b482518

Not that it matters much, but I wholeheartedly agree. I see nothing wrong in striving to make this particular feature work a bit better than its original implementation.

Heads up, ...\\VC90\\MoveMapGen.vcproj seems to be missing the closing tag for <Filter Name="pathfinding" >

See also: http://udbforums.kicks-ass.net/index.php?topic=16056

If you are using Firefox, get QuickPageZoom. If you aren't... Well, you really should be.

How about a Scrabble addon for WoW, while you at it?

I run a very quick test of the latest revision and it looked like it had some new problems with certain areas. Zul'Gurub (309), and at least parts of Stranglethorn (try Booty Bay) and Elwynn Forest returned numerous "Path Build failed: invalid start or end polygon". I'll try to rebuild 000 when I come back, but rebuilding 309 didn't help.

I wasn't talking about 'lavafalls' but about pools of liquid lava in Molten Core that look nearly identical to ones in Blackrock Depths. BRD's lava streams (i.e., liquid pools with animated surface textures) hurt. Molten Core's lava didn't seem to do so. I thought that it might be useful, but if the issue is not related to vmaps, I really don't care about that particular map. The only time I ever touched that lava unintentionally was on during my first fight with Ragnaros on the official servers gods know how many years ago.

Ah, ok. From the following quote I assumed...

Try opening the solution and compiling g3dlite and zlib first.

With non-updated vmaps, everything seems to work almost perfectly. Thanks again, Faramir. Oh, in the unlikely case some of you didn't think of it, if you have a multicore processor, you can shorten your total building time by half if you run two or more (depending on the number of cores) instances of the generator on separate processor affinities at the same time. Just run one for --skipContinents true, and then another four for maps 0, 1, 530, and 571.

Molten Core's lava streams still don't seem to inflict any damage.

Hmm, I get numerous "No vertices to build tile!" warnings with various maps. The fastest way to reproduce is to try to build maptiles for 409:

Seriously though, how hard is it to get and compile SD2? Do it once, and with your approach, you won't even have to update it.

Great news. Is it necessary to rebuild the previously generated maps? REPLY TO BELOW. If you won't specify the parameters, it'll build with: skipContinents = true, skipJunkMaps = true, skipBattlegrounds = true, hiResHeightmaps = false, shredHeightmaps = true, debugOutput = false.

If this works, the same logic can be applied on Glyph of Blessing of Wisdom www.wowhead.com/spell=57979 www.wowhead.com/spell=19742#see-also-ability http://mangos.lighthouseapp.com/projects/18208/tickets/503 and Glyph of Blessing of Might www.wowhead.com/spell=57958 www.wowhead.com/spell=19740#see-also-ability http://mangos.lighthouseapp.com/projects/18208/tickets/504

Wikipedia has a policy that says that if a rule prevents you from improving or maintaining Wikipedia, you should ignore it. I don't see why Mangos should operate in a different manner. It seems like a highly useful feature. If someone doesn't like it, he can always choose not to enable it in mangosd.conf.

I noticed that HomeMovementGenerator ignores the navmesh data. Is there a reason for that or you just haven't gotten to it yet?

From what I see, MapBuilder::build() doesn't use shouldSkipMap checks and specifying the number directly seems to override skipContinents. Re to below: I finally see your point, but since there is absolutely no downside in specifying the argument, it seems rather moot. I use a batch file. Editing skipContinents out for no reason only to write it in later would be a waste of time.

Unless you have a monster machine, 3 hours seems way too quick. It took me about the same time to build the instance maps, and I'm sitting on Core 2 Duo 3.0GHz. Again, please try running --skipContinents true --hiResHeightmaps false 001, or at least actually see if your mmaps directory contains any map tiles that start with 001.

Clearly. This is why my original post explicitly says that I'm talking about continents 001, 530 and 571. Again, low res Eastern Kingdoms and all of the instances seem to build just fine. The instances can also be built with hiResHeightmaps set to true.

I assume you used --skipContinents true? Try building with: MoveMapGen.exe --skipContinents true --hiResHeightmaps false --debugOutput false 001 In my case, it crashes after loading the heightmaps, on "[33,42]: Intersecting meshes" (~1 min). EDIT: Are you on Linux or Windows?

Map 000 builds fine, but continents 001, 530 and 571 return immediate out of memory errors even with --hiResHeightmaps false. Tested on two machines, Win OS 32/64-bit, plenty of memory. @willisoady That information is classified.

Will, what does your log.txt say? Was it a writing error? See if your DataDir has an empty 'mmaps' folder, and if it doesn't, try creating one manually. I also had a "Memory manager returned NULL: out of memory" crash when I used --hiResHeightmaps true with --skipContinents false. If you get it, try setting hiResHeightmaps to false as a temporary workaround.

After finally getting it to work and testing it for about two hours, I strongly feel that this should be made priority number one and merged with the master as soon as it's humanly possible. The difference in gameplay is absolutely staggering and I can't think of any other underdeveloped feature more important than this. Faramir118, thanks for keeping the project alive.

Does anyone else has problems compiling the new vmap extractor in VS90? Something about stdint.h missing. EDIT: I found missing stdint.h in a different folder, but when I tried to copy it, I got a whole bunch of LNK2005 errors about recurrent function definitions. EDIT2: Never mind all this crap above. I found the source of the problem and it was local.