burek for breakfast

My work e-mail address has changed - the @sun.com part is now @oracle.com. Careful readers of the OpenJDK mailing lists have already noticed that, of course, but since not everyone is on the various mailing lists, I thought I'd mention it over here, too, like I did over on my twitter stream.

After talking on OpenJDK & JDK 7 at the European Java Roadshow, FOSDEM, JAX, GeeCON, JUG Münster & LinuxTag so far this year, my next stop will be the Java One and Oracle Develop in September, where I'll be doing an OpenJDK BOF with Kelly O'Hair.

Browsing the 2010 content catalog, there are five sessions mentioning OpenJDK in their title or abstract, which is up from two at JavaOne back in 2009. The other directly OpenJDK-related sessions are "Dual-Pivot Quicksort and Timsort, or Sorting on Steroids", "OpenJDK at Google: A Year in the Life", "Performance and Debugging Advancements in OpenJDK", "Project Lambda: To Multicore and Beyond", and feature speakers from Google & AMD, beyond the usual suspects (i.e. Oracle & Sun Microsystems).

To over here: http://blogs.sun.com/openjdk/ . Future roundups will be posted over there.

The last roundup was two weeks ago, so it's time for another one.

The JDK 7 project released build 79. The list of changes for this build has bug fixes for compressed oops, G1, more work on JSR 292 and build infrastructure for modules.

In the Jigsaw project, Mark Reinhold posted a draft for a simple file format for modules.

Over in the Da Vinci VM project, Lukas Stadler, currently busy hacking on coroutines, wrote a post introducing the concept, it's potential uses, and the status of his patch in the mlvm repository.

The schedule for the Free Java devroom at the FOSDEM conference has been published. It contains the familiar mix of OpenJDK-related topics, packaging themes and open source projects running on top of the JVM.

Sean Mullan announced that a new version of the Secure Coding Guidelines for the Java Programming Language is available.

Finally, over in the BSD porting project, David Green published an update on the state of the remaining bugs that prevented Eclipse from launching with OpenJDK from MacPorts - it's looking good for the upcoming Eclipse 3.6 release.

The schedule for the Free Java devroom at FOSDEM has been published.

It contains the by now familiar mix of OpenJDK-related topics, packaging themes and open source projects running on top of the JVM.

The schedule is a bit lighter then in the previous years, where we tried to cram as many talks as possible into the devroom schedule at expense of breaks between sessions. This year, since there is one open source Java related talk in the FOSDEM main track every day (about Systemtap on Saturday, and about Apache Hadoop on Sunday), there is a strategically placed longer break every day in the devroom to allow visitors to head to the main track and see that talk, or stick around in the devroom and continue working with each other on topics that interest them.

As usual, there were some last minute requests for talk slots. A spot was found for most of them, so the schedule on the FOSDEM site will likely see a small update for Sunday in the coming days - but there is also a special Sunday morning slot for last minute lightning talks. I'm hoping to see a very fast, spontaneous version of last year's VM rumble, given that IKVM, JamVM, JikesRVM, Cacao VM & Jato VM have seen new releases since last FOSDEM, assuming their maintainers attend FOSDEM, of course.

The InfoQ team has posted new videos from the JVM Languages Summit:

• David Pollak on Scala Basics - Byte-code Fancypants


• Charles Nutter on JRuby, Duby, and Surinx: Building a Better Ruby


• Neal Gafter on Statically Dynamic Typing


• Cliff Click on A Crash Course in Modern Hardware

Updste: Thanks to InfoQ's Chief Editor Ryan Slobojan for tweeting me that Cliff's talk is now online, too.

Hiroshi Yamauchi provided instructions to speed up the OpenJDK build from the perspective of a developer trying to debug code freshly written code fast by using environment variables to turn off bits of the build that a developer may or may not be interested in.

The Haiku port has reached the stage where the HotSpot VM can be built natively on Haiku OS. Andrew Bachmann published the build instructions.

The BSD port started using Zero to get OpenJDK working on some new platforms, too. Landon Fuller posted his build instructions for OpenJDK on Leopard PPC. Greg Lewis posted a bootstrap build for OpenJDK on FreeBSD/sparc64 8.0.

It's not all new ports and platforms, of course. The Compiler Group sponsored a new Project, adding Project Lambda to OpenJDK's line up of projects. Mark Reinhold published a straw-man proposal to kick the discussion off, which is now taking place on the lambda-dev mailing list.

Finally, Alan Bateman wrote a great in-depth post about the current state of the work to break up the JDK into pieces.

It's that time of the year again - the FOSDEM developer rooms have been assigned.

The Java libre community has been using FOSDEM for five years in a row to meet, chat, collaborate, and discuss its work from way back in 2004, when GNU Classpath brought different open source VMs together, to last year's packed edition with 24 fast paced sessions ranging from OpenJDK & IcedTea to JNode and Grails in two days. So I'm glad to announce that thanks to the FOSDEM devroom team, we'll be able to have another Free Java devroom at next year's FOSDEM in Brussels, on February 6th & 7th.

Like every year, there is some basic information on the Free Java devroom in the wiki. If you'd like to talk about your Java libre related free software project in this devroom, please add your proposal until Sunday, January 3rd 2010 here.

Let's start with videos from Devoxx - the JDK 7 related videos are online now:

• Joe Darcy on Project Coin


• Mark Reinhold on JDK 7 & Java SE 7


• Michael Ernst on detecting and preventing bugs with pluggable type checking


• Alex Buckley & Brian Goetz on the road towards a universal VM


• Mark Reinhold on the Modular Java Platform & Project Jigsaw

In the OpenJDK BSD Port Wiki, S.P.Zeidler reported that there is an OpenJDK 7 package in DragonFly BSD. If there is a *BSD OS with an OpenJDK package not listed on the wiki, please add it there.

The BSD porters have been busy in the past week getting the Zero backend integrated into their port. With the implementation of the necessary atomic access primitives for *BSD operating systems, Zero now builds on FreeBSD on i386, amd64 and sparc64 architectures, on Mac OS X and on NetBSD. It'll take a bit for things to solidify, but it looks like the BSD porters team is making good progress here.

Following in the footprints of the BSD porters team, Simon Lewis published how to boostrap the 64bit version of OpenJDK 7 on Mac OS X, and along the same lines, how to do it for the 32 bit version as well. Getting OpenJDK 7 built is a precondition for him to be able to experiment with invokedynamic and method handles, and with a working OpenJDK 7 build on his machine, Simon starts weaving a set of blogs exploring how deep the invokedynamic rabbit hole goes.

Continuing with the invokedynamic theme, John Rose has published his thoughts on how invokedynamic could interact with a tail calls feature in the JVM.

Switching tracks from VM features to language feature developments last week: Following his Devoxx talk, Mark Reinhold published a blog post on closures for Java in JDK 7. And Joe Darcy pointed developers interested in trying out small language changes from Project Coin included in JDK 7 M5 to a developer build of NetBeans.

For new and old hat OpenJDK developers wanting to build OpenJDK themselves, Kelly O'Hair provides a hatful of tricks to speed up builds. Finally, Andrew Cowie shares the joy of having debug symbols for OpenJDK in self built binaries or distributions for developers debugging crashes in software reaching deep down into native code from Java, like java-gnome.

It's another fine piece by Stephenson. Where Pratchett in Nation uses multiverses as a way of making the historical, geographical and cultural backdrop of the story as similar to this universe as necessary to make the differences stand out, Stephenson dives with delight into the implications of multiverses on philosophy, mathematics, information theory, language and other sources of playful interaction with fun subjects, while meshing in cultural references from Greek philosophy up to spam filtering & Star Trek into the stream of information connecting his fictional multiverse with the reader's. It's pretty amusing to read a story about multiverses being told using a post-apocalyptical world divided between causal domains of worldly power and secluded scientific monasteries with limited information flow between them in this way, and Stepehenson takes care of adding a lot of lovely miniatures of right out of 'scary devil monastery', academic labs, and teenage love life to keep it all entertaining until the big finale, and some pages after that.

I enjoyed the made up words like 'speely' from bits and pieces in other languages, and the linguistic twists employed to come up with something like 'mathic' & 'incanters' to create a nod to Clarke and Stephenson's earlier works. All in all, very enjoyable, and even has, contrary to Stephenson's usual style, a 'proper' ending, as much as the multiverse setup permits.