halcyonXIII

Have the 0AD developers ever considered implementing terrain bonuses and penalties? Unit stats like speed, defence, attack range, health, stamina or perhaps even line of sight could be influenced by the terrain. I think roads would have more purpose if 0AD had such a system, allowing players to flatten the land and create an unhindered path to travel along through the rougher areas.

myconid, I think your changes to particle.vs has introduced a bug where the particles no longer rotate to always face the screen.

Ah, I see. It just threw me off a bit when I saw it was still working. Anyway, I haven't encountered any problems with ARB either. There was something in Atlas though (with preferglsl set back to true), with the "Grass/Grass1 spring fancy" brush texture. When I draw this on to the terrain with the left mouse button it blends with the surrounding textures as expected, if I then draw a different texture over part of this grass texture (also with the left mouse button) then they don't blend together properly. Likewise if I draw this grass with the right mouse button, then the grass doesn't blend nicely with the texture under it. The terrain highlighting around the cursor disappears over this texture as well.

Is that by setting preferglsl and gentangents to false? Incidentally, is smoothlos supposed to work when preferglsl is set to false?

Yes, both problems have ceased now ^^

That might explain the second time it happened to me, if my CPU opponent built a mill. (They were set to random, so they may have ended up being the Britons or the Gauls.) I selected "continue" in this error window to resume the game, and the mill's terrain decal hadn't loaded and 0AD was emitting a constant stream of errors for the duration of the mill's existence. ERROR: Terrain renderer failed to load shader effect (model_transparent)

I've come across a couple of issues so far. Firstly, in Atlas (but not in-game), the Roman Civ Centre has a large black square on the terrain around it. This square doesn't rotate when I rotate the structure, but it does scale so the structure fits inside it's bounderies. See attached image. Secondly, 0AD crashed. It happened once in Atlas whilst looking at the temple_mars_2/rome_civil_centre, and once when playing on a random map using the Romans. I copied the error message, but I haven't been able to get it to crash again when running under gdb. Assertion failed: "0 && (L"Can't find a usable technique")" Location: ShaderManager.cpp:474 (NewEffect) Call stack: (0x81e9ea) ./pyrogenesis() [0x81e9ea] (0x7c7a81) ./pyrogenesis() [0x7c7a81] (0x7c875e) ./pyrogenesis() [0x7c875e] (0x7c8a33) ./pyrogenesis() [0x7c8a33] (0x6137cc) ./pyrogenesis() [0x6137cc] (0x614f6d) ./pyrogenesis() [0x614f6d] (0x6e974a) ./pyrogenesis() [0x6e974a] (0x68a081) ./pyrogenesis() [0x68a081] (0x6569bb) ./pyrogenesis() [0x6569bb] (0x659da2) ./pyrogenesis() [0x659da2] (0x65a73f) ./pyrogenesis() [0x65a73f] (0x590771) ./pyrogenesis() [0x590771] (0x41df63) ./pyrogenesis() [0x41df63] (0x413a47) ./pyrogenesis() [0x413a47] (0x7f198322076d) /lib/x86_64-linux-gnu/libc.so.6(__libc_start_main+0xed) [0x7f198322076d] (0x41c575) ./pyrogenesis() [0x41c575] errno = 0 (Try again later) OS error = ?

Okay, thank you for the explaination ^^ In Atlas, the problem tree seems to be "gaia/flora_tree_pine_animated", which according to the "pine_animated.xml" file in art/actors/flora/trees/, uses the mesh "tree_pine_01.dae" in art/meshes/gaia/. What defines the root point, is it the object's origin or something else?

Thanks, that fixed it. I've noticed one small issue with the windy trees. You can see this on the Deep Forest random map, take a look at the body of the Pine trees (the green-blue coloured ones). They deform too strongly, almost as though they are pulsating. Do the Pine tree model/s require more verticies?

I see, no worries. Now I know what the -p option does. I probably should have already read up on patch if I'm going to make use of it. The compilation completed. When I load a map I get the following messages. ERROR: Failed to compile shader 'shaders/glsl/model_common.fs': 0(205) : error C7505: OpenGL does not allow swizzles on scalar expressions ERROR: Failed to compile shader 'shaders/glsl/model_common.fs': 0(205) : error C7505: OpenGL does not allow swizzles on scalar expressions

I'm getting this now. In the diff, all of the file paths are written as such. diff --git a/binaries/data/config/default.cfg b/binaries/data/config/default.cfg index 06af9c0..e9bbaf8 100644 --- a/binaries/data/config/default.cfg +++ b/binaries/data/config/default.cfg I'm not familiar with the syntax of diff files, but the "a/..." and "b/..." shouldn't be preceding the file paths, should they? I modified your diff file, removing them, and the patch now seems to have applied properly, bar a bit of fuzz. I've attached the modified diff in case anyone else needs it. mmpatch_2.diff.zip

The patch failed to successfully apply all changes. This was on a clean copy of 0AD, updated to revision 12212.

I was thinking about this matter a while ago, but instead being able to build defensive structures (at the least) within allied territory. I've been using wooden outposts but they can't really cut it against large armies, and they become completely ineffective once the enemy has made it deeper inside the territory as they are too far away, outside the borders. An alternative but similar idea I had was that any adjoining territories between allies would become shared. (That is, each ally would be able to build in those territories connected to theirs.) But regardless, I do think that co-operative play could benefit from any of the ideas mentioned so far.

Actually, your changes seem to have worked. 0AD is no longer crashing when loading any scenario maps, and mars_temple2 is displaying as expected in Atlas with no problems ^^

When adding one patch at a time, 0AD didn't start crashing until I copied the glsl shaders into "binaries/data/mods/public/shaders/glsl", at step 8 of your instructions. Setting "gentangents = false" and "preferglsl = false" also stopped the crashing.

I'm getting a segfault when trying to load the mars_temple2 model in Atlas, and also when loading certain scenario maps although not always. For example if I first try loading "Acropolis 2" then 0AD crashes, however if I load a map which doesn't crash, such as "Savanna Ravine", before loading "Acropolis 2", then "Acropolis 2" won't crash either. I've attached the call stack generated by gdb after reproducing the crash in Atlas. 0ad_atlascrash_modelmappingpatch_callstack.txt

Hi vts. I've just updated to revision 11983 and was able to set a chain of ~100 rally points with no problems, so I can confirm that your fix has worked. Thanks!

The selection circles do have a white border (see here), but they do look a little jagged even at the maximum allowed zoom. Perhaps a softer transition between the white border and the colour would look better..

This has fixed it, thanks! And I think that's all of them ^^

I knew I forgot something.. I'm using Kubuntu 12.04 (Linux 3.2.0-24-generic). I've attatched the call stack. 0ad_gdb_callstack.txt

Unfortunately it didn't fix the problem. With that change the terrain/entities are always black, just like if I had turned 'Enable Water Reflections' off. I saw something else whilst testing. When I rotate the view so that a black entity is in front of some grass, that entity's texture gets blended onto the grass. This happens with and without the changes to Renderer.cpp.

I encountered a crash when setting a sequence of rally points for a structure, using svn rev. 11951. Steps to reproduce: 1) Select a building which can produce or garrison units. 2) Hold down the Shift key, and set a chain of rally points anywhere. 3) Usually on about the 5th rally point, 0AD crashes. Occasionally, the crash produces this log: Function call failed: return value was -120000 (Handle index completely out of bounds) Location: h_mgr.cpp:197 (h_data_from_idx) Call stack: (0x7f4caa) ./pyrogenesis() [0x7f4caa] (0x79fc11) ./pyrogenesis() [0x79fc11] (0x7a08ee) ./pyrogenesis() [0x7a08ee] (0x7a00e9) ./pyrogenesis() [0x7a00e9] (0x7c33f3) ./pyrogenesis() [0x7c33f3] (0x7cad95) ./pyrogenesis() [0x7cad95] (0x5e81e4) ./pyrogenesis() [0x5e81e4] (0x675537) ./pyrogenesis() [0x675537] (0x65aade) ./pyrogenesis() [0x65aade] (0x65ad39) ./pyrogenesis() [0x65ad39] (0x65b62c) ./pyrogenesis() [0x65b62c] (0x6510c4) ./pyrogenesis() [0x6510c4] (0x65191f) ./pyrogenesis() [0x65191f] (0x587fa1) ./pyrogenesis() [0x587fa1] (0x41d9b3) ./pyrogenesis() [0x41d9b3] (0x4135d7) ./pyrogenesis() [0x4135d7] errno = 0 (No error reported here) OS error = ?

Okay, the water is displaying correctly now, but the black terrain and entities issue still persists. I'm not sure, I'm not too familiar with what LoS (Line of Sight?) is. Is it what your units can see? Panning the screen to an area where water is, even if that part of the map has not been revealed yet, will still cause everything to render correctly until I pan away from the water again. Although I did noticed that disabling the 'Enable Water Reflections' option in the menu causes the black terrain and entities to show regardless of whether the water is in view or not.

The framebuffer warnings have gone after applying your changes to PostprocManager.cpp, and the replacement files have fixed the errors. I have new one for you though: ERROR: Failed to compile shader 'shaders/glsl/water_high.fs': 0(56) : error C7011: implicit cast from "float" to "vec4" This is presumably what is causing the water to look undecorated, however I'm seeing another problem in addition to this, perhaps it is related. The terrain and some entities are black, but this only seems to happen when no water is being rendered. If some water is visible, then everything appears correctly. (except the water, of course.) I'm not seeing any other error messages (except that model_common_norm.vs one), nor framebuffer warnings. Hopefully that's the last of them.

I'm getting the following error messages on the main menu screen, with your new patch: WARNING: Framebuffer object incomplete (A): 0x8CDD WARNING: Framebuffer object incomplete (: 0x8CDD ERROR: Failed to compile shader 'shaders/glsl/fog.fs': 0(23) : error C7011: implicit cast from "vec4" to "vec3" 0(28) : error C7011: implicit cast from "vec4" to "vec3" 0(34) : error C7011: implicit cast from "int" to "float" 0(34) : error C7011: implicit cast from "int" to "float" ERROR: Failed to compile shader 'shaders/glsl/los.fs': 0(23) : error C7011: implicit cast from "vec4" to "vec3" 0(25) : error C7011: implicit cast from "vec4" to "vec3" And then these when I load a map: ERROR: Failed to compile shader 'shaders/glsl/model_common.fs': 0(132) : error C7011: implicit cast from "float" to "vec3" ERROR: Failed to compile shader 'shaders/glsl/model_common.fs': 0(132) : error C7011: implicit cast from "float" to "vec3" ERROR: Failed to compile shader 'shaders/glsl/model_common.fs': 0(132) : error C7011: implicit cast from "float" to "vec3" ERROR: Failed to compile shader 'shaders/glsl/model_common.fs': 0(132) : error C7011: implicit cast from "float" to "vec3" ERROR: Failed to compile shader 'shaders/glsl/model_common.fs': 0(132) : error C7011: implicit cast from "float" to "vec3" ERROR: Failed to compile shader 'shaders/glsl/model_common.fs': 0(132) : error C7011: implicit cast from "float" to "vec3" ERROR: Failed to compile shader 'shaders/glsl/model_common.fs': 0(132) : error C7011: implicit cast from "float" to "vec3" ERROR: Failed to compile shader 'shaders/glsl/model_common.fs': 0(132) : error C7011: implicit cast from "float" to "vec3" ERROR: Failed to compile shader 'shaders/glsl/terrain_common.fs': 0(156) : error C7011: implicit cast from "float" to "vec3" ERROR: Failed to compile shader 'shaders/glsl/terrain_common.fs': 0(156) : error C7011: implicit cast from "float" to "vec3" ERROR: Failed to compile shader 'shaders/glsl/terrain_common.fs': 0(156) : error C7011: implicit cast from "float" to "vec3" ERROR: Failed to compile shader 'shaders/glsl/model_common.fs': 0(132) : error C7011: implicit cast from "float" to "vec3" ERROR: Failed to compile shader 'shaders/glsl/model_common.fs': 0(132) : error C7011: implicit cast from "float" to "vec3" ERROR: CRenderer::EndFrame: GL errors occurred I'm not sure about the two warnings, but the errors look similar to the one I encountered when trying your last patch.