Bugs item #678034, was opened at 2003-01-31 10:19 Message generated for change (Settings changed) made by sjoerd You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=482468&aid=678034&group_id=56967
Please note that this message will contain a full copy of the comment thread, including the initial issue submission, for this request, not just the latest update. Category: Modules Group: (zombie: MonetDB 4.3 CVS Head) >Status: Closed >Resolution: Out of Date Priority: 1 Private: No Submitted By: Sjoerd Mullender (sjoerd) Assigned to: Sjoerd Mullender (sjoerd) Summary: NaN / Inf test on Windows give wrong result Initial Comment: The tests of the isnan / isinf / finite functions on Windows give incorrect results. The reason (or in any case, one reason) is that the method used to obtain a NaN doesn't work on Windows. The test uses log(-1.0) to obtain a NaN, however this doesn't seem to do the trick. Use the attached program to obtain the below results: Cygwin: + ./a.exe 0 0: isnan 0 isinf 0 finite 1 fpclassify 2 0000000000000000 NAN NaN: isnan 1 isinf 0 finite 0 fpclassify 0 000000000000f87f z/z NaN: isnan 1 isinf 0 finite 0 fpclassify 0 000000000000f8ff log(-1.0) -Inf: isnan 0 isinf 1 finite 0 fpclassify 1 000000000000f0ff pow(2.0, 5000.0) Inf: isnan 0 isinf 1 finite 0 fpclassify 1 000000000000f07f pow(-2.0, 5001.0) -Inf: isnan 0 isinf 1 finite 0 fpclassify 1 000000000000f0ff Native Windows: C:\Documents and Settings\sjoerd\My Documents>x.exe 0 0: isnan 0 isinf 0 finite 1 fpclassify 64 0000000000000000 z/z -1.#IND: isnan 1 isinf 0 finite 0 fpclassify 2 000000000000f8ff log(-1.0) -1.#IND: isnan 1 isinf 0 finite 0 fpclassify 2 000000000000f8ff pow(2.0, 5000.0) 1.#INF: isnan 0 isinf 1 finite 0 fpclassify 512 000000000000f07f pow(-2.0, 5001.0) -1.#INF: isnan 0 isinf -1 finite 0 fpclassify 4 000000000000f0ff ---------------------------------------------------------------------- >Comment By: Sjoerd Mullender (sjoerd) Date: 2009-10-02 18:07 Message: The test output has long since been changed because the method used to produce the NaNs produces a range error instead. If I create a NaN in C and use that to test the mmath function isnan() it correctly reports that the value is a NaN. Therefore I now close this bug. ---------------------------------------------------------------------- Comment By: Niels Nes (nielsnes) Date: 2005-10-06 13:29 Message: Logged In: YES user_id=43556 BugDay_2005-10-06: Claimed by niels BugDay_2005-10-06: TEST exists Test exists as src/modules/plain/Tests/mmath.milM ---------------------------------------------------------------------- Comment By: Niels Nes (nielsnes) Date: 2004-03-11 08:54 Message: Logged In: YES user_id=43556 windows short comming ---------------------------------------------------------------------- Comment By: Sjoerd Mullender (sjoerd) Date: 2003-01-31 10:22 Message: Logged In: YES user_id=43607 I did attacht the file, and I did check the checkmark, but somehow the file was still not attached. Bloody Windows or bloody Mozilla? ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=482468&aid=678034&group_id=56967 ------------------------------------------------------------------------------ Come build with us! The BlackBerry® Developer Conference in SF, CA is the only developer event you need to attend this year. Jumpstart your developing skills, take BlackBerry mobile applications to market and stay ahead of the curve. Join us from November 9-12, 2009. Register now! http://p.sf.net/sfu/devconf _______________________________________________ Monetdb-bugs mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/monetdb-bugs
