Shadowhide on 20/5/2010 at 17:13
Dromed set out an error after I computing database pathfinding :mad:
Problem solved after I deleted one secret door
hmmm.....
LarryG on 7/9/2010 at 15:50
Here's a new one to me. I get it when I have Light Bright on, am standing in a particular spot, and attempt to look at a map! (Nothing about the light of the new moon or standing on one foot while whistling La Marseillaise, but I'm sure that's just because DromEd is saving that up for a really special occasion:laff:) If I turn Light Bright off or stand someplace else when I look at the map, I don't get the crash, so I'm not really worried by it. But it is odd.
[ATTACH]642[/ATTACH]
Yandros on 8/9/2010 at 01:18
Quote Posted by LarryG
Here's a new one to me. I get it when I have Light Bright on, am standing in a particular spot, and attempt to look at a map! (Nothing about the light of the new moon or standing on one foot while whistling La Marseillaise, but I'm sure that's just because DromEd is saving that up for a really special occasion:laff:) If I turn Light Bright off or stand someplace else when I look at the map, I don't get the crash, so I'm not really worried by it. But it is odd.
Generally, assertions are not seen when playing through Thief2.exe, but in that case it suggests you're at or near the palette limit. Have you checked how many you have or done any palette conservation? given your high number of custom objects, it's something you'll no doubt have to contend with in your campaign.
LarryG on 8/9/2010 at 02:33
I've done some conservation, and was surprised to find just now that I am at 254. :weird: That's way more than I thought. Time to do some more common palette work with my textures. I thought I was doing a good job of reusing palettes, but I guess not good enough. :tsktsk:
Edit: I used NV's PaletteCounter.exe to id the 284 onesies-twosies in my txt16 folder. I then converted them all to 256 x 256 BMPs using IfranView's batch conversion utility, then Bright183 to create 284 .pcx files with 50 common palettes, then back to gif format using IfranView's batch conversion utility again. So far, I don't see any color or other degradations in the images ... crossed fingers. I'm now down to 220 palettes reported by DromEd.
LarryG on 13/9/2010 at 05:32
This one is another new one.
[ATTACH]649[/ATTACH]
It's associated with this line while running complete processing: "[linkarry.cpp@301] LINKID_RELATION(id) == Vec.Relation()"
After selecting cancel, re-doing complete processing and it doesn't happen again ... ???
It started happening after adding the Will-o-wisp, but it does have a TPathInit link to a TerrPt, and the TerrPts are all TPath linked in a loop.
LarryG on 13/9/2010 at 05:42
Right. I saw that. But that's not the case. I'm sure it is somehow related to the Will-o-wisp, but the indicated cause of a missing TPathInit or TPath link is not the case in this instance.
[ATTACH]650[/ATTACH]
EDIT: If it is any help, even though the Assertion Failed message is not repeating, I still see this in my Monolog:
Adding moving terrain objects.......................................................................................[linkarry.cpp@301] LINKID_RELATION(id) == Vec.Relation()
........................................................................................................................................................................................................
So, yeah, I'm sure its the Will o'Wisp. But the bloody thing is set up correctly and works just fine. Should I worry? Could it be because the winding path the thing takes between TerrPts sometime intersects objects and solid brushes? It seems to do no harm in-game, and if it want straight line like an elevator, there would not be any intersections with solids ... is there a way to narrow that wandering?
EDIT #2: I bet I know the reason: I haven't room brushed that area yet. I suspect that it wants a room to pathfind in. I'll verify that tomorrow.
Edit #3: Well that wasn't it. I'll try deleting and recreating the Will o'Wisp and its path. Thing is, despite the error message, the wisp seems to work just fine ...
