Alan-Jowett / LexonGraph / 27376573373

82%

push

github

Add Zip64 support to zip block store (#91)

* Add Zip64 support to zip block store

Requirements:
- require Zip64 support in the rust-zip-block-store spec when the zip reader can open the archive
- update design and validation docs to cover successful Zip64 read paths and duplicate detection

Implementation:
- teach ZipBlockStore's central-directory scan to parse Zip64 EOCD metadata while preserving duplicate recognized-entry detection
- add Zip64 success coverage for constructor, get, and iter_block_ids

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

* Handle max-comment Zip64 locator scan

Widen the EOCD tail scan so Zip64 locators remain visible when a near-max EOCD comment pushes the EOCD to the start of the search window, and add a regression test covering that case.

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

* Tighten PR 91 Zip64 follow-ups

Cache the archive offset in ZipBlockStore so read-path operations stop reparsing the archive just to recover that value, and add Zip64 duplicate-entry coverage to preserve the duplicate detection requirement on the Zip64 path.

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

* Harden Zip64 central-directory parsing

Parse the central directory directly from the file stream instead of allocating buffers from untrusted Zip64 metadata, avoid preallocating from the reported entry count, and add prefixed Zip64 regression coverage for archive offsets and duplicate detection.

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

* Fix Zip64 locator fallback and spec wording

Fall back to classic EOCD metadata when a pre-EOCD locator signature does not resolve to a valid Zip64 EOCD for archives that do not require Zip64 sentinel fields, add a regression test for that case, and align the design spec wording with the implementation's direct central-directory inspection model.

Co-authored-by: Copilot <223556219+Copilot@users.nore... (continued)

192 of 272 new or added lines in 1 file covered. (70.59%)

9282 of 11138 relevant lines covered (83.34%)

246.52 hits per line