Re-enabled the api for player sense by bot authors (for @Exia) More useful debug information is now available from the Random class when there's an exception relating to the bounds of the arguments, more specifically the nextGaussian method. Fix for the loading of bots that choose to include the word class at the beginning of one of their middle packages for their main class, for example, if a bot did ".classes.MainClass.class" it wouldn't load. Enjoy @Bertrand Improved the handling of our resource managers usage of temporary files to try stop the "IOException: Access Denied" issue a few people are having. Added an About dialog with useful information for debugging so end users can access it more easily. Fixed a bug in our Linux support, you should now be able to bot smoothly Improved the debugging in our Linux support so that if you're missing a necessary package, it'll notify you what it is. Catch ServerExceptions that occur when a user closes the game and doesn't stop the bot first, previously they were being sent to bot authors and there logs were getting spammed with them. There's still some similar exceptions that can end up there, but we believe most of them are relevant and should likely be escalated to us. Applied an optimization to InterfaceComponent.getLayer() based on some new research. If this causes any bugs than please notify us immediately. Stopped caching root index arrays from InterfaceComponents because it may have broken isVisible() in some cases. Added support for the usage of root indices 100-103 on RS3 which were added at some point in the last year I believe. This didn't likely affect anyone but could prevent a few null returns from getBounds() in rare situations. InterfaceComponent#hashCode() was surprisingly slow (IntelliJ generated it suboptimally a while back), so I rewrote it to be fast like it should be. Fixed an NPE that could occur when projecting a coordinate to the minimap on RS3. Please note that with this release I've cleared my entire exception log. Developers should re-escalate issues to me if they believe it's the client's fault, but only if they're checked the bottom of the exception and verified it occurred with version V2.0.4.0, if it occurred on V2.0.3.1 (which some people will continue to run for a while), you are NOT to report it.
Sounds awesome man, so many fixes. Loving it! What does InterfaceComponent.getLayer() do? I've never noticed it before.
The standard tree view you see of contrainer -> component -> component is a poor representation of what's really going on. In reality, visibility and bounds are calculated based off of other components, sometimes in entirely different containers. The real tree is actually quite interesting to see.
@Cloud That's a big list of fixes! Also, I've noticed that all the bots that I run have a similar issue in that they have long pauses between each step. For example, the bot will click to walk a ways, then pause there for a bit and then click again. Or there will be long pauses between performing actions, they all seem to just be running slowly. I do not have any specific errors I have noticed from any logs, and I am sure whatever the cause(s) are they are well above my head, but I must assume this has to do with the client since this problem extends to all the bots that I have tried. I should probably add I am running Windows 10, on RS3.
Path navigation is still slower than it was before spectre's release but that'll likely change as spectre matures. In fact, in the next release there are some pretty significant performance improvements. Since I started to offer to profile bots for bot authors Are your users complaining that your bot is slow? Let me profile it for you. | Community | RuneMate I've also been noting the way things are frequently used and been able to apply optimizations to boost the performance of their bots on my end too.
