Hi Mario, On Fri, 2006-11-24 at 19:50 +0100, Mario Torre wrote: > The class is now complete for 1.5 and I would like to take it into the > next classpath version, even though it misses a couple of little things > in functionality.
Yes, it looks good. Thanks for your hard work on this. > It passes all the mauve tests, except few minor ones, the only real > stopper could be this case: > > apply (harness, df, "##0.####E0"); > harness.check (df.format (12345), "12.345E3"); > > fails with this result: > > FAIL: gnu.testlet.java.text.DecimalFormat.formatExp: pattern ##.###E0 > (number 1) > got 1.234E4 but expected 1.2345E4 > > As you can see, it misses the last digit. I believe this is a corner > case, so I would not mind right now for that (though it is clearly > wrong, as the pattern ask for 3 digits after the dot, they have to be > present if the result is nonzero.) But it this isn't a regression, before we weren't doing the right thing either. So I wouldn't worry too much about it. > For the formatToCharacterIterator stuff I got some hints from the ICU4J > class, this is stated on top of the file, and is the only method I did > not wrote by myself. And I have to say that this is also pretty > straightforward, but still it gave me clues about what to do. Is this > ok? The license is compatible and this seems small enough to actually be fair use. But it is good to mention it. Please do put the comment around the exact bit of code, not at the top of the file though. That makes clear what exact bits we are talking about. With that I think you should commit this since it is a clear improvement over what we had. thanks, Mark
