Re: GNU Classpath and JDK1.0 completeness and compatibility

[Roman Kennke]

> FWIW, and you may already know this, nearly all the errors thrown up by
> japi for 1.0 (i.e. the binary compatibility) are differences between
> this and later versions (e.g. final/non-final, etc.).  It's worth
> checking stuff like this against the docs for Sun's 1.4.2 to see if they
> implement the same change.

Yes. I will do this in my audit and accept such changes in favour of the
newer spec.

>   As I recall, I only found the fields in font
> to truly differ from 1.0 and Classpath (and the visibility of these
> still seems a little dumb to me).

You mean java.awt.Font?

> I think this is a great idea on the whole, and should give a better idea
> of how well Classpath matches the Java APIs.  If you want others to help
> out (myself included), perhaps it might be best to tell us which order
> you intend to tackle the packages, and then others can elect to take
> some on too.

Ok, I have checked java.lang now (look at the pages for the results).

Next on my list are the AWT packages (I hope that the RH folks fill in
the missing pieces quickly).

after this I'll look at the rest in this order:
java.applet
java.io
java.util
java.net

I welcome everybody to help. This can be done by auditing the listed
classes and testcases as well as filling in the missing pieces in
already audited classes. Oh, and discussing stuff that is red and yellow
in the list may also help in some cases :)

At the moment all changes (proposals) in the list should be sent
directly to me. As soon as I have have CVS access to the GNU Classpath
webpages, I think these should go there.

Cheers, Roman

signature.asc
Description: Dies ist ein digital signierter Nachrichtenteil