Again - please don't confuse by posting on the stable side. ===================================================== Update 19 (Build 337-unstable, 338-release) Changelog (October 17, 2013): ===================================================== FEATURES: ———————— Now allowing user vehicle data (like FFB multiplier) to be stored separately from mods. To enable, set the PLR file option "User Vehicle Data" to 1. The current default is 0, which is the original behavior. If it works well, we will change the default. Allowing virtual mirrors to be used with onboard cameras now. Also made mirror key apply to swingman and onboards. Added UI options to create a "virtual" vehicle...based off an existing vehicle. Goto Tuning Page with an existing vehicle...select a loose alternative skin. (if no loose alternative skins are availible, hit the "create dir" button, then the "reload" button and it'll generate one for you (use the skin spinner and select "alt"). Once you select an alternative loose skin, the "create" new driver button should be enabled. hit that...fill in the new driver name and team, and next time you enter the vehicle selection screen you should be able to select your new vehicle. (new "virtual" vehicle will also be available as an opponent) (Doesn't currently work for multiplayer). Added a more generic way for physics modders to attach pushrods (or spring/damper units); in particular they can be attached to A-arms now. Added tire compounds to results file. Allow brake bias to be mapped from a controller axis with the controller.ini setting "Brake Bias Axis". GRAPHICS: ———————— Changed baseline turbidity to 2.0 Now able to exclude specified objects (such as cockpits) from HDR scene avg luminance computation. New particle cloud system is being phased in. Improvements to HDR downsampling. Reflection maps set to Off now uses a near-static map. Some improvements to reflection map rendering pipeline. BUG FIXES / OPTIMIZATIONS: ———————— Fixed case where incorrect password would be saved after a failed download. Fixed incident reporting and crash sounds, not coincidentally! Fixed mistake when we added support for 6 controllers. Garage UI captions that are overridden by one vehicle now get reset to the default if a different vehicle is chosen. Fixed time length of replay reported in Replay Information box. Improved handling of the situation where there are more vehicles than garage spots (added warning message, and prevented a possible CTD). Added a controller.ini variable "Detect Rearranged Controllers" which was and is on by default, but can now be disabled if it seems to cause problems. Computing main vehicle body extents using collision body only now, since collision is its primary purpose. For some cars, this fixes a problem where the AI thought the car was really wide due to the ambient shadow object. Fixed a problem where the grid at the monitor would look messed up if adding/removing vehicles while paused. A hopeful fix for the problem where Windows 8 users are reporting occasional lost controllers. MULTIPLAYER: ———————— Allow /restartwarmup to go back to the warmup session briefly even if it is disabled. The purpose is to allow pending drivers to join. Optimized the initialization of temp cars in multiplayer, which reduces the slight hitch when other clients join while you are driving. Fixed bug where server messages were being labeled as "Your Name:" instead of "[server]:" MODDING / PUBLIC DEV ———————— Fixed crash from entering game with no AIW file. Increased rate for plugin scoring output (and commentary requests) from 2Hz to 5Hz. AI: ———————— Introduced version 1.0 of new AI learning code. When Autocalibrate AI Mode="2" in the playerfile, AIs will run laps in practice trying to optimize their operating parameters on a per corner basis. They create a database (*.wis file...for "wisdom", in the UserData\player\Settings\componentName\VehicleFile directory) with the results & use these results to adjust their parameters on the fly in a race. Because this is the first release, it's not completely full-featured...for best results, during practice, disable fuel usage, tire wear, and mechanical failures, and try to keep out of their way (they will not currently ignore a result because the the AI was held up by another car). Results are stored between games, so you may run a race with fuel usage, tire wear, and mechanical failures once the databases are created.
Probably not a bad idea to recommend testers use a separate 'clean' build with only rF2 official content. That will keep other (third-party) track issues from potentially 'contaminating' your testing results.
I'd agree except the Autocalibrate AI should probably be used with third party content specifically and ISI might want to know about issues that caused them to do things like quickly put out release 300.
Thought the idea was that public testing was all about what the internal testers didn't have time to test, ie all those third party tracks and cars and what effect they may have on any given build (like the recent long name bug) Exactly the kind of "contamination this is set up to catch, I thought...
Harsh as it might sound...and please don't take this the wrong way; ISI's priority is to guarantee the official software works as bug-free as possible. It doesn't guarantee third-party add-ons. That said, I'm sure the'd do everything to assist with their integration, since they're benefits to having them. I'd venture a guess their internal testing was concentrated on rF2 only...without add-ons. Ultimately, it's up to the add-on track providers to fix any issues to ensure compatibility. Unless a specific instruction set was put out by ISI (which followed the official ISI track build-up), they'd just be too many variables...anyone of which could negatively impact testing results.
See your point. But, the long name bug was solved by altering something in the ISI code, the internal testers missed it because they are testing ISI content, not third party. Now had this been in place at the time, it would have been caught by those testing the "unstable" release, and fixed before being released as an official build. See my point? Internal testers testing ISI content, everyone else testing everything else. Sure there is nothing stopping you just testing ISI content, but if everyone does that? Things like the long name bug still don't get caught and nothing has really been gained. Also any other bugs caught can be highlighted on a build release, ie "Croft will not work with this build, please uninstall until it is updated" Saving a lot of hassle, and alerting the modder to a problem that needs fixing early.
I can see your point of view as well, which is one of the reasons I'd like to see a post/poll indicating simply (YES/NO) to add-ons which have successfully loaded and worked in rF2. In that way, It'd be very easy for users and track builders alike to rapidly identify if a particular track has issues. The fact that ISI adjusted things to address the issue with non-ISI tracks, confirms what I said earlier about aiding integration.
If you're testing the unstable Builds then you should be testing all aspects of the game, regardless as to whether it's ISI or non-ISI content, if you dont then the bug mentioned above would'nt be found...that's a perfect example right there!
Tim, we try this out to help out a bit. Isn't it great we just want to be thorough? I don't see your post as the nicest one ever. Some other thing; if you want it tested as good as we can do, maybe you could explain some things like the AI learning mode (2) so we know how it works and test it for you?
from the changelog: "Because this is the first release, it's not completely full-featured" I read that as "it will be improved upon in the future" so nothing to really test right now, as long as it works in basic form and does'nt cause the game to crash (for example) then that'd be all ISI would want to know. I sort of look at these unstable Builds as Us helping ISI find Build Breakers, as - unfortunatley - little bugs will always get through but as long as we can help find the main ones (Real Road not working, Reflections acting weird, Resume From Replay not working, Major crashes, Major slowdowns etc etc etc) then that's gotta be good enough, they will of course log every bug found and deal with it accordingly but the fix may miss the next Build coming out.
Jeremy saying on twitter that bugs were found and another unstable build might be out Friday. I wonder what the showstopper was, seeing as no one here has reported any CTD or critical bugs.
