IWTE - World editing - (General Discussion)

[wilddog]

Hi, the only other speed thing that may help you (initially) is change the bounding sphere calc to basic. I don't know if that's really causing a problem (it does show a timing for it) but may help a bit.

To get around the only the walls bit. They shouldn't be split across structures. You may find it quicker to change textures of unwanted structures to the wall texture. Combine those structures and then remove the bits you don't want. Its still unlikely to be that fast but should overcome the issue with the split structures.

[Arkay]

Hi, the only other speed thing that may help you (initially) is change the bounding sphere calc to basic. I don't know if that's really causing a problem (it does show a timing for it) but may help a bit.

Hmmm that did not really help much. Thank you anyway.

To get around the only the walls bit. They shouldn't be split across structures. You may find it quicker to change textures of unwanted structures to the wall texture. Combine those structures and then remove the bits you don't want. Its still unlikely to be that fast but should overcome the issue with the split structures.

Sorry wilddog but can't understand. What split structures? But it's me, I think I haven't been clear, sorry. All I want is to delete everything but walls and stuff (ambient_buildings, litter_pass). Using delete complex could create problems? Or is it fine?

[wilddog]

Hi Arkay
The crash you said you were having is because an object is split so it has (normally) an undamaged group in structure A and its damaged group in structure B.

eg this will show in the structure column with something like 2,3 as per object 949, 950 (where the two structures A and B are 2 and 3 in this case).

To get around this you could assign the same texture to both (ie assign the texture reference that structure A is using to structure B.
Then you could combine (just using the button) these structures. This should then show that the objects causing the problem were in the same structure.
You should now be able to delete that structure without getting the index error.

Deleting a complex in its entirety will not hit this problem as an objects groups are always within the same complex.

Which option you need would depend on which structures have walls etc that you want to retain.

Not really sure on the speed thing other than its a problem tied just to the exe creation rather than the code. If I can work out how to resolve that it should bring the speed back down again.

@JOX25 - I tracked down where the issue of the crash in setting that specific table block is. Just need to figure out a way of generating the block outlines that handles it correctly. Its tied to handling intersecting lines from the damage triangles in the 3rd group. Hopefully will resolve by end of the week.

[Mr.Jox]

Thanks, WD! Would it be possible to make invert function for collisions and animations(by X and Y). It just makes me a bit trouble to make my walls symmetric. (Perhaps you remember about my trouble from 0* to 90*?).

[makanyane]

Thanks, WD! Would it be possible to make invert function for collisions and animations(by X and Y). It just makes me a bit trouble to make my walls symmetric. (Perhaps you remember about my trouble from 0* to 90*?).

there already is the one for the collisions - it's the reflect X or reflect y buttons mentioned earlier - using both gives you same thing as rotating 180*

did you try using that for both assign collision and anim? That's the bit I'm not sure of (if the anim will obey the instruction to be mirrored like the collision - generally it follows the rotation/location of the collision but I'm not sure about that one). If the anim doesn't invert, then there isn't anything we can do about it in the near future - we (well wilddog actually) spent a lot of time trying to work out how to do new anims - but there's a sticking point because the animinstances content is compressed by a weird and unfriendly bit of software.

EDIT: Have tried using the reflection on an animated wall - it doesn't work - it attempts to invert the animation - unfortunately it looks like it confuses it into inverting everything and you get a wall with the battlements facing the correct way around, but sort of inside-out-ish, see pic;



so you can invert the collision to re-use them, but not if it's for anything animated. (you can if you want use un-animated walls, it just jumps between damaged states without the moving stuff...)

Just a question: Is it really a dial: Vanilla Med2 settlements have same textures (same type, path etc.) but they are like copied and used for different structures. This way i ahve few NE_litter_pass textures. Each one assigned to different structure but these texture don't have any difference. Also does it affect perfomance or any other thing?

It obviously doesn't seem to matter in vanilla - I'd think though that the best thing is to just reference the same texture instead of listing it again. Only exception would be if you need the alpha layer to be read in some and not others, where it'd probably be better to only use the 'alpha' setting for the structures that actually need it.

[Arkay]

@wilddog Thank you very much sir, got it now.

[Mr.Jox]

Just a question: Is it really a dial: Vanilla Med2 settlements have same textures (same type, path etc.) but they are like copied and used for different structures. This way i ahve few NE_litter_pass textures. Each one assigned to different structure but these texture don't have any difference. Also does it affect perfomance or any other thing?

[wilddog]

Just a question: Is it really a dial: Vanilla Med2 settlements have same textures (same type, path etc.) but they are like copied and used for different structures. This way i ahve few NE_litter_pass textures. Each one assigned to different structure but these texture don't have any difference. Also does it affect perfomance or any other thing?

A bit of explanation.
Complexes are set up in a tree like structure which is basically always the same number of levels (0,1 and 2) (except for models created via the DBM tools where they tend to use a variable tree size unlike vanilla). This level can be thought of a s a set of cubes in groups of 8 set as two layers of 4 around a common centre point. Level 0 is just the parent level, 1 means 8 cubes and 2 means each of those cubes is split down into 8 so now you have a 4 by 4 by 4 cube. Interestingly he DBM team exceeded this in some places even though there are fields to hold the value which are only a byte in size.

Each of these complexes has a number of objects which are split into groups and these object groups are spread over different numbers of structures. A different structure simply means a set of object groups which belong to the same complex and have a common texture. The first time a texture is referenced it needs the full set of details but thereafter a reference is often used.

Complexes also have a specific number of effects that can affect the object/groups that are in that complex. This is a maximum 69.

Collisions have a similar structure but break the tree down to specific sizes so may have more or less of a tree structure but are typically more. The tree structure is then constant for that model. Again the DBM team didn't keep the trees constant but instead I think just decided to use a tree structure that fitted the particular collision and then used that for the complex. This didn't seem to cause any harm but was just completely non standard which is why IWTE couldn't handle the DBM models for the EDB team initially.

In terms of performance the engine may well use some of the characteristics of the tree sizes to help in specific areas e.g. like lighting effects or collision setting. We only really noticed some issues when complexes were set up which overlapped each other considerably. We did get some weird lighting effects (disconnects of the models) but that was a few years back. After that we never really tried making 'none standard' complexes or collisions. Its only looking at Leos and King Kong's TATW that we saw that they hadn't followed the vanilla setup and instead used basically a single complex rather than following the vanilla setup.

Hope that explains things a bit more.


On the animations thing, I thought that the collision matrix might of handled the animation as well but the problem on the animations is that each animation is made up of different groups of vertices/triangles. These groups then have different types of animations. The standard displacement and quaternions that other movements have and then some have just rotations but with different parameters which control the swing movements e.g. gates all have this type and different falling/rotating broken bits of walls. I think its these latter bits which are mostly causing the issues on the animation and confused the hell out of me when trying to connect to milkshake (I know they still don't work correctly). They also have lighting effects built into them which may or may not apply. So all in all the animation bits seem way to complicated o get too involved in other than some fairly simple animations. I was planning on using some but decided it was too much work (for my limited artistic skills) to try and put together 6 different sets of tumbling walls

[wilddog]

@ Jox25
I'm going to change the process on the setting blocks. Think I found the issue causing the problem you encountered so just want to test it out further against some other walls. Should hopefully bring this 'bug' to an end. I also want to address the non deletion of the breaches but that should be pretty straight forward by comparison.

Issue here was trying to reduce the connected points and I think I was doing it a stage too early causing a mismatch later if similar points ended up 'splitting further apart'. They then failed a continuous join test leading to the error.

[leo.civil.uefs]

Some information I've collected after my last post on the possible techtree issue (I said IWTE was deleting my tectrees without any reason and then they wont load ingame).

Not exactly that, the techtree dissapears aparently after I make any changes on the settlement using IWTE, the tech tree wont show ingame. However, closing the game, deleting map.rwm and running the game again and then the same custom settlements makes the techtree to appear again.

that happened in different settlements.

strange.

[makanyane]

Some information I've collected after my last post on the possible techtree issue (I said IWTE was deleting my tectrees without any reason and then they wont load ingame).

Not exactly that, the techtree dissapears aparently after I make any changes on the settlement using IWTE, the tech tree wont show ingame. However, closing the game, deleting map.rwm and running the game again and then the same custom settlements makes the techtree to appear again.

OK so that doesn't sound like it can actually be down to IWTE. Presumably each time you open the settlement with IWTE it still shows the techtree/s you added?

Two things;
A. Are you using the techtrees to load 'slot-fillers' or actual buildings named in the export_descr_buildings file? If actual buildings, then the confusion is probably down to whether the game thinks that building is built in the settlement or not.

B. If you change a .worldpkgdesc file it's not actually read until M2TW launches from scratch - if you leave M2 up and just start new game/battle then changes won't show - and perhaps if you have a cache problem they might not show - if it is a cache problem that might be why deleting map.rwm fixes it.


(one other thing - if you're trying to load an actual building you can only have one sized slot working for it - so you can't have a 32x32 version and a 64x32 version in a different settlement for the same actual EDB building, only one of them will show up - this does not affect slot-fillers)

[leo.civil.uefs]

Presumably each time you open the settlement with IWTE it still shows the techtree/s you added?

O wait... I remember having to add it again in gundabad as it was deleted. :/
There is two of the same techtree in gundabad, the mountains you see on each side.
And in case of gundabad yes, it was not showing up even in IWTE anymore, it was deleted.

Ask me for more information if you need/want and I can test it again.

A. Are you using the techtrees to load 'slot-fillers' or actual buildings named in the export_descr_buildings file?

slot filler it is.

B. If you change a .worldpkgdesc file it's not actually read until M2TW launches from scratch - if you leave M2 up and just start new game/battle then changes won't show - and perhaps if you have a cache problem they might not show - if it is a cache problem that might be why deleting map.rwm fixes it.

Then this will be not a problem for my mod as it automatialy deletes .rwm on installation

(one other thing - if you're trying to load an actual building you can only have one sized slot working for it - so you can't have a 32x32 version and a 64x32 version in a different settlement for the same actual EDB building, only one of them will show up - this does not affect slot-fillers)

not the case here.

[wilddog]

Some information I've collected after my last post on the possible techtree issue (I said IWTE was deleting my tectrees without any reason and then they wont load ingame).

If you were to generate the text file and simply look at it and search for techtree you would be able to know if the file contained the techtree or not and whether IWTE deleted it or not. Hope that helps but its pretty simple to do.

@ Jox25 - I found the issue on the the deployment block and need to test it then add in one new routine. Had hoped to finish it for Sunday but haven't had any time this week as I messed up something in the other mod I was doing.. and its the Rugby this weekend. Expect sometime next week to test and finish it. Should also address the non deletion of the breaches.

Edit :Think I've finally sorted out the deployment block issues. Will try some more tests in the week and hopefully do the breach deletion.

[Mr.Jox]

Thank you, wd for your hard job! Btw i wanted to find out: looks like med2 engine supports destruction of building. But how can i initializate destruction of building in game? Catapults etc. dont destroy buldings..

[makanyane]

Btw i wanted to find out: looks like med2 engine supports destruction of building. But how can i initializate destruction of building in game? Catapults etc. dont destroy buldings..

Catapults can target techtree buildings i.e. when you have a techtree slot and the building that shows up there is a named thing built in EDB like "milita barracks", you can't target the slot-fillers that Leo has been using as a trick to get decorative settlements to appear outside settlements...

the damage to proper techtree slot EDB buildings will appear in the settlement screen after the battle, but it'll only damage the building I don't think they can be totally destroyed that way (so a lvl 5 barracks can be repaired, the owner is not forced to start from scratch again)


there's ways via script to totally destroy a range of buildings throughout a faction, but I don't think that's what you want...

[Mr.Jox]

All i want is just make demage to buildings. Only techtrees and no others ?

[makanyane]

All i want is just make demage to buildings. Only techtrees and no others ?

The only things you can ever actually TARGET are the 'main objects', gates, walls, gatehouses, arrow_towers. and then the Techtree slots that have EDB linked buildings in them.

You can make any structure damageable by giving the objects two groups, one for initial state and one for damaged state, and making/assigning a collision to it. With that you can also assign the nice burning/smoking effects to the damaged state so they appear when it's hit. But you still can't actually target it! So the damaged state will only appear when artillery misses a nearby target that it was aiming at (any of the above structure types or units). I think that sort of damage is only decorative and doesn't add to the amount of repairs needed afterwards - not sure though...

Gatehouses, (and Arrow Towers that aren't part of the walls structure), don't seem to have a lot of effect on the ai, so you could if you wanted to get something else blown up make it a gatehouse/tower... but it still needs to have two groups per object, or it's got nothing to switch to once hit.

[Mr.Jox]

Thanks, this is what i needed to be explained

[leo.civil.uefs]

may I suggest something.

As mak recently discovered the height of the settlement can totally compromise the AI functionality (see here).

The way we found to fix this is through IWTE by world vector change, very simple and functional, thanks IWTE.

Problem is that you also need to elevate the terrain itself otherwise the buildings would be floating. Its simple to do it by creating a terrain .ms3d but what about when you do have a settlement larger than 179x179 (ms3d limit) like 311x311? You need then to resize your terrain reducing it to ms3d limit (179x179), then create the ms3d terrain file from this resized terrain, elevate it in ms3d, use the 3d file to create a new .worldterrain file and then resize it again to to its original size (311x311), but as we know this way we will loose all the edition beyond the 179x179 borders, needing to do it all again.

Resuming, a command to "elevate" the terrain directly through IWTE itself would be welcome. If possible.

[makanyane]

It's already in there Leo - somewhere around that specials tab and the vector change stuff - am 'multi-tasking' at the moment and can't load IWTE - will look and give exact details later...


EDIT; the button next to "world vector change" called "world terrain vec/rotation" does height change and rotation by 90degrees
(only thing it doesn't do is a horizontal shift as that gets too dodgy about what to put in/crop and it can only move in 8m jumps which might not tie up with world change)