Redesign texture pipeline around one canonical handle

[JeanPhilippeKernel]

Summary

RRM's own ImageHandle/m_image_slots texture system had zero real consumers and was disconnected from what actually renders (materials sample via a raw bindless index into Device->GlobalTextures). This deletes it in favor of Rendering::Textures::TextureHandle everywhere, and fixes the three structural gaps that came with it:

• No working texture disposal — TextureHandleToDispose had a consumer but no producer. VulkanDevice::DestroyTexture is now the sole producer, timeline-gating both the VkImage free and the bindless slot reclaim in Present().
• No real hot-reload trigger — AssetManager::IngestTexture's dedup silently blocked re-ingest. A new TextureImporter routes texture files through ImportCoordinator like every other asset type, and a dedup hit now calls RenderResourceManager::ScheduleTextureReload instead of no-op'ing.
• No reference safety — materials stored texture refs as a bare uint64_t index with no generation. AssetManager::ReleaseTexture/FlushTextureReleases patch every referencing material to the INVALID_MAP_HANDLE sentinel before the underlying bindless slot can ever be reused.

Also included:

• VulkanDevice::ReconstructTexture generalizes the in-place resize pattern (same handle, same slot) that was previously duplicated inline in RenderGraph::Resize.
• TextureHandleToDispose is now a lock-free SPSC queue since producer and consumer are both render-thread only.
• AssetRegistry::InferTypeFromExtension recognizes all 8 raster extensions TextureImporter claims, plus the pre-existing .exr gap.
• Image2DBuffer renamed to ImageBuffer — it holds 2D, cube, and array images, not just 2D.
• Removed the fully dead Texture2D.h/.cpp (a superseded Ref<T>-based texture class, zero callers anywhere).

Adversarially reviewed across 4 parallel passes; 2 real bugs found and fixed (an unlocked concurrent read of AssetManager::Textures[], and a missing arena Clear() in TextureImporter that would have grown unboundedly and crashed on exhaustion).

Test plan

• zEngineLib, Obelisk, ZEngineTests all build and link cleanly
• 554/554 tests passing (5 new: AssetRegistry extension-classification + OnRemoved callback firing, TextureImporter::CanImport coverage)
• Verified live in Obelisk under an aggressive resize stress test (3 escalating rounds, ~100+ resize events including near-zero dimensions) — no leaks, no crashes, same PID throughout
• Manual texture hot-reload verification in the running editor (tracked separately — the headless VulkanDevice test fixture needed to un-skip the 4 RenderResourceManagerHotReloadTest cases is still open)