto:tydeman@tybor.com" class="">tydeman@tybor.com</a>> wrote:</div><br class="Apple-interchange-newline"><div class=""><div class="">Thanks for reviewing.<br class=""><br class="">On Sun, 25 Aug 2019 15:15:34 -0700 Jim Thomas wrote:<br class=""><blockquote type="cite" class=""><br class=""><blockquote type="cite" class="">Some implementations consider f(infinity) as a range error based upon the text: A range error occurs if [the magnitude<br class=""></blockquote>of] x is too<br class=""><blockquote type="cite" class=""><br class=""></blockquote>Consider changing to: "Some implementations incorrectly treat ".<br class=""></blockquote><br class="">Done.<br class=""><br class=""><blockquote type="cite" class=""><blockquote type="cite" class="">The exp10m1 functions<br class="">Change 7.12.6.3#2 from<br class=""><br class="">A range error occurs if finite x is too large or if the magnitude of nonzero x is too small.<br class="">to<br class=""><br class="">A range error occurs if finite positive x is too large or if the magnitude of nonzero x is too small.<br class=""></blockquote>I recall David suggesting this be "A range error occurs if the magnitude of finite positive x is too large ." but am not sure<br class="">what we agreed to.<br class=""></blockquote><br class="">I also do not recall what we agreed to.<br class=""></div></div></blockquote><div><br class=""></div><div>As Rajan points out, the resolution is in the minutes. </div><div><br class=""><blockquote type="cite" class=""></blockquote></div><blockquote type="cite" class=""><div class=""><div class=""><br class=""><blockquote type="cite" class="">With that change we'd be saying "magnitude" wherever talking about too large or too small. Without<br class="">that change we'd be saying "magnitude" just where needed for correctness.<br class=""></blockquote><br class=""><blockquote type="cite" class=""><blockquote type="cite" class="">The compoundn functions<br class="">Change 7.12.7.2#2 from<br class=""><br class="">A range error may occur if n is too large, depending on x.<br class="">to<br class=""><br class="">A range error may occur if n is too large, depending on finite x.<br class=""></blockquote>Do we mean if the magnitude of n is too large?<br class=""></blockquote><br class="">Yes. Changed.<br class=""><br class=""><blockquote type="cite" class=""><blockquote type="cite" class="">The tgamma functions<br class="">Change 7.12.8.4#2<br class=""><br class="">A range error occurs if the magnitude of x is too large and may occur if the magnitude of x is too small.<br class="">to:<br class=""><br class="">A range error occurs if the magnitude of finite x is too large and may occur if the magnitude of non-zero x is too small.<br class=""></blockquote>How is there a range error if the magnitude of negative x is too large?<br class=""></blockquote><br class="">With the FP systems still in use that I know about, all large FP values are integers. So, large negative values would<br class="">be integers, so would be domain or pole errors.<br class=""><br class=""><blockquote type="cite" class="">Can't there be range errors for moderate magnitude<br class="">negative x?<br class=""></blockquote><br class="">Yes. I think "moderate negative x" could still be considered "too large". But, I am open to<br class="">other wording.<br class=""></div></div></blockquote><div><br class=""></div><font size="2" class="">How about: "A range error occurs if the magnitude of finite positive x is too large and for some negative finite x and …”?<br class=""></font><blockquote type="cite" class=""></blockquote><blockquote type="cite" class=""><div class=""><div class=""><br class=""><blockquote type="cite" class="">A general question: What does "may occur" mean?<br class=""></blockquote><br class="">Does not always happen. For example, on IBM S/360 hex FP, there are no subnormal numbers,<br class="">so underflows will not happen for many functions, that do happen when subnormals are supported.<br class=""></div></div></blockquote><div><br class=""></div><div>Do we have consistent wording around this? sinpi has "A range error occurs if the magnitude of nonzero x is too small”, but range errors occur for sinpi only for subnormal input. I thought this was ok, because if the implementation didn’t support subnormals, |x| would never be too small and the statement would be vacuously true.</div><div><br class=""></div><div><blockquote type="cite" class=""></blockquote></div><blockquote type="cite" class=""><div class=""><div class=""><br class="">An alternative wording: A range errors occurs for some finite arguments.<br class=""></div></div></blockquote><div><br class=""></div><div>That approach might be better in some cases. Instances of “may occur” might need more consideration.</div><div><br class=""></div><div class="">- Jim Thomas</div><blockquote type="cite" class=""><div class=""><div class=""><br class=""><br class="">---<br class="">Fred J. Tydeman Tydeman Consulting<br class=""><a href="mailto:tydeman@tybor.com" class="">tydeman@tybor.com</a> Testing, numerics, programming<br class="">+1 (702) 608-6093 Vice-chair of PL22.11 (ANSI "C")<br class="">Sample C99+FPCE tests: <a href="http://www.tybor.com" class="">http://www.tybor.com</a><br class="">Savers sleep well, investors eat well, spenders work forever.<br class=""></div></div></blockquote></div><br class=""></body></html>