https://issues.apache.org/bugzilla/show_bug.cgi?id=43940
--- Comment #40 from Glenn Adams <[email protected]> --- (In reply to comment #39) > (In reply to comment #38) > > (In reply to comment #36) > > > The bug is trivial: > > > "if (source >= 10e-3 && source < 1e7) {" is wrong ! > > > "if (source >= 1e-3 && source < 1e7) {" is right. > > > > > > My mistake when reading 10 exp -3 in Double.toString javadoc and writing > > > 10e-3 instead of 1e-3. > > > > > > Patch will follow. > > > > i'm not satisfied with this patch, since it depends upon the implementation > > specific nature of Double.toString(); i would like to see better checking in > > the code I listed in comment 34 to remove assumptions about the use of > > scientific notation (and assuming that 'E' is present) > > for example, what happens if source is NaN? just looking at the code makes > me think the same else clause cited in comment 34 will apply, which means > yet again, that parseInt will be invoked on the string "NaN" in this regard, please add NaN, +Infinity, and -Infinity to your test cases; also add the following test cases: * 1E-3 + Double.MIN_VALUE * 1E-3 * 1E-3 - Double.MIN_VALUE * 1E7 + Double.MIN_VALUE * 1E7 * 1E7 - Double.MIN_VALUE -- You are receiving this mail because: You are the assignee for the bug.
