[chdre]

EmailDocumentBackend._msg_to_rfc822_bytes converts an Outlook .msg into an RFC 822 message and copies every attachment's bytes into it. The backend only ever reads the filename and MIME type of attachments, so the copy is wasted. The attachment content is held in memory several times: in the .msg object, in the EmailMessage, as the serialized bytes, and as the parsed message.

This change passes an empty payload to add_attachment and keeps the filename and MIME type. Output is unchanged, list_attachments still works, and no new option is added.

Checklist

• Documentation has been updated, if necessary. (not needed)
• Examples have been added, if necessary. (not needed)
• Tests have been added, if necessary.

Review evidence

Measured with a full DocumentConverter conversion of a 60 MB .msg with one large attachment, list_attachments on, median of 3, tracemalloc, Windows, Intel Core Ultra 9 285HX. The file is private and not part of the PR.

	Peak memory	Wall time	Markdown
Before	723.1 MB	7.3 s	identical (sha256 prefix)
After	185.9 MB	1.6 s	identical

With list_attachments off, peak memory is also 723.1 MB before and 185.9 MB after, with identical output.

New tests in tests/test_backend_email.py:

• test_msg_to_rfc822_keeps_attachment_metadata_without_payload: filenames and content types remain, and neither the raw payload nor its base64 form is in the output. It fails on main.
• test_msg_attachment_listing_leaves_the_body_unchanged_end_to_end: the body text is the same with listing on and off, and the listed attachments are still correct.

Commands run:

• uv run pytest tests/test_backend_email.py: 29 passed.
• uv run prek run --files ...: all hooks passed.

Not run:

• The full test suite.

Not included:

• .msg files with embedded message attachments were not tested.

[github-actions]

✅ DCO Check Passed

Thanks @chdre, all your commits are properly signed off. 🎉

[mergify]

Merge Protections

🟢 Merge protection satisfied — ready to merge.

Show 1 satisfied protection

🟢 Enforce conventional commit

Make sure that we follow https://www.conventionalcommits.org/en/v1.0.0/

• title ~= ^(fix|feat|docs|style|refactor|perf|test|build|ci|chore|revert)(?:\(.+\))?(!)?:

[github-actions]

AI triage (advisory)

Checked commit 5d6d919152. The analysis read the diff only. It did not run code from this pull request.

This pull request avoids copying attachment file bytes when EmailDocumentBackend converts .msg files to RFC 822 format. It retains only attachment filenames and MIME types to reduce peak memory usage.

Topic labels added: performance.

CI safety: no concern found. A maintainer must still approve CI.

_{The ai:* labels are advisory. They do not approve CI and they do not add tests:full. A new push runs this check again.}