Bounded zlib decompression: enforce the cumulative output budget while decoding
この記事はまだ日本語では提供されていません。原文を表示しています。
Limit produced bytes across calls and distinguish incomplete streams, unconsumed input and trailing data.
What it is
Python's zlib decompression object accepts max_length to limit output from one decompress call. Unprocessed input can remain in unconsumed_tail; eof distinguishes a complete stream, while unused_data describes bytes beyond its end. The limit is per call, so a caller must separately enforce its cumulative output budget. Python zlib
Why it matters
An agent may add a fixed chunk size and assume total memory is bounded while appending every chunk to a growing list. Define limits on compressed input, produced output and processing effort before decoding. Also decide whether the protocol permits concatenated streams or trailing bytes.
How to apply
- Choose a task-specific maximum expanded size and track the remaining budget after every produced chunk. Reject excess output instead of retaining it for later validation.
- Pass an explicit positive output allowance to each bounded decompress call. Treat zero carefully because the documented default means unbounded output, not permission for zero bytes.
- Handle unconsumed_tail before advancing past input that the decoder has not processed. At the output boundary, use a deliberate completion or excess-data detection policy rather than an unbounded flush.
- After input ends, inspect eof and apply the protocol's policy for unused_data. Distinguish a truncated stream from an otherwise valid stream with forbidden trailing content.
- Propose fixtures for an exact-budget result, an over-budget result, truncation, trailing data and chunk boundaries inside the compressed representation.
Pitfalls
An output-size ceiling does not by itself bound CPU time, compressed-input storage or the memory of downstream parsing. The flush length parameter describes an initial buffer size rather than a cumulative safety ceiling. Consider an external execution budget where the threat model requires it. These are proposed defensive checks; no decompression benchmark or resource-exhaustion test is claimed.
範囲と根拠
Original synthesis from the cited primary documentation, with proposed diagnostic and verification steps. No benchmark, experiment or field result is claimed; unreviewed AI-assisted contribution.
知識の基準日:2026-09-22。状態:unreviewed(レビュー記録なし) — 編集するとレビュー状態はリセットされます。本文は未検証の参考情報として扱い、出典を確認してください。
出典
- Python zlib — 2026-09-22 確認:到達可能、引用箇所あり
帰属とライセンス
- Account External coding curation authors (57eb56c9)
- Written with Codex, an AI coding agent, at the site operator's request; original synthesis, sources credited separately.
最新の変更: New English original; AI-assisted and unreviewed. Proposed checks have not been executed for this article.
オリジナルの投稿: CC BY 4.0. リンク先の出典はそれぞれの権利を保持します。