Add reject_continue to PlaceResult and make SKIP short-circuit in ofChain (#63)

ofChain survival placement had a bug where SKIP (returned when a block
already exists) would not stop the chain, causing fallback alternatives
to replace already valid blocks , for example:
<img width="436" height="477" alt="image"
src="https://github.com/user-attachments/assets/7ff3d6fa-de98-4716-823f-3ad2abab7394"
/>

SKIP now short-circuits the chain, remaining alternatives are skipped
because the position is already valid.
To keep correct behavior for elements without autobuild (which
previously returned SKIP as a "no implementation" signal), a new
REJECT_CONTINUE value is added, which: continues silently like SKIP (no
error, not counted) in the walker and in the chain tries the next
alternative like REJECT
Default survivalPlaceBlock and IStructureElementCheckOnly now return
REJECT_CONTINUE instead of SKIP for unhandled paths.

Tested this with the example above, and fixed the issue:
<img width="435" height="498" alt="image"
src="https://github.com/user-attachments/assets/d92bce4d-b5ad-4051-af79-ffd8c819f9f7"
/>
survival placement also still works:
<img width="1234" height="916" alt="image"
src="https://github.com/user-attachments/assets/632d5556-524d-4dc5-a7f5-75c859cb6b70"
/>

this should ideally merged before
https://github.com/GTNewHorizons/GT5-Unofficial/pull/6922 , to fix this
underlying bug, but if possible not be tagged without the other PR,
which would cause all NEI previews of multis with water to not show any
water anymore

---------

Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
This commit is contained in:
Yoshyandgithub-actions[bot] authored and GitHub committed 2026-06-06 18:10:11 +02:00
1 parent 95d1533342
commit 287e8317c9
5 files changed
+37 -27

No files matched your search

@@ -96,7 +96,7 @@ public interface IStructureElement<T> {
"Default implementation of survivalPlaceBlock hit! Things aren't going to work well! IStructureElement class: {}",
getClass().getName());
if (!StructureLibAPI.isBlockTriviallyReplaceable(world, x, y, z, actor)) return PlaceResult.REJECT;
return PlaceResult.SKIP;
return PlaceResult.REJECT_CONTINUE;
}
/**
@@ -206,7 +206,7 @@ public interface IStructureElement<T> {
if (StructureLibAPI.isDebugEnabled()) LOGGER.info(
"Fallback shim code of survivalPlaceBlock hit! Things aren't going to work well! IStructureElement class: {}",
getClass().getName());
return PlaceResult.SKIP;
return PlaceResult.REJECT_CONTINUE;
}
/**
@@ -268,16 +268,21 @@ public interface IStructureElement<T> {
enum PlaceResult {
/**
* This element either exists already, or does not yet have an implementation for survivalPlaceBlock, or some
* other unforeseen situations. TODO this definition doesn't seem right. Should we separate SKIP from EXISTS, or
* SKIP from ERROR?
* The element's check passed, the position is already valid. No placement is needed. In the walker: continue
* iteration (no error, not counted as placed). In ofChain: short-circuits, remaining alternatives are skipped.
*/
SKIP,
/**
* This element's space is occupied by other stuff and require player attention, or player missing required
* resource, or the block cannot be placed due to obscure mechanisms, or some other unforeseen situations.
* The element cannot handle this position (occupied, missing resources, etc.). In the walker: stop iteration
* with error particles. In ofChain: try next alternative.
*/
REJECT,
/**
* This element has no survivalPlaceBlock implementation (default fallback), or otherwise cannot place but is
* not an error condition. In the walker: continue iteration silently (like SKIP). In ofChain: try next
* alternative (like REJECT), so a better-suited element can handle it.
*/
REJECT_CONTINUE,
/**
* Autoplace cannot proceed within this tick. To proceed further would require stopping the placement and wait
* for next round.