Add korge-fleks extension to gradle plugin #2352
Conversation
- Start to add more frame and slice info into ASEInfo - Update NewTexturePacker writeImage function
- All tilesets shall have the same size of 4096 tiles
🤖 Augment PR SummarySummary: Adds a new Changes:
Technical Notes: The tileset packer now supports duplicate-tile removal and represents empty tiles as 0x0 rects in atlas metadata. 🤖 Was this summary useful? React with 👍 or 👎 |
![augmentcode[bot]](https://avatars.githubusercontent.com/in/1027498?s=60&v=4){kind=small}
|
|
||
| // (1) Check if this frame is an image (animation) | ||
| imagesInfo[frameTag]?.let { imageInfo -> | ||
| saveImageInfo(imageInfo as LinkedHashMap<String, Any>, frameEntry as Map<String, Int>, frameTag, idx, animIndex) |
There was a problem hiding this comment.
saveImageInfo(..., frameEntry as Map<String, Int>, ...) makes lookups like frameEntry["spriteSourceSize"] try to cast a nested map into an Int, which will throw a ClassCastException at runtime. This likely needs a non-Map<String, Int> type for the frame entry data.
🤖 Was this useful? React with 👍 or 👎, or 🚀 if it prevented an incident/outage.
| } | ||
| else -> error("ParallaxImageBuilder - Parallax mode must be set to HORIZONTAL_PLANE or VERTICAL_PLANE in parallax plane '$planeName'!") | ||
| } | ||
| val index = if (attachBottomRight) offset else offset |
There was a problem hiding this comment.
val index = if (attachBottomRight) offset else offset ignores attachBottomRight (both branches are identical), so bottom-attached layers won’t differ in index computation. This looks like a logic bug given the parameter name and the differing indexForSpeed calculation.
🤖 Was this useful? React with 👍 or 👎, or 🚀 if it prevented an incident/outage.
| tileFramesInfo[tileIndex] = tileInfo | ||
| } | ||
|
|
||
| // Check if all tilesets consists of the expected amount of tiles |
There was a problem hiding this comment.
The tileSetMap validation runs inside the per-atlas loop, so it can fail when tiles from a tileset are distributed across multiple atlas pages. As written, it will require each tileset to be fully present after processing the current page.
🤖 Was this useful? React with 👍 or 👎, or 🚀 if it prevented an incident/outage.
| } | ||
| println(tileSetDataByUid) | ||
|
|
||
| return |
There was a problem hiding this comment.
This return exits exportTileMapLDtk before any level/layer processing, so the method is effectively a no-op beyond logging tileset info. If this isn’t intentional, it will prevent LDtk tilemaps from being exported.
🤖 Was this useful? React with 👍 or 👎, or 🚀 if it prevented an incident/outage.
There was a problem hiding this comment.
Intended, code beyond is not yet ready.
| val stackedTile = stackedTileMapData[tileIndex] | ||
|
|
||
| for ((idx, tile) in stackedTile.withIndex()) { | ||
| if (tile == 0) { |
There was a problem hiding this comment.
Using 0 as the “empty slot” marker means tile id 0 (which is a valid LDtk tileset index) can’t be represented and will be dropped when stacking tiles. The same tile == 0 sentinel logic appears in the other stackTilesIntoTileMap overload too.
Other Locations
korge-gradle-plugin/src/main/kotlin/korlibs/korge/gradle/korgefleks/AssetLevelMapExporter.kt:511
🤖 Was this useful? React with 👍 or 👎, or 🚀 if it prevented an incident/outage.
| fun SimpleBitmap.isEmpty(): Boolean { | ||
| for (y in 0 until height) { | ||
| for (x in 0 until width) { | ||
| if (this[x, y].r != 0) return false |
There was a problem hiding this comment.
SimpleBitmap.isEmpty() treats any pixel with RGB=0 as empty, so an opaque black tile would be considered “empty” and packed as a 0x0 rect. If the intent is “fully transparent”, this likely needs to consider alpha instead of RGB.
🤖 Was this useful? React with 👍 or 👎, or 🚀 if it prevented an incident/outage.
| val ninepatch = linkedMapOf<String, Any>() | ||
| // For nine-patches we only store the frame info of the first frame | ||
| if (slicename.isEmpty()) { | ||
| val ninePatchslice = aseInfo.slices.first() |
There was a problem hiding this comment.
aseInfo.slices.first() will throw NoSuchElementException if an ASE file has no slices, which bypasses the more descriptive error handling below. This could fail with a confusing stacktrace for missing-slice nine-patch inputs.
🤖 Was this useful? React with 👍 or 👎, or 🚀 if it prevented an incident/outage.
| ) | ||
| ) | ||
|
|
||
| assetLevelMapExporter.exportTileMapLDtk( |
There was a problem hiding this comment.
This test currently has no assertions, so it will pass even if exportTileMapLDtk produces no output (including the current early-return behavior). It may be worth asserting that expected entries are added to assetInfo for at least one layer/tile.
🤖 Was this useful? React with 👍 or 👎, or 🚀 if it prevented an incident/outage.
No description provided.