render3d: cut-out padding on every core - tex_dilate's rows per pass through Job.parallel_for, bytes unchanged

Within a pass a row writes only its own still-masked texels and reads only neighbours already let go,
which no row writes that pass, so the result is the single-threaded one. The worker takes a DilateJob
of plain buffers and allocates nothing. tex_dilate_bytes (safe_api) and examples/rendering/dilate.ludic,
which checks it against the old loop on RGB and RGBA atlases of sizes that do not divide (DILATE OK).

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

packages/ludic.render3d/safe_api.ludic

@@ -39,3 +39,8 @@ function png_decode_bytes(render3d_st: mut Render3dState, path: string, reuse: [
 function tex_upload_bytes(render3d_st: mut Render3dState, px: []byte, srgb: bool, mips: bool) -> int {
   return tex_upload(render3d_st, data_of(px), srgb, mips)
 }
+# cut-out edge padding over samples laid out as the last decode left them (tex_dilate)
+function tex_dilate_bytes(render3d_st: Render3dState, px: []byte, thresh: int, passes: int) -> void {
+  if px == null or len(px) < render3d_st.tex_w * render3d_st.tex_h * render3d_st.tex_channels { return }
+  tex_dilate(render3d_st, data_of(px), thresh, passes)
+}