harmony-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Mark Wielaard <m...@klomp.org>
Subject Re: Assembled comments from the web
Date Sun, 08 May 2005 09:43:37 GMT
Hi,

On Sat, 2005-05-07 at 12:19 -0400, Henri Yandell wrote:
> 2) Pity for those who have to write the AWT libraries.

:) We can confirm that is a lot of hard work. Especially if you want to
be compatible with all the quirks (bugs?) in other AWT implementations.
An interesting experiment in this area is Odonata which only needs
access to a framebuffer device. The author, Stephane Meslin-Weber, gives
feedback to the GNU Classpath team to make sure our AWT is "pluggable".
The GNU Classpath default AWT implementation is based on native GTK+
peers. Kaffe comes with two other sets of AWT peers, one based on Qt.

> 3) Use JamVM. (It's GPL).

JamVM + GNU Classpath + Jikes compiler is probably one of the easier
methods to see what the current state of the free implementations is.
Another is of course using kaffe which is already used by the Apache
Gump project to monitor progress:
http://brutus.apache.org/gump/kaffe/
Or getting a distribution that already ships GCC 4.0 (with gcj). (In
respect to gcc 4.0 and gcj, take a look at this list of new packages
that will ship with Fedora core 4 based, most new ones are based on gcj:
http://people.redhat.com/sopwith/new-packages.txt

> 5) Have Sun open-source things to Harmony, or IBM.

I think everybody would welcome additional resources.

> 10) Does the TCK test serialization compatibility, or should that be
> something that gets created, a large test that compares two J2SEs for
> compatibility?

Mauve has some tests for that. See http://www.sourceware.org/mauve/
You might want to contact Brian Jones (cbj-at-gnu.org) to see how you
can help.

> 11) It would be interesting to see a document with areas for vendor
> improvement. For example, I assume we can't just put Jakarta ORO
> behind the java.util.regex package as ORO supports a richer dialect of
> regular expressions and I assume the dialect is in the spec.

This is one of the things to watch out for. Currently the GNU Classpath
util.regex implementation is based on  GNU regexp which has a slightly
different syntax (seem every package, library or tool doing regular
expressions comes with its own syntax, sigh). We have tried to translate
between the syntaxes, but there are bugs (see our and the gcc
bugdatabase).

> 15) Last, my favourite is the blog entry that doesn't comment but just
> provides a link to the homepages of everyone mentioned in the email: 
> http://www.weiqigao.com/blog/2005/05/07/re_who_are_these_people_was_re_open_source_jvm_proposal_on_the_table_at_apache_software_foundation.html

Nice! That puts things a bit more in perspective.

Cheers,

Mark

-- 
Escape the Java Trap with GNU Classpath!
http://www.gnu.org/philosophy/java-trap.html

Join the community at http://planet.classpath.org/

Mime
View raw message