It's that damn Builder Brush causing my crashes! (Testers needed to verify)

Destrier on 8/3/2005 at 03:57

As promised prior- here is the results of installing a nVidia 6800 GT in the Dell XPS referenced before- After some initial configuration oddities, we had the system up and running- the card edging the ATI in performance on three programs: flight sim 2004 (FS9) being one of the most noticible.

After establishing that all was normal, we brought up Thief 3- and it ran without problem. We then moved on to the editor tests and performed the exact same procedures conducted prior on this machine under the ATI (where the crashes were constant and consistant) and the HP under an earlier nVidia card (where it ran with no crashes).

Results:
The Dell XPS with the nVidia 6800 GT ran the editor flawlessly- there were no crashes and no glitches. Both extensive brush subtraction and vertex manipulation were utilized and not once was there any problem whatsoever. We conclude therefore that the ATI 9800 Pro (XT) with the drivers listed prior are the cause of the subtraction crash. I conclude then that the Thief3 Editor and the ATI cards identified in this thread are incompatable with the editor.

One note of interest regarding the nVidia 6800 GT- for reasons unclear- despite impressive game graphics and a clear edge over the eariler ATI product in performance- it causes a strange artifact on the screen with black text on white backgrounds- as if the RGB guns in the tube are not quite aligned- a very noticible blue cast just right of the black letters. This monitor is shared with a Macintosh and I can switch back and forth between the two- the artifact is quite obvious with the XPS & nVidia. I will continue to investigate. Tests tomorrow will focus in on the video issue mentioned before.

In summary- the ATI Radeon 9800 with the 4.11 and 5.2 drivers are, to the best of our testing ability, apparently incompatable with the Thief 3 editor.

Hope this is has been of help.

Good Thieving to All.

Destrier

scumble on 8/3/2005 at 13:01

It's still possible to work with it if you're careful - it's a relatively annoying glitch we're talking about on the whole, considering the world of game editing in general. We can add the 6800GT to the list then...

Mandrake on 8/3/2005 at 13:18

Although I havn't done *extensive* brush manipulation testing, my Nvidia FX5900XT seems to be fine in the editor also, I havn't had any crashes while adding or subtracting the builder brush.

I AM able to get the editor to crash by subtracting or adding a brush, doing a build all, and then doing Undo (it crashes as I choose undo) but I don't know if this is a related bug or not....(doesn't always do it either, depending on what is in the map etc, but on a given map it is reproducable)

The error I get is:

Code:

App-Specific Logging Text
________________________________________________________________________________
Init: Name subsystem initialized
DestroyValue: Tried to destroy unlinked array property EnginePackages
Log: bspBuild built 10 convex polys into 10 nodes
Log: bspBuildBounds: Generated 9 bounds, 106 hulls
Log: bspBuild built 10 convex polys into 10 nodes
Log: Found 0 coplanar sets in 10
Log: BspMergeCoplanars reduced 10->10
Log: Polys: 10 -> 10
Log: Nodes: 10 -> 10
Log: Points: 16 -> 9
Log: Vectors: 10 -> 10
Log: Vertex stream 0: 48 -> 0
Log: Polys: 10 -> 10
Log: Nodes: 10 -> 10
Log: Points: 24 -> 24
Log: Vectors: 10 -> 10
Log: Vertex stream 0: 48 -> 48
Log: bspBuildBounds: Generated 9 bounds, 106 hulls
Log: bspBuild built 10 convex polys into 10 nodes
Log: Polys: 10 -> 10
Log: Nodes: 10 -> 10
Log: Points: 24 -> 24
Log: Vectors: 10 -> 10
Log: Vertex stream 0: 48 -> 48
Log: Found 1 zones
Log: Polys: 10 -> 10
Log: Nodes: 20 -> 10
Log: Points: 24 -> 24
Log: Vectors: 10 -> 10
Log: Vertex stream 0: 96 -> 48
Log: bspBuildBounds: Generated 9 bounds, 106 hulls
Log: BuildZoneInfo: 0 ZoneInfo actors, 0 duplicates, 0 zoneless
Log: Portalized: 0 portals, 0 zone portals (0 fragments), 1 leaves, 10 nodes
Log: BspOptGeom begin
Log: MergeNearPoints merged 8/24, collapsed 0
Log: Polys: 10 -> 10
Log: Nodes: 10 -> 10
Log: Points: 24 -> 16
Log: Vectors: 10 -> 10
Log: Vertex stream 0: 48 -> 48
Log: BspOptGeom building sidelinks
Log: BspOptGeom end
Log: Processed 0 T-points, linked: 48/48 sides
Log: Polys: 10 -> 10
Log: Nodes: 10 -> 10
Log: Points: 16 -> 16
Log: Vectors: 10 -> 10
Log: Vertex stream 0: 48 -> 48
Log: bspBuildBounds: Generated 9 bounds, 106 hulls
Log: Found 0 coplanar sets in 10
Log: BspMergeCoplanars reduced 10->10
Log: Polys: 10 -> 10
Log: Nodes: 10 -> 10
Log: Points: 16 -> 9
Log: Vectors: 10 -> 10
Log: Vertex stream 0: 48 -> 0
Log: Polys: 10 -> 10
Log: Nodes: 10 -> 10
Log: Points: 16 -> 16
Log: Vectors: 10 -> 10
Log: Vertex stream 0: 48 -> 48
Log: bspBuildBounds: Generated 9 bounds, 106 hulls
Log: bspBuild built 10 convex polys into 10 nodes
Log: Found 1 zones
Log: Polys: 10 -> 10
Log: Nodes: 20 -> 10
Log: Points: 16 -> 16
Log: Vectors: 10 -> 10
Log: Vertex stream 0: 96 -> 48
Log: bspBuildBounds: Generated 9 bounds, 106 hulls
Log: BuildZoneInfo: 0 ZoneInfo actors, 0 duplicates, 0 zoneless
Log: Portalized: 0 portals, 0 zone portals (0 fragments), 1 leaves, 10 nodes
Log: BspOptGeom begin
Log: MergeNearPoints merged 0/16, collapsed 0
Log: Polys: 10 -> 10
Log: Nodes: 10 -> 10
Log: Points: 16 -> 16
Log: Vectors: 10 -> 10
Log: Vertex stream 0: 48 -> 48
Log: BspOptGeom building sidelinks
Log: BspOptGeom end
Log: Processed 0 T-points, linked: 48/48 sides
Log: Polys: 10 -> 10
Log: Nodes: 10 -> 10
Log: Points: 16 -> 16
Log: Vectors: 10 -> 10
Log: Vertex stream 0: 48 -> 48
**** Mesh nodes: 1 (time: 0.00)
**** Mesh nodes after adding terrain: 1 (time: 0.00)
**** Mesh nodes after adding static mesh surfaces: 1 (time: 0.00)
**** Mesh nodes after pre-merge: 1 (time: 0.00)
**** Mesh nodes after height split: 1 (time: 0.02)
**** Mesh nodes after pre-merge: 1 (time: 0.00)
**** Mesh nodes after subtracting static meshes: 42 (time: 0.01)
**** Mesh nodes after removing nodes under terrain: 42 (time: 0.00)
**** Vertex elimination: 0 (time: 0.00)
**** Set node heights: 42 (time: 0.00)
**** TOTAL TIME: 0 (time: 0.05)
**** NODE BIN SIZE: 12 (time: 0.00)
**** NODE BINGRID SIZE: 64 (time: 0.00)
**** Links after adjacent link addition: 102 (time: 0.06)
**** Found stairs: 0 (time: 0.00)
**** Link data calculation (adjacent): 0 (time: 0.20)
**** Removed stair side-rails; nodes: 42 (time: 0.00)
**** Z-delta calculation (far): 0 (time: 0.01)
**** TOTAL: 102 (time: 0.27)
*** TOTAL NAVMESH BUILD TIME: 0.324219
**** Spatial DB build node matches: 0 (time: 0.00)
**** Spatial DB subdivide polys: 0 (time: 0.01)
**** node matching / dbg_find_lowest_unmatched_time: 0 (time: 0.00)
**** node matching / dbg_match_node_time: 0 (time: 0.00)
**** Spatial DB node matching: 0 (time: 0.00)
**** Spatial DB split and height adjust: 0 (time: 0.00)
**** Spatial DB small gridlet removal: 0 (time: 0.00)
**** Total # Gridlets: 30 (time: 0.00)
**** Gridlet size: 84 (time: 0.00)
**** Total Spatial Database size: 4440 (time: 0.00)
**** Spatial DB calc blocked cells: 0 (time: 0.08)
**** CalculateOpenness() / GetCellsWithValue(): 0 (time: 0.00)
**** CalculateOpenness() / dbg_no_neighbors_time: 0 (time: 0.01)
**** CalculateOpenness() / dbg_expand_all_time: 0 (time: 0.06)
**** CalculateOpenness() / dbg_delete_and_append_time: 0 (time: 0.00)
**** Spatial DB calc openness: 0 (time: 0.08)
Final NavMesh size: 9768 bytes (4860 bytes for nodes, 4908 bytes for links)
Log: bspBuild built 10 convex polys into 10 nodes
Log: bspBuildBounds: Generated 68 bounds, 587 hulls
Log: bspBuild built 10 convex polys into 10 nodes
Log: bspBuildBounds: Generated 68 bounds, 587 hulls
Log: bspBuild built 6 convex polys into 6 nodes
Log: bspBuild built 6 convex polys into 6 nodes
Log: bspBuild built 34 convex polys into 34 nodes
Log: bspBuild built 34 convex polys into 34 nodes
Log: bspBuild built 34 convex polys into 34 nodes
Log: bspBuild built 10 convex polys into 10 nodes
Log: bspBuild built 10 convex polys into 10 nodes
Log: Found 13 coplanar sets in 116
Log: BspMergeCoplanars reduced 116->94
Log: Polys: 123 -> 123
Log: Nodes: 153 -> 116
Log: Points: 252 -> 120
Log: Vectors: 10 -> 10
Log: Vertex stream 0: 696 -> 0
Log: Polys: 123 -> 123
Log: Nodes: 109 -> 109
Log: Points: 263 -> 263
Log: Vectors: 10 -> 10
Log: Vertex stream 0: 444 -> 444
Log: bspBuildBounds: Generated 67 bounds, 577 hulls
Log: bspBuild built 94 convex polys into 109 nodes
Log: Polys: 123 -> 123
Log: Nodes: 109 -> 109
Log: Points: 263 -> 263
Log: Vectors: 10 -> 10
Log: Vertex stream 0: 444 -> 444
Log: Found 2 zones
Log: Polys: 123 -> 123
Log: Nodes: 292 -> 109
Log: Points: 367 -> 263
Log: Vectors: 10 -> 10
Log: Vertex stream 0: 1184 -> 444
Log: bspBuildBounds: Generated 67 bounds, 577 hulls
Log: BuildZoneInfo: 0 ZoneInfo actors, 0 duplicates, 0 zoneless
Log: Portalized: 168 portals, 0 zone portals (0 fragments), 49 leaves, 109 nodes
Log: BspOptGeom begin
Log: MergeNearPoints merged 102/263, collapsed 0
Log: Polys: 123 -> 88
Log: Nodes: 109 -> 109
Log: Points: 263 -> 161
Log: Vectors: 10 -> 10
Log: Vertex stream 0: 444 -> 444
Log: BspOptGeom building sidelinks
Log: BspOptGeom end
Log: Processed 87 T-points, linked: 526/526 sides
Log: Polys: 88 -> 88
Log: Nodes: 109 -> 109
Log: Points: 161 -> 161
Log: Vectors: 10 -> 10
Log: Vertex stream 0: 950 -> 526
Log: bspBuildBounds: Generated 67 bounds, 577 hulls
Log: Found 7 coplanar sets in 109
Log: BspMergeCoplanars reduced 109->94
Log: Polys: 88 -> 88
Log: Nodes: 109 -> 109
Log: Points: 161 -> 52
Log: Vectors: 10 -> 10
Log: Vertex stream 0: 526 -> 0
Log: Polys: 88 -> 88
Log: Nodes: 113 -> 113
Log: Points: 164 -> 164
Log: Vectors: 10 -> 10
Log: Vertex stream 0: 460 -> 460
Log: bspBuildBounds: Generated 59 bounds, 599 hulls
Log: bspBuild built 94 convex polys into 113 nodes
Log: Found 2 zones
Log: Polys: 88 -> 88
Log: Nodes: 294 -> 113
Log: Points: 242 -> 164
Log: Vectors: 10 -> 10
Log: Vertex stream 0: 1192 -> 460
Log: bspBuildBounds: Generated 59 bounds, 599 hulls
Log: BuildZoneInfo: 0 ZoneInfo actors, 0 duplicates, 0 zoneless
Log: Portalized: 119 portals, 0 zone portals (0 fragments), 37 leaves, 113 nodes
Log: BspOptGeom begin
Log: MergeNearPoints merged 0/164, collapsed 0
Log: Polys: 88 -> 88
Log: Nodes: 113 -> 113
Log: Points: 164 -> 164
Log: Vectors: 10 -> 10
Log: Vertex stream 0: 460 -> 460
Log: BspOptGeom building sidelinks
Log: BspOptGeom end
Log: Processed 78 T-points, linked: 540/540 sides
Log: Polys: 88 -> 88
Log: Nodes: 113 -> 113
Log: Points: 164 -> 164
Log: Vectors: 10 -> 10
Log: Vertex stream 0: 950 -> 540
**** Mesh nodes: 24 (time: 0.00)
**** Mesh nodes after adding terrain: 24 (time: 0.00)
**** Mesh nodes after adding static mesh surfaces: 86 (time: 0.00)
**** Mesh nodes after pre-merge: 58 (time: 0.16)
**** Mesh nodes after height split: 111 (time: 0.12)
**** Mesh nodes after pre-merge: 66 (time: 0.27)
**** Mesh nodes after subtracting static meshes: 555 (time: 0.53)
**** Mesh nodes after removing nodes under terrain: 555 (time: 0.04)
**** Vertex elimination

So far I havn't had an editor crash while building a map. (I would never normally do an add/subtract brush, buildall, undo sequence)

bukary on 8/3/2005 at 13:37

Is there anyone who uses Cats 4.9 with Rad 9800 Pro? Perhaps you should try these drivers...

Bumbleson on 8/3/2005 at 23:27

Quote Posted by Destrier

I conclude then that the Thief3 Editor and the ATI cards identified in this thread are incompatable with the editor.

Could there be one last bit of hope (apart from a possible but improbable fix from ATI) in (un-)setting some hidden driver options with a tweaker? I remember how ATI users didn't have fog in Thief 2 for a long time until somebody (wasn't it Hit Deity?

Inline Image:

http://www.mainzelahr.de/smile/traurig/scratchchin.gif) discovered that setting table fog and Z fog together was the solution. Could it possibly be that easy?

Hit Deity on 9/3/2005 at 00:24

Actually, that was Northon, and then BlackRuin determined what it took to get it to work on our cards.

(http://www.ttlg.com/forums/showthread.php?s=&threadid=60992) First post by Northon called "Salvation for ATI Radeon owners..."

(http://www.ttlg.com/forums/showthread.php?t=66851&highlight=fog+table+radeon) BlackRuin's helpful fix.

It would be nice if ATI would jump on this right away. How about a mass emailing? :ebil:

jay pettitt on 9/3/2005 at 00:30

Thanks Durinda,

I've repeatedly tested my Radeon 9600 using wireframe mode and as many brush subtractions as I can cope with before going potty (ie > 50) followed by a build followed by switching back to textured render modes. And it works well.

While it would be nice not to have to worry about a workaround; considering that Thief 3 levels are mostly populated with static meshes rather than complex brush geometry, the bug really shouldn't prove to be much more than an annoyance to ATi users. We can edit FMs too. Hooray.

Hit Deity on 9/3/2005 at 02:02

I emailed Rage3d and ATI/support.....now we wait... :bored:

scumble on 9/3/2005 at 09:17

It's worth a try.

Bumbleson on 9/3/2005 at 23:41

Quote Posted by Hit Deity

I emailed Rage3d and ATI/support.....now we wait... :bored:

Do they still force you through a load of FAQ pages before allowing you to send your request, and then bombard you with even more FAQs instead of an answer? :eww: Good luck! :D :thumb: