Tuesday, January 28, 2003

Java/ATI bug exposes configuration woes I just lost a couple of hours to the Java/ATI video card incompatibility. Ouch! When was the last time you saw a Java-related bug crash Windows so hard that you got a blue screen and had to remove the laptop battery before restarting! The various workarounds show how far we have to go to make Java configuration workable, especially in code-download scenarios. For example, cforster says: Aahhh, what a nightmare! We'd *really* like to move our apps/applets to 1.4.1 (using 1.3.1 now) but how can we if something like this is around? Re: "Is there any way to achieve the equivalent of -Dsun.java2d.noddraw=true for every application (including applets and webstarts)?" - 1.4.1 Webstart apps can put the noddraw prop in their JNLP file (see forums, etc. for details). But, you'll be forcing *everyone* running the app to run with slow 2D... Savvy plug-in *users* can put the sun.java2d.noddraw runtime param in their JPI control panel, BUT this is totally UNWORKABLE for JApplet deployers (esp. those with no control of desktop hardware, drivers, etc... Eg. most of us:) Seems, at minimum, JPI applet deployers need control of java2d props at HTML tag/param level in order to consider rolling out under 1.4.1. Would *also* be nice if JRE could detect where it cannot reliably render & automatically revert to noddraw, etc. Regardless, deployers need control of props affecting the succesful use of their apps & applets. Seems like the configuration issues I wrote about aren't getting any better over time. 12:56:38 PM    comment []