TES4Gecko 15.0 [Public read-only thread]

[sandor]

Superb. :goodjob:

I'll run some extensive tests tomorrow (customized test cases).



This is a major step forward!

[TheImperialDragon]

The exterior merging feature sounds like it will save a LOT of time! :goodjob: It takes both a very long time and a tremendous amount of effort to manually clean region files.

[SACarrow]

Oddly enough, it was quite simple. Looking at the code, it was a bit cavalier with the order of the cell child groups: temp was always last, but the VWD and persistent groups could be in either order. Looking at the OB ESM, that was not the case; persistent was always first, followed by VWD and then by temp. Once I changed the merge code to force that order, lo and behold! Areya's Palace was no longer falling into Cygnus X-1! Been on a Rush kick lately.

It obviously need more testing, but my test was to merge the VWD ESP with ST v3.12, and both Areya's place and the Steadhelm castle had actual architecture.

As a side note, I added the same "expand cell tree" functionality to worldspaces that the CELL group has; just right-click on the child group just below the WRLD record you want to expand on the plugin side.

Tell me what you think,
Steve

[sandor]

Originally posted by SACarrow


It obviously need more testing, but my test was to merge the VWD ESP with ST v3.12, and both Areya's place and the Steadhelm castle had actual architecture.

After sandor puts this through the wringer, I think this is the version to send to dev_akm. As a side note, I added the same "expand cell tree" functionality to worldspaces that the CELL group has; just right-click on the child group just below the WRLD record you want to expand on the plugin side.

Tell me what you think,
Steve

:goodjob:

I'll run a few tests, seeing your description (changed order) it should be fine.

[SACarrow]

Something I just added to my working version of Gecko is the ability to get a rudimentary report of all "objects" added by a plugin. I added this partly to have an offline way of searching while having the CS running and also as a possible steppingstone to some cell pre-population features I'm mulling over. It is very simple: a little header info, an intro for each group, and a list of form & editor IDs. It includes everything but INFOs & CELLs.

Here's a brief snippet:

TES4GECKO FORM ID REPORT for plugin <Silgrad_Tower.esm>
Object type: Global Variables [GLOB], 9 new form IDs found.
010054D7 STMusicPlaying
010054D8 STMusicTimer
010C0CB2 STFightersGuildReputation
010C1FF3 StSelbBashMe
011C346B StJWPaintercolor
011C346C StJWPainterPaid
011C346D StJWPainterChosen
011C346E StJWPainterTracker
011C346F StJWPainterWeapon
Object type: Classes [CLAS], 8 new form IDs found.
010C10B5 StSrSol0501Priest
01000DE2 STKESHC17priest
0118D2D8 StRFSHC38RetiredGuard
011C1160 StTKMerchSpellPotion
011C1CA5 STIJEbb11MerchAlcSmithNoRepair
011C1CA6 STIJEbb11MerchSmith
011C1CA7 STIJEbb11MerchAlcSmith
011C2A9D STBDSoluthisGuard

Thanks,
Steve

[sandor]

Originally posted by SACarrow
..... a popup appears that gives the options to:

• Merge all of the exterior cells,
  
• Merge none of the exterior cells,
  
• Merge some of the exterior cells, choosing by region. A list of all regions to which the cells belong is shown in plugin-worldspace-region format. The user can choose more than one region.
  

Since a cell can belong to multiple regions, if any of those regions match the chosen list, that cell is merged. This does not apply to persistent references, which are stored in a pseudo-cell, one per worldspace, and get merged regardless.

<p>I attached a few test files.

<p>#1 I used an existing region area and created a second region area outside the original area which overlaps with the Steadhelm area and generated the content.

<p>#2 I created a new region which overlaps with the Steadhelm area and generated the content.

<p>#3 I created a new region which overlaps with the Steadhelm area and added a castle (manually) to the SH area and one castle oustide the SH area within the new area.

<p>When merging these files to the ST esm I can't select the region areas.

<p>If I merge your ExteriorCellTest.esp (altering Tamriel) to the ST esm it's possible to select the regions. If I merge the same file to the OB esm I can't select the regions.

[SACarrow]

Yep, I found a couple of cut-n-paster errors that made it so no regions were found from the very master to be merged to. I did see a couple of oddities that I need your opinion on:

• One of the samples contained many cells without regions. How should those be handled? Right now they are only merged if ALL is selected.
  
• In #3, you update a region without a worldspace (StIdDeepenMeadowsRuggedRegion) with one and attempt to add cells with that region. The problem is that when Gecko looks at the region info, that is pre-merge, so there was no WNAM yet.
  

So the questions here are how should I treat cells w/o regions and regions w/o worldspaces? I'dthought that regions required worldspaces.

Dead on my feet,
Steve

[sandor]

Originally posted by SACarrow
So the questions here are how should I treat cells w/o regions and regions w/o worldspaces? I'dthought that regions required worldspaces.

Dead on my feet,
Steve

Only the cells within the desired region should be merged.
Cells without a region shouldn't be merged.

For regions w/o a WS --> Normally a region should be linked to a WS.

The presets plus the region areas (as per the coordinates of the claim description) have to be merged to the esm before the generation starts (as per the region guidelines). This ensures the preset is linked to the WS.

For some reason the Deepen Meadows areas weren't merged to the esm. The whole area has been generated but not merged.

[sandor]

Originally posted by SACarrow
I'll still be working on it, but only as requests (and bug reports) arise.

Steve

I could prepare an esp for another Gecko bug.

[SACarrow]

Please do so. It's not at that release stage, 'cause I've discovered another outbreak of the disappearing land bug. I'm trying a test case right now.

Needless to say, ready for release it is not.

Steve

Edit: As Arte Johnson would've said in the early '70s, "Veddy interesting..." I fired up a test using 6 cells: -33 & -34 for X and -9, -10, -11 for Y. The -33 column is in the region with ID 0109C472; the -34 column is senza ragione. I added a statue to the -9 and -11 cells and did minor landscaping to the -10 and -11 cells. I then merged and said to merge no exterior cells; this merge removed everything from the -10 and -11 cells. So just ignoring in Gecko (my current approach) cells in the plugin that have a LAND subrecord is totally frakked. :O Investigation continues.