fabio

Do you know about the recently open sourced, MIT licensed (GPL compatible) Torque engine? It looks to have nice water rendering, maybe there is something that could be useful for 0 A.D.: EDIT: it is Direct 3D only apparently.

Yes, thanks, I know about it but I am using older r300 driver, the patch is only for r600 driver. I opened an interesting similar bug affecting both r300 and r600 with 0 A.D. some time ago, however: https://bugs.freedesktop.org/show_bug.cgi?id=44536

Sorry the performance drop was with superfancywater. Fancywater doesn't have a noticeable FPS drop. They depends on the distance and IIRC they move, however I am not sure since I cannot test again until next week. (Also a note that with superfancywater I don't see error by the graphics driver about unassigned input).

I had some problem with postproc effects see here and later posts for the full story. FPS went from 20-21 to 13-14 with fancywater here (RV530 with git mesa driver).

You may want to try my PPA, see my signature (packages for 12.04 are currently unavailable due to mesa build problems, but they are working for 12.10).

12811 break this test: Running 270 tests.....................................................................................................................................................................................................ERROR: JavaScript error: simulation/components/GuiInterface.js line 77 TypeError: cmpPlayer.WantAlly is not a function ()@simulation/components/GuiInterface.js:77 ()@simulation/components/tests/test_GuiInterface.js:162 In TestComponentScripts::test_scripts: ../../../source/pch/test/../../../source/simulation2/components/tests/test_scripts.h:43: Error: Test failed: L"Running script simulation/components/tests/test_GuiInterface.js" ../../../source/pch/test/../../../source/simulation2/components/tests/test_scripts.h:43: Error: Assertion failed: scriptInterface.LoadScript(pathname, content) ........................................................................ Failed 1 of 270 tests Success rate: 99%

Yes for the "black areas", I suppose yes also for the zigzagged shoreline, but it's probably less noticeable on some maps (clearly reproducible at least on two maps). A last design issue is that waves always go to the beach, while foam goes back and forth: in some maps it doesn't seems very realistic, maybe in this case waves should be disabled?

The menu now works indeed. A couple of issues in this screenshot (Cycladic Arcipelago 1), only reproducible with superfancywater: zigzagged shoreline; little black areas (little points) when zooming far, see the top left seas.

What I mean is: is superfancywater local.cfg option matching "Enable advanced water" in the in-game options menu? If yes it's not working properly because the in-game menu option is always enabled no matter of the local.cfg option.

Have superfancywater config option an equivalent in the setting menu? I am asking because whatever setting I have in the config file for superfancywater, the water options in the game menu are always enabled. Also it would be nice to have a sound effect for waves (with volume depending on zoom level, no sound when zoomed too far).

I already suggested something like that here.

Thanks for the update! I think that at this stage it would be cool to have a proper in game option menu for managing all these amazing new effects, without editing the config file. Similar to what was done on the match setup GUI.

(Sorry for the delay) The error just appears with preferglsl=true (and no other settings), even if I add materialmgr.quality=0.0 Without preferglsl=true: Removing <PostEffect>hdr</PostEffect> from Zagros Mountains 01 removes the corruption. However I cannot notice any other effect (fog included). No errors are printed (also with hdr restored). With preferglsl=true: Removing <PostEffect>hdr</PostEffect> from Zagros Mountains 01 removes the corruption and I can notice some effect (at least fog). Errors are printed.

That problem also happens on current selection rings. See #1368 (#1517 also related).

The WPOS error: is triggered by this code: http://cgit.freedesk..._derived.c#n552 while with: http://cgit.freedesktop.org/mesa/mesa/tree/src/gallium/drivers/r300/r300_state_derived.c#n488

No, but is there a way to enable the selection box on single parts? This is probably the key that trigger the problem.

Hmm, I'll try tomorrow with my current driver (from mesa 9.1) but according to the 0 A.D. OpenGL capability report for GL_MAX_VARYING_FLOATS_ARB (although it's not updated since long) all ATI cards have at least 32 slots (or also 40 on some version of RV530, my codename). EDIT: also that error message only appears with glsl mode, while the corruption happens also on standard mode.

On Mesa R500 has complete OpenGL 2.1 (the hardware doesn't support NPOT textures, but they are implemented in software, the only problem is a slowdown when they are used). It's likely a driver bug however.

Sure, this should obey usual code standard. Eventually a requirement could be to work tracking progress on a ticket or a thread forum rather than doing a single code drop. Also the reward should be given after the code is approved and committed and at least after two weeks without reported bugs. Good news with the pathfinding! Other example bounty could be: usable Android port; cleanup and commit tickets like #885; improvement of 5% speed of a reference game replay, working on an area of choice (excluding pathfinding if it's being rewritten); use a thread to get at least a 10% speed up of a reference game replay; other improvements few ones have interested in doing but that can benefit the game.

OK I tried, but I cannot reproduce the problem with just waterfall and/or water. Note that the problem only happens when the mouse is over the gardens and the selection box appears (even if I don't click on it). When using the lonely water or waterfall the selection box doesn't appear over them, so this is probably why the problem cannot be reproduced with just these.

Postproc effects are not working properly o my Radeon X1600 with mesa drivers, see screnshots below. If I also enable preferglsl I get similar screen but also with these:

I measured the gathering speed: citizen: 1 food/s cavalry: 15 food/s It doesn't seems very realistic. Cavalry also has advantages of faster speed of movement and can transport 30 food vs 20 of a citizen (but has the disadvantage of gathering only meat and not vegetables).

Yes, I created a map with just the gardens and it happens whenever I hover the mouse on it. What do you mean with the "which prop is causing the problem"?

You maybe could set up an open source bounty to who is able to improve the pathfinding (or other areas).

Why don't just put in the game a check for the latest version (just check a file on wildfiregames.com containing latest release) and if > of installed version show a button with "New release available, check the 0 A.D. site!" that links to http://trac.wildfiregames.com/wiki/LatestRelease (or even better one of the subpage depending on the OS)?