TES4Gecko 15.0 [Public read-only thread]

[sandor]

I'll post the file tomorrow including a short description.

[SACarrow]

To edit my edit: Nothing is actually missing but the LAND record, so the land is set to the default. Trees and rocks are still there.

Steve

[SACarrow]

Here is a new 7Z that seems to fix the issue I mentioned above; I've also included the ESP I descibed above for testing. The problem lay in the fact that merging occurs in two phases:

• Mapping, where new form IDs and questions of what overrides what are determined,
  
• Updating, where the answers decided upon above are executed.
  

The thing is that there are some anticipatory actions taken in mapping, like deleting one cell's LAND record if it is going to be "merged over" by another plugin. My approach was to simple ignore all desired exterior cells and associated groups; this worked in the updating phase but not the mapping phase. This meant that a LAND record could be deleted in anticipation of a merge update that would never come. Now ignored records stay ignored throughout the merge process. To replicate the error, take the ESP here and merge it to ST with the previous version, selecting that no exterior cells be merged. Lots of pink and floating stuff will be visible in the 4 cells with landscaping.

<p>Steve

[sandor]

Sometimes modders want to delete cells e.g. Razorwing when he's reorganizing Soluthis etc.

The attached test file deletes a cell within the ST esm.
After the merge the content of the cell has been deleted, the cell remains.

It's not a huge problem, but I always have to manually delete the cell using the Gecko.

[SACarrow]

I think. Here the issue seems to be that upon merging, the CELL, WRLD & DIAL objects are not ignored when merging even if they are deleted, when is exactly when that object both in the plugin and master need to be ignored. I remedied this and your test case worked. I attempted to delete a WRLD from ST as another test, but they were all so interconnected, being in real use and all. So I deleted TestTree and eyeballed the result, which looked good. As for DIALs, removing them in the CS doesn't actually mark them as deleted, so I have no idea of what to do there.

HTH,
Steve

[TheImperialDragon]

My bad about the region areas for Deepen Meadows, they'll be in the main ESP. I hope that's okay.

I don't think my Oblivion plane worldspace (STIDOblivionPlane01) is hooked up to anything yet, so you might be able to use it in your tests if you need to delete a worldspace from the ST ESM.

[SACarrow]

Originally posted by TheImperialDragon
My bad about the region areas for Deepen Meadows, they'll be in the main ESP. I hope that's okay.

Nah, we're good. When one is working on something like Gecko, one needs to know what's possible, not just what's recommended. I didn't even know it was possible to have CELLs sans REGNs or REGNs sans WRLDs until this. Now both cases are (kindasorta) dealt with; you can't merge only CELLs senza ragione, but since sandor says those aren't recommended, I'm OK with that.

Thanks for the WRLD offer BTW. The concern in the code was, I believe, because those 3 types all have corresponding groups that immediately follow the objects. These groups cannot be ignored simply because their lead object has been ignored; however for WRLDs and CELLs, deleting the lead object in the CS does in fact also delete the following group. As stated above, I could not find a way to delete a DIAL via the CS.

Later,
Steve

[sandor]

I merged the Deepen Moor area with the latest Gecko version.
There were alterations to 4 existing region areas (Soluthis and Silgrad city).

After checking all of the border cells, the merge was perfect (no rips, floating stuff and so forth).


SACarrow this a major breakthrough!! :bananarock:

I'll run a few more tests regarding the things you mentioned in your previous post(s).

[SACarrow]

I'm playing with the idea of being able to clone a worldspace; it would be similar to the quest-cloning approach, involving the copying of exterior cells and their groups, followed by renaming and renumbering. I thought of this when thinking about cloning the area around FishIsland to a WS for LOD generation. Does this seem like an idea with application or a solution in search of a problem?

Steve

[sandor]

Originally posted by SACarrow
I'm playing with the idea of being able to clone a worldspace; it would be similar to the quest-cloning approach, involving the copying of exterior cells and their groups, followed by renaming and renumbering. I thought of this when thinking about cloning the area around FishIsland to a WS for LOD generation. Does this seem like an idea with application or a solution in search of a problem?

Steve

It could be useful for city building (creating backups/for reference purposes).