[s0767][mangos-one] vmaps v3 backport

[faramir118]

https://github.com/faramir118/mangos/tree/one_vmap_rewrite'>

https://github.com/faramir118/mangos/tree/one_vmap_rewrite (make sure you're on the one_vmap_rewrite branch)

I backported vmap v3 to mangos-one, along with some related code.

• Extractor works on client 2.4.3
  
• assembler produces valid collision data
  
• Core still needs to be tested (I don't have a db for this part)
  

Look on github, or read this summary of the commits backported.

[stfx]

Awesome. One thing I noticed when looking at the code is that https://github.com/mangos/mangos/commit/6a6cde8d284370b160952582378190b92c5ac2d4 has not been backported to mangos-one yet. Is it for 3.x only?

I think we can add atleast a few of those changes. For example the normals variable of model.h is no where really used.

Also

fwrite(&nIndices,sizeof(uint32), 1, output);

should rather be

fwrite(&nIndexes,sizeof(uint32), 1, output);

in model.cpp:88

[faramir118]

6a6cde8d284370b160952582378190b92c5ac2d4 is from Nov 22, 2008 - this was right around WotLK release, so I decided that commit was for 3.x compatibility.

My guess was confirmed when I tested mmaps with the 2.4.3 vmap data produced by my current github branch - navmeshes and debug output look perfect on the spot-checked maps. (except the seem to have reversed the .mdx triangle vertex order for client >= 3.x)

Fixes like the one you wrote are probably good to include though

[stfx]

Alright here is the partial backport for 6a6cde8d284370b160952582378190b92c5ac2d4: https://gist.github.com/802722

It does not change anything except renaming a few modelheader variables and removing the not used normals variable as well as the not used ModelTexUnit struct.

Changes left out in the commit:

- The removal of ModelVertex and ModelView structs as those are needed for 2.4.3 AFAIK

- The ofsViews entry which has those views in it

- BB_vertices, BB_indices and offsBB_vertices as they are not used

- The removal of origVertices as we need them to get the modelviews

Please tell me what you think

EDIT: Readded nD, ofsD, nI and ofsI based on http://www.madx.dk/wowdev/wiki/index.php?title=M2. The format now perfectly matches with the one on that page

[stfx]

Indoor/outdoor checks are working fine in town/bg/world which has always been an issue in my port

The only big issue which existed on mangos-one for like forever is non working liquid damage (lava in molten core, searing gorge) but that could also be .map related. Would be awesome if you can look into that as well

Tested with patch added in previous post and those small typo fixes I commented in github

[mns]

Great backporting

[VladimirMangos]

Great work! I not look yet code but I think this one from most important missing backport features. So i will prio review this now.

Big thanks!

[VladimirMangos]

Backporting package must also include

43f5bc2d0dc1e8952ee49b7b7228ba842df67d3e [9998] Fix *nix build.

For make it compile possible for Unix.

For simplify later backporting to zero (ofc will been more fixes for make this possible) i will backport

per master commit + your fixes. At least where possible (in pre-required commits for example) with cross check with your version.

[Lynx3d]

You accidently backported 3.x format of WorldMapArea.dbc too...core won't start with 2.4.3 dbc files.

After fixing this and a few other minor things line of sight and liquids are working, but indoor check does not work.

I think someone tried to backport it before and had the same problem. Unfortunately i don't really have a clue how arrai and tom_rus actually figured out the correct algorithm for 3.x so i'm pretty clueless on how to fix it.

[VladimirMangos]

ok, base at possibility disable this part by option we can live with it. And this will not worse that current Mangos One state.

If backport then this will local problem that can be fixed at some moment in future but other chnages and features dependent or implemented by vmap3 will not blocked.

[Oniryck]

Really really nice job. Big thanks to you guys! *bow*

[stfx]

You accidently backported 3.x format of WorldMapArea.dbc too...core won't start with 2.4.3 dbc files.

After fixing this and a few other minor things line of sight and liquids are working, but indoor check does not work.

I think someone tried to backport it before and had the same problem. Unfortunately i don't really have a clue how arrai and tom_rus actually figured out the correct algorithm for 3.x so i'm pretty clueless on how to fix it.

Yes, there are a few typos and I already commented them in github:

https://github.com/faramir118/mangos/commit/62ec65c676b90acbb9675c3c5a76bc95cf5d617a#src/game/GridMap.cpp-P29

https://github.com/faramir118/mangos/commit/583066196273f49422b575cbf5263a8ae7bca769#src/game/DBCfmt.h-P6

https://github.com/faramir118/mangos/commit/9fa9121af59527bbb27c8ace3e11accc06c45abe#contrib/vmap_extractor_v2/vmapextract/vmapexport.cpp-P503

https://github.com/faramir118/mangos/commit/9fa9121af59527bbb27c8ace3e11accc06c45abe#contrib/vmap_extractor_v2/vmapextract/adtfile.cpp-P65

Well the indoor check seemed to work in my tests but they were rather quick. Btw I was the one who contacted you about the indoor check issue a few months earlier when I ported vmap3 to oregoncore (2.4.3)

[VladimirMangos]

In https://github.com/VladimirMangos/mangos master - version with some fixes and changes.

I plan tommorow finish tests

[Oniryck]

Thanks Vladimir, I'll try to find the time to test it before tomorrow, check if I find any problem before release

[VladimirMangos]

vmaps 3.0 added to Mangos One (with disabled indoor check) in [s0767]. Thank you

indoor check problem as i look in debug in result different flags used in 2.x for AreaInfo::flags field

that get from class GroupModel

            uint32 iMogpFlags;// 0x8 outdor; 0x2000 indoor

Commented values not valid for 2.x

[stfx]

vmaps 3.0 added to Mangos One (with disabled indoor check) in [s0767]. Thank you

Good job

Btw did you check my partial backport of 6a6cde8d284370b160952582378190b92c5ac2d4 already? I think its pretty good one. It backports as much as possible without breaking 2.4.3 extraction. Patch can be found here: https://gist.github.com/802722

[VladimirMangos]

hmm, ok, i look at original commit and because hard backporting skip it from including into backport package . But i see you cleanup it so i will look

[Oniryck]

I tryied the https://github.com/VladimirMangos/mangos version of vmap3 and didn't noticed any change : all the bugs are still there (for example being sheeped near the flag in warsong gulch is 100% undermap), the mobs still half falls underground when climbing an inclined floor and the lava in and near molten core still doesn't do any damage. Is that normal or is there something wrong at my side? Or maybe all of this isn't vmap related, i'm not sure.

[VladimirMangos]

vmaps 3.0 added to Mangos One (with disabled indoor check) in [s0767]. Thank you

Good job

Btw did you check my partial backport of 6a6cde8d284370b160952582378190b92c5ac2d4 already? I think its pretty good one. It backports as much as possible without breaking 2.4.3 extraction. Patch can be found here: https://gist.github.com/802722

in [s0769] thank you.

[VladimirMangos]

I tryied the https://github.com/VladimirMangos/mangos version of vmap3 and didn't noticed any change : all the bugs are still there (for example being sheeped near the flag in warsong gulch is 100% undermap), the mobs still half falls underground when climbing an inclined floor and the lava in and near molten core still doesn't do any damage. Is that normal or is there something wrong at my side? Or maybe all of this isn't vmap related, i'm not sure.

But not new bugs? So all ok.

I more interesting that as i test: fishing work in cities. city and caves now correctly return area info. So many quests related things start work in normal way. And etc.

[VladimirMangos]

Just for note: i work currently for backport One -> Zero

[added]In [z1307]. Base at some testes problem only already noted for One: indoor checks.

[VladimirMangos]

In [s0772][s0773] restored work indoor checks so its enabled by default as in master.

[Oniryck]

In [s0772][s0773] restored work indoor checks so its enabled by default as in master.

Well done ! Thanks a lot!

[stfx]

In [s0772][s0773] restored work indoor checks so its enabled by default as in master.

Nice, but:

1. you can still mount up in flagarea in WS (tunnel works but flag area itself not)

2. you can mount up a few meters in front of the gate in AV (some more meters back into the cave does not work)

3. you dont get dismounted in the slime in Undercity.

The bg indoor checks worked with old flag checks and the slime issue is probably related to environmental damage missing (lava, slime) due to .map update (not vmap).

The only problem with the old checks was the fact that you got dismounted in cities like Orgrimmar or Undercity but everything else worked as far as I know.

[VladimirMangos]

3. you dont get dismounted in the slime in Undercity.

This is not related to vmaps indoor. It must work as normal dismount in liquid.

Old data no any indoor checks so if something work this has been not related to vmaps work in like cases.

Indoor check as feature exist only in vmap 3.0 version

About mounting: cleant by self control mount allowed check in indoor/outdoor case.

So server just not get any casts in like case if client think that are indoor.

vmap3 indoor feature just make accessable same info for server side.

This feature allow recheck mount casting (but only as anti-cheating because client by self not allow it)

When feature really used at server side only: when you attempt enter to indoor area at mount. This part not control by client.

If in point mount cast allowed by client then area isn't indoor area from client (and then server) point and wmaps indoor check not related to it.

mount not allowed but you can enter to are at mount - vmaps really detected bug. Please report exactly .gps coordinates where this problem (you enter mounted and stay in point are where you after unmount can't cast mount by client check protection).