fix: gadfly review — video-sized response cap, stale comment

maxVideoResponseBytes (512MB) replaces the shared 64MB JSON cap on the
/v1/videos/sync read path (doRaw now takes the cap per call) — 3/6
models flagged that a legitimate long/high-bitrate clip would be
discarded after minutes of GPU work. Plus a stale stable-diffusion
comment in initImageFilename and a test-handler early return.

Co-Authored-By: Claude Fable 5 <[email protected]>
Claude-Session: https://claude.ai/code/session_01AXQxVhXBw8PwFAtsVrXSmj

provider/llamaswap/llamaswap.go

@@ -47,6 +47,12 @@ const DefaultName = "llama-swap"
 // can't make a decode allocate without limit.
 const maxResponseBytes = 64 << 20
 
+// maxVideoResponseBytes caps the /v1/videos/sync body — the response IS an
+// encoded clip, which legitimately dwarfs any JSON/audio payload (a long
+// high-bitrate generation can pass 64MB). Still bounded so a buggy upstream
+// can't allocate without limit.
+const maxVideoResponseBytes = 512 << 20
+
 // Provider is a llama-swap client. It satisfies llm.Provider (chat, delegated
 // to provider/openai) and imagegen.Provider (image generation), and exposes
 // llama-swap's management endpoints as concrete methods.