How does TubeNotes handle canvas memory management and page splitting during long video transcript PDF generation?

[PrinceBad]

When compiling multi-hour university lecture notes into structured PDFs via jsPDF on the client side, what architecture prevents DOM canvas memory leaks and page boundary overflows?

[PrinceBad]

Canvas Offscreen Management & Vector Splitting in TubeNotes

1. Virtual Page Height Buffering:
   TubeNotes renders document sections in a virtual OffscreenCanvas context rather than the active DOM tree. This ensures zero layout thrashing or browser DOM memory buildup during massive document compiles.

2. Dynamic Flow Splitter:
   Text paragraphs, code blocks, and diagrams calculate line-height heights before rendering. If currentY + elementHeight > PAGE_MAX_Y, an explicit pdf.addPage() is pushed and cursors reset to margin bounds.

3. Garbage Collection & Blob Disposal:
   Once binary array buffers are passed to URL.createObjectURL(), canvas contexts are manually cleared and image handles nullified to trigger browser V8 GC collection immediately.