Move canvas mode GL mappings to renderer gl

This commit is contained in:
2026-06-02 09:00:53 +02:00
parent 4a7eff24bf
commit 466c1d0cc0
3 changed files with 72 additions and 32 deletions

View File

@@ -202,6 +202,9 @@ Known local toolchain state:
Desktop VR drawing also consumes backend-owned scissor/depth/blend state,
depth clear masks, active texture units, and fallback 2D texture unbind
targets while retaining the existing VR SDK/platform bridge shape.
Canvas mode overlay, mask, and transform paths also consume backend-owned
blend/depth state, active texture units, 2D texture copy targets, and RGBA8
readback format tokens.
- `windows-msvc-vcpkg-headless` validates manifest install/configure/build/test
for the current headless component matrix; see DEBT-0007 for remaining app
and platform triplet migration.

View File

@@ -746,6 +746,9 @@ Results:
depth clears, active texture units, and fallback 2D texture unbinds through
the renderer GL backend mapping; platform VR SDK bridges remain isolated for
later platform-shell extraction.
- Canvas mode overlay, mask, and transform paths now route generic OpenGL
blend/depth state, active texture units, 2D copy targets, and RGBA8
readback formats through the renderer GL backend mapping.
- Known remaining warnings: legacy project/vendor diagnostics, Visual Studio
vcpkg-manifest warning, `LNK4099` missing libyuv PDBs, and `LNK4098` runtime
library conflict from retained vendor binaries.