The Source for Java Technology Collaboration
User: Password:
Register | Login help    

Search

Online Books:
java.net on MarkMail:


JDK 6u14-b02

Posted by opinali on February 27, 2009 at 6:27 AM PST

The fixed bug list shows a few highlights:

  • Yet another LW/HW bugfix. Will this ever end? :-) It’s a hard problem, good thing that Sun seems motivated to fix it completely even if it takes several updates.
  • Two startup time fixes: Make sun.misc.FloatingDecimal.hexFloatPattern static field initialized lazily (the JRE would initialize the whole regex API with no reason!) and Avoid loading of Normalizer resources for simple uses (hits JAWS in particular). You would consider such bugs “low hanging fruit” that Sun should have fixed aeons ago… you’d be wrong.
  • Two RFEs to tune the plugin cache access.  This is significant for the “warm” startup time of applets and JAWS apps with significant resources. The bugtracks mention issues like lots of separate read requests to the same file and suboptimal indexing due to bad handling of buffering and metadata. It seems the JPI cache was not as mature as you’d expect, even at the plugin2 release.
  • Another 5 plugin2 bugs, including an important fix for better supporting multiple browsers (other than IE and Firefox). This is important as the browser ecosystem becomes more complex, with entries like Google Chrome and Safari for Windows (P.S.: I never liked Safari remarkably on Windows, but the v4 beta rocks. You’re reading that from a blogger who loves Firefox and usually bashes Apple for several reasons).

Meanwhile, in the JavaFX JIRA I found that only now they are paying attention to efficient handling of deployment resources. Check this bug: minify and compress deployJava.js. Quoting description: “deployJava.js can be "minified" and gzip HTTP compression needs to enabled to cut download from 32k to 3k ...”, same for several other .js dependencies. And this is not all; other bug asks to Lazy Load deployJava.js, description: “A large portion of the html page loading time is loading deployJava.js which is only ever used in the case that no JRE is installed. Make this file "lazy" or "dynamic" loaded so that it is only loaded when required.”. So, part of the loading time of JavaFX applets (and also sommon Java applets using the new 6u10 deployment toolkit) are just your browser fetching multiple big JavaScript files, often unnecessary files... gripes!

Well, that’s a lot of startup-time bugs discovered in a single day. The JRE bugs are fixed only in 6u14b02, the JavaScript issues in the JavaFX bugs are still unfixed (promised for 1.5 – but hopefully 1.1.1 since these should be pretty simple fixes?). Anyway, better later than never…

Zero GC fixes in this build. which is probably a good thing. My next plan is evaluation the G1 collector on JavaFX, stay tuned.

Related Topics >> J2SE      
Comments
Comments are listed in date ascending order (oldest first)