Ping! Any opinion on this one? /Roman
Am Donnerstag, den 27.09.2007, 21:01 +0200 schrieb Roman Kennke: > Hello, > > while implementing GapContent back then in GNU Classpath, I noticed that > the JDK implementation does not exactly conform to the specs in > AbstractDocument.Content. There, all the methods are specified to throw > a BadLocationException for invalid locations. However, the GapContent > doesn't throw an exception there. For example, it is possible to create > a Position for a location that is outside the GapContent object. > > I'm not sure if I'm mis-interpreting the specification here. It could be > argued that it means, the implementation MAY throw an exception in such > a case, but doesn't HAVE TO. OTOH, not doing so in the reference > implementation might lead application implementors to rely on less > strict behaviour which can easily break on different implementations. > > What do you think? > > BTW, we observed a number of other places where the spec and the > behaviour diverged in Swing. David, do you recall some of these? I think > you even filed bug reports about some, right? > > Cheers, Roman -- http://kennke.org/blog/
