mirror of
https://github.com/storytold/vectorcraft.git
synced 2026-11-20 01:14:23 +01:00
The import path set the reader's max_input_bytes to usize::MAX, so it accepted a PDF of any size and copied it whole before parsing. A real ceiling of 1 GiB keeps ordinary documents while refusing a file that would exhaust memory; stream decoding keeps its own per-document limits.