jME 1.0rc1 is now officially released. Here is a collection of various details of the changes and differences between 0.11 and 1.0rc1: (mostly a digest of details from other posts.)
New additions and changes:
All logging calls are now handled by jul.
The common xml classes are now split out to a jmex.xml package.
Multitexturing and other GL13+ features have been changed to use their ARB equivalents in an effort to be more compatible with older drivers.
GLSL shaders have been revamped to use less memory and work better with uniforms and attributes.
Texture.anisoLevel - is now a percent (0.0f -> 1.0f) representing a range from no aniso filtering up to the max aniso level supported by the user's card. Additionally, you can now change aniso level after creating your texture if you so choose.
Texture parameters now use the various unit caps (fixed, fragment texs, fragment coords, etc.) to determine which texture calls are allowed to be called on a given card.
TextureKeys do not hold filter settings anymore. Filter settings can be changed at any time now, so they are no longer unique texture identifiers.
com.jme.util.resource - A new set of util classes for making resource locating easier has been added to jme core.
All previous ways of specifying a texture directory to a model loader are now replaced with the use of ResourceLocatorTool.
Now using lwjgl 1.1.2.
More complete javadoc. (yes, I know there's still a lot more to be done here.)
Lots of bugs fixes.
Removed API sections:
com.jmex.sound.* - This section has long been abandoned by its original author and was replaced in the last release with jmex.audio. Because jmex.audio only uses OpenAL (for now), the fmod libs have also been removed from jme/lib. You will probably need to update your build path and remove the fmod jar listing to make your IDE happy. StandardGame has also been ported to using jmex.audio. For information on porting to jmex.audio, see jmetest.audio tests and tests like HelloIntersection and AppletTestSounds which have been ported to use jmex.audio.
com.jmex.model.XMLParser.* - The model converters are now in jmex.model.converters, everything else like JmeBinaryReader/Writer and the XMLToBinary, etc. have been dropped. (Use BinaryExporter/Importer instead.)
TextureKey.LocationOverride - This interface is gone - replaced with the ResourceLocator and ResourceLocatorTool. See above for more details.
com.jme.util.LoggingSystem and com.jme.util.ErrorManager - (replaced with pure jul usage)
cvs cleaned: removed most files under jme/data, compressed and cleaned jmetest/data - CVS syncing and checkouts should now be faster.
yeah, it would be nice to have a minimal set of functional conversion tools when 1.0 is out. is there any kind of plan when the release comes or is it just "as soon as possible". i ask just because there may still be bugs to find and things to fix.
edit: for example should i bother to make Text work in non-ortho? or isn't that so important because of the other various workarounds/text-techniques?
Another long awaited feature we have added(at least for those doing production quality games that need to work on lots of comps) is that the render to texture with FBO's now fallbacks to PBuffer when needed
@sfera: I would be very interested in Text working in non-ortho, since I use a a non-ortho parallel projection render pass for my HUD to allow for hud elements with some thickness in the z axis (like cylindrical dials that spin around the x axis, etc.). At the moment I have to use ortho for the text written on non-ortho indicators, which is odd
then i might look into that. but i think it requires some small changes to the renderer and i'm not sure that they are so welcome just before the realease
Ok. I'd imagine our Text class will be replaced in 2.0 anyhow. Since we're in bug fix mode for 1.0 release, maybe a workaround would suffice. I personally use ortho UI textlabels with world->screen translation and scaling to do this trick. I could probably whip up an example to illustrate if interested.
Another long awaited feature we have added(at least for those doing production quality games that need to work on lots of comps) is that the render to texture with FBO's now fallbacks to PBuffer when needed ;)
(And don't forget to try TestNormalmap)
:-o Eye-popping, visually exquisite... Oblivion quality right there! }:-@