On Tue, Oct 07, 2014 at 04:53:44PM +0000, Andreas Gustafsson wrote: > Module Name: src > Committed By: gson > Date: Tue Oct 7 16:53:44 UTC 2014 > > Modified Files: > src/tests/lib/libm: t_exp.c > > Log Message: > In the exp2_values test case, provide separate expected return values > for the float case, reflecting the actual exp2f() argument value after > rounding to float precision. Fixes PR lib/49256. Thanks to Makoto > Kamada and Tetsuya Isaki for the analysis.
The reason I left the tests failing is that the results should be more accurate than the values that are actually returned. Changing the 'expected' values so that the tests pass is just wrong. What I should have got araound to doing is using the x87 instruction to generate accurate 64bit mantissa (long double) results for the values, and then looked at the sizes of the error terms. I think the values ought to be accurate to one or two counts on the lsb of the mantissa. David -- David Laight: da...@l8s.co.uk