Showing posts with label bugs. Show all posts
Showing posts with label bugs. Show all posts

Monday, 7 June 2010

MacOS hotfix for crash in shader loading

If you run into shader problems on your Mac, plese use this new fixpack (unzip in game's dir overwriting all the stuff): link. Probably other nvidia owners might be affected by this, if you get an error in your log related to 'mod' (modulo), try this hotfix.

Saturday, 5 June 2010

Fix pack download is available

Fix pack for 'Almost Pristine' release is here for you to download and test!

Fixing a few things:

- MacOS nVidia crash (const int in shaders for float values are not acceptable with MacOS nv driver)
- Background terrain color flickering with nvidia cards
- Depth of field shader for lower precision cards is now working, without blurring whole screen
- Zooming unit texts are now not oversized after multiple zooming
- Adding missing binary cache for OBJ models
- Mouse look is now switched OFF when entering combat

Thanks for all your testing. Please give this fix pack another try soon, so I can roll out and announce the release later for major publicity (happypenguin, freshmeat etc.) as well...

Grab the package, unzip it in the game's folder, overwriting all (including your jCRPG-engine dir and all its content). This one is only for the latest release.

Linux intel vga quarter year release still NO-GO

No luck with the new driver release (im using the testing repo of Arch to grab it for my notebook): gma965 still crashes: java: brw_wm_surface_state.c:591: brw_update_renderbuffer_surface: Assertion `brw->has_surface_tile_offset || (tile_x == 0 && tile_y == 0)' failed.

Rant, rant, rant, why did they spoil this with crashy assertions. :/
(Commenting out the assertion, things run flawlessly. With this...it crashes)

Thursday, 3 June 2010

Fixing technical issues

Currently working on ironing out nVidia related shader issues. Like... Depth of Field precision issues came into picture. Solved. Next thing is : light is flickering on the terrain with nVidia cards. Another fatal error with character creation is fixed. After fixing these i'll roll out another release this time with bigger publicity. Thanks for all the folks who helped to discover the problems with the new release! (especially: qubodup, tidbit, people at Hungarian Unix Portal.)

And yeah, happy 3rd birthday jClassicRPG. :D

Wednesday, 5 May 2010

Things are pretty close to release

Time spent on squasing the few remaining bugs and adding parameterization to the starting view distance of the new clipmap terrain (which made possible the usual arbitrary setting of the old tiles view distance) means that all of things related to the Ardor3D based release are in place now. I still want to do some profiling and updating the content hieararchy before commiting the stuff and rolling out the release. Stay tuned... :)

Sunday, 25 April 2010

Forging a perfect weapon takes a lot of time

It takes a heap of time to get the things really polished and perfectly working. I've stumbled upon a dozen of smaller things that block the way. 80% of them eliminated, so the code is getting into shape. A few still left on the list but I hope to get those done in the next week or so. Well, just to name one heavier problem that took hours to conquer was Terrain reloading... starting at a different point than new game's origo rendered it ruined..now it's origo is the saved coordinate of the player and that way it's working flawlessly.

tidbit is restlessly helping out the content side. This time a new guardpost has been raised from the ground. Thanks tidbit!

I have some bad news, some newly inserted 'assert' codes in the current stable and development version of Intel GPU driver for linux (gma965 that I know of) is making the game stop just upon starting... I hate to say, but without removing those assertions directly in the driver game won't start for you. If removed, it works okay... ATi and nVidia drivers are both working nice under Linux/Windows - except ATi's framebuffer doesnt render texture shadows (Rikard @ ardor3d pointed me to some sites were people had the same issue, thanks), so you can assume that the shadows of the foliage won't be rendered (under both Win/Linux). nVidia is perfectly fine again.

Thursday, 8 April 2010

New buildings by tidbit + further testing/bugsquashing with Ardor3D version

Pleased to inform you that in the recent weeks tidbit created some nice new 3D content, namely new buildings for a temple and a shop. Thanks! Meanwhile I've been pushing the tests/validation of the ported code, squashing some upcoming problems and bugs. It looks like it is steadily becoming more and more finalized. Latest improvement was the restructuring of the directories containing textures and creating a separated media-raw hierarchy which servers as convenient base for batch converting images to the different detail levels and ultimately to DDS file format. I'm happy to tell that now 98% of the images are converted to DDS with mipmaps with AMD ATi's Compressonator and that all of these finally got their proper detail level versions (high/middle/low/low_png) obsoleting most of the textures found in the 'common' category. The DDS part was hiding some traps for me, for now I had to change to use the Compressonator instead of the GIMP DDS plugin as that one was generating DDS with missing mipmap levels somehow - interestingly it was okay under Linux + ATi, that way I didnt notice until I tested it under Windows where it presented small gaps in the texture. (Another thing that I should rant about is Intel's driver -- works under Linux but the UI part crashes under Windows + Intel GPU... this is still unresolved, funny thing is that if I downgraded to a 3 years old Windows Intel GPU driver the crash went away.)

Monday, 23 November 2009

Trac on sf.net activated - for task/bug tracking

I've decided to check out the task/bug/project tracker utility used by many projects. I've looked at it many times, lately related to PARPG project. It's looking neat and useful with all the project management/svn/task tracking integration it has. Form now on jClassicRPG project will use this facility to manage such things.

URL: http://sourceforge.net/apps/trac/javacrpg/roadmap.

Unfortunately it needs sf.net account to be editable (at least that filters spamming), so if you want to report a bug you'll need to register on the sourceforge.net site. (You can still browse it without an account though.) After registering that you'll receive email notifications about updates for your reported bug, so it might be worthy to register - as the jCRPG forum has no email facility. Those who prefer to use the forum - you can still use that as a less 'technical' form to report problems. (Old bug tracker is deactivated on sf.net project site.)

Saturday, 21 November 2009

Important quick fix again and great news - donation received!

Hello, folks!

A problem experienced by some of the testers and also reported on happypenguin.org, is now hunted down and exterminated! It was related to the logging of X11 events and a badly created filter pattern in Jcrpg.java. Now it's rooted out and committed. You can find a zip with the fixed ./bin directory of the release here: link. I will provide a patch release on sf.net later (hopefully tomorrow).

Another problem experienced on some weaker video cards (like 9600 mobility or intel video chip) can be resolved by first running the game with the jcrpg-lowest-debug.sh. (Then go to option menu and probably switch off the Normal maps totally too, save it, and later you can use the jcrpg.sh to play the game as the low settings got saved that way.) So if it looks like you are having troubles with starting, try to fall back to that jcrpg-lowest-debug.sh/bat file.

Very good news is that the project got a new donator! We have received donation of $5. Big thanks go out! I'm contactacting the donator to get info if I can share the name or nickname (desired one) of the donator before I can post it here and add to the DONATORS file of the project.

EDIT: The new donator is Chris Noffsinger from the USA. Thanks Chris!!
Donator granted me permission to share his name.

Wednesday, 4 November 2009

Quick fix needed - fatal error in loading reported

Here you can download the quick fix in case you have downloaded the game before this post (20091102 version) - those who have the new version (20091104) can skip this post.
The bug prevents you from loading the save games. Also the save game saved without this fix will not be working even with the fix. So it's recommended to update ASAP, before saving games! :) Soon I'll upload a full package too. Installation: unzip in the folder of the game, overwriting previous files.

EDIT: Full package uploaded now. Fresh downloaders use the new full package - you shouldnt do anything special, just go to download page and now everything should be okay. :)

Sunday, 18 October 2009

Contributions, shaders, new normal maps

Polishing, polishing, polishing - that's what's being done right now. And nice contributions by our new community member getter77, an aspiring and talented game developer. Right now he has created a pair of male voicepacks (1 for psychotic and 1 for bully personality). Thanks getter77 (aka Brian)!

Meanwhile I've fired up gimp-normalmap + gimp-dds plugin again on my laptop, and started to spit out a dozen of normal/height/spec maps for the already present building/road/tree textures. The result is awesome. The heavy use of such mappings culminated a series of interesting bugs/detection of missing features of the shaders. Most of these problems are now resolved, also a new level of normal map shader detail is added to the Options menu (detailed shader can handle 2 lights blended). Also fog is now supported by those shaders - and the shadow render pass is also fixed to work correctly with those. Check the screenshots for shaders/mappings in action! :) On the second one you can catch tidbit's new stone base square for the anatipion towns. Release is very near now...

Saturday, 10 October 2009

Screenies, polishing, testing multiple platforms

Options menu is again in focus. A few related bugs resolved now. Meanwhile tidbit has added cool new animations for the new humanoids (yeti and anatipion [which is walking, scorpion like people of desert]), all those got into the SVN too. Anatipions even have their own building models (done also by tidbit). Thank you!

More testing is under way. Even Windows 7 64 bit test version got its round after Vista64. (You'll have to be careful with Windows 7, like installing latest nvidia driver for nvidia cards, otherwise it doesn't work. Also it's recommended to use 32 bit Java under Windows, seems like 64 bit doesnt want to work with the supplied 64 bit lwjgl dlls.) Well, with my new mobo + CPU and my good 'old' 8800GTS, I've tried to boost the limits of view in game, game works nicely with a view distance of 100 now. Check screenshots.

Friday, 2 October 2009

Bughunting, profiling, polishing

In brief: smooth camera turning added for combat plus zooming activated enemy title text for better understanding what's going on the battlefield; left over memory reference hunting to allow restarting/loading games multiple times without lag or OutOfMemory; new male character voice pack by qubodup added by qubodup personally, thank you; a dozen other smaller bugfixes; correcly working mouselook while moving; added starting-at-closest-shrine at starting new game; saving of Options menu settings implemented (now on you can use jcrpg.sh/bat instead of the special scripts for different detail levels and use Options menu to select/tweak the needed detal level and save). New content done by tidbit awaits integration (brand new humanoid base model and its home buildings). After adding those I'll start to give the really really final polishing to the release package (reviewing scripts and documents) and then......

Monday, 24 August 2009

Fixing Windows mouse issues, Vista 64 bits note

I have just fixed two issues related to Windows + the new mouse input. First, cursors under Windows rather be 32x32 or displayed corrupted - googled it up after a series of experimentation in Gimp with no success. Now all cursors resized to 32x32, to let it work as fine as under Linux. Second, main menu mouse input didnt work under Windows - solving it was to prepare the code for Windows specific file path separator. All committed into SVN repo.

MysticSkript visited us on IRC chat and told about a problem related to running latest release under Vista64bit. I figured out just today that the solution right now is to use x86 (32-bit) Java installation if you want to run the previous release. (SVN version should work with 64 bit Java and Vista64 tho' I haven't tested it yet.)

Wednesday, 29 July 2009

jME and lwjgl updates in SVN

I'm currently catching up with what has been going on in the jMonkeyEngine development. It's still a rather active community there, I'm glad to see - patches are incorporated in the jme svn, dependent libraries updated. So I took the time and updated jcrpg's dependencies.

Next thing I must have a look at before the planning is to debug a recent in game text related bugreport - fonts are getting totally unreadable with some GPUs under Windows. Have any of you experienced the same thing (messed up fonts in-game) with latest release - if so what GPU/OS do you have?

PS: Just attached the bug's screenshot. Anyone else with it?

Friday, 20 February 2009

Intel video driver 3D issues under Linux

Currently I'm busy with resolving the recently detected Intel openGL + Linux related issues of jClassicRPG. I'm progressing well, I could identify a way to prevent instant and notorious crashes with xf86-video-intel driver under jME. Also I've detected that DDS support is a missing feature in the driver, so I'll have to add alternative textures for the ones in DDS format and add some tricky resource locator configuration to override them when DISABLE_DDS is specified in the configuration file. I'm happy that now I'm able to test the game with Intel driver as well, though performance of the GM965 is quite mediocre compared to those of the Nvidia GF6200Go (of course it's not a surprise, these video chips are known to be less performant 3D wise).
PS: patching the jME TextureKey and TextureManager to fall back to png/jpg if DDS is disabled, and adding fallback textures as well, I've finally brought up the project to shape with intel graphics...though it still is a lot more unstable than on any card of the two other major brand of video cards - probably due to the overload of 3D and driver problems?!

Monday, 9 February 2009

Polishing, ironing out bugs, recent donations

I was concentrating on testing the new features and fixing the bugs that got into the code. At the moment I'm in the process of fixing the world map thingy which seems to be incorrect regarding the geography representation. I plan to iron out some remaining bugs while freezing the addition of new features into the game till the release of the next prealpha test package. Recently mvrasseli has come to the forums of jcrpg and has started to work on a new piece of music for the mountain caves. You can check the progress here.

Good news is that the latest donation campaign is starting to spin up with two recent donators (Juan and Josh, thanks!) sending some coins (till now a total of $30), catalyzing the 'Free the CC-BY-SA-NC soundtracks to CC-BY-SA' thing that is going on. After these now $20 is requested to free the first track from the NonCommercial term. I've added a little section, so that you can see the virtual budget of the project on the left side now under the 'Project Budget' panel.

PS: World map fixed. Now positioning is correct and I managed to fix the population pixels too. White shows normal villages, yellow shows dangerous maze populations - check it on the screenshot. :D
PS2: First $50 reached, tomorrow I will relicense the first track (The March On Ice) under CC-BY-SA (it's too late here now 11PM :)).

Monday, 12 January 2009

Back To Coding - AudioSystem

All compositions are ready, two for each climate zone. So now I'm back to coding, at the moment trying to add background music in game playback to the AudioSystem class and its dependencies. It's going nice, but a serious problem with it is still blocking the road. Playback and change of tracks by zone is working well, but after a while of walking around background music doesn't play at all. Seems like a problem with the underlying library openal + jorbis. I will investigate it further and after that I'll try to finish the treasure chest codes, getting close to a long awaited release. :)
UPDATE: It's getting better, seems like I could identify the issue: OpenAL Sources (practically sound 'channels') are not freed/reused correctly if I try streamed and preloaded content playback on reused Sources after playing the other way on it before. I've modified JME's openal audio system to separate them. I'm testing it further, so far seems to be better now!

Friday, 17 October 2008

Small steps - shaders, bugfixing, integrating contributions

A short list what's being done: I've been working to provide normal maps for already used textures with gimp's normal map plugin. It's not as good as I would like it, but it's usable with some experimenting. A few bugs (like slowly flying arrow/knife bug) on the list has been hunted down meanwhile. Parallax Shader has got some improvement like pointlight/attenuation support. Also tidbit has provided his nice new model green lizard (labeled 'Lizzie' on the forum :)) with a texture and after some community feedback and eyeballing it has become a part of the SVN repository! :) Thanks go to tidbit, and ppl on the forum!

If you happen to be good at 2D texture creation (for 3d models), and/or have talent or already created normal/specular map textures for some generally usable textures + you would like to help jClassicRPG then it would be highly appreciated if you help the project with some new wall etc. textures with normal/specular maps.

Also animators would be much welcome as we have multiple models that would need rig and/or animation! Come visit to forum for more info.

Tuesday, 9 September 2008

Background thread for area loading, optimizations, bugfixes

Finally after two previous attempt to realize a working and fast parallel load of map I have succeeded. Now a separate thread is starting to load the area when you reach the middle of the previous render distance, so while you keep walking around it will load the things and with high chance you won't reach the limit of loaded area before the load of the newly needed area finishes in the background thread. It generally doesn't happen yet the rendering is prepared for that, if you reach the limit the background thread is stopped and a new foreground loading starts halting you walk. But this case happens really rarely if you selected the right level of detail for your PC's performance. Thus you will be able to walk freely without pauses with a dynamic background loading - in the next release that will come. Or if you check out the latest SVN you can test it right now.

Bugs...Long lurking culling problem of tree foliage now I could eliminate - GeometryBatchMesh spatial instance needed a buildMatrices to have the correct world rotation before being locked. Also I could fix some other annoying bugs - foliage of bushes rendered at high above the trunk sometimes and texturestate vegetations were wrongly repositioned while you walked around (both problems were related to reuse and wrong modification of position vectors).

A new RenderedArea implementation was born to fasten the loading of area - and it was a successful rewrite! Now it's 3 or 4 times or even 8 times faster than before in some cases. It includes a smarter calculation of areas to newly load and to remove. It calculates the newly needed and removable areas based on the disposition to the last rendered location vector. The other big thing was to define and implement a new getFilledZonesOfY for the Place superclass. Every Place extension (Geographies, Economics, Waters, etc.) now tells what Y height intervals are parts of them at a given X/Z location of the World. The information is collected and combed together and is used to determine what Y intervals to include in RenderedArea's rendering. This way about 4 or 5 times less World.getCube call is needed - only the really filled cubes are being requested by the loading process, empty positions are skipped. Naturally this leads to a much better loading time because of the nature how a world's surface is only consisting of a few cubes height even considering economic upon them.

Cool! :) Listing the things done since starting the optimization session: much more stable memory use (less overall memory needed), continuous appending of 3D scenario, background process for loading areas with no pauses, important bugfixes, better 3D performance/loading times.

Twitter