fix(stdio): handle bufferless text streams in fallback

[sidinsearch]

Summary

Fixes a crash in stdio_server() when sys.stdin/sys.stdout are replaced with bufferless text streams (e.g., io.StringIO).

The _claim_fd function's fallback path unconditionally dereferences stream.buffer, but _is_backed_by_fd returns False when a stream has no .buffer attribute. This raises AttributeError before the transport serves a single message.

This is a distinct issue from the original #1933 (which was fixed by #3117). The original issue was about the wrapper closing the real stdio handles; this is about bufferless text streams (like io.StringIO) crashing the fallback path.

Changes

• src/mcp/server/stdio.py: Check for the buffer attribute in _claim_fd's fallback path; return None when the stream is bufferless. The caller (stdio_server) then wraps the text stream in place since there is no binary layer to re-encode.

• tests/server/test_stdio.py: Added test_stdio_server_serves_bufferless_text_streams_in_place which monkeypatches sys.stdin/sys.stdout to io.StringIO and verifies the transport serves messages correctly.

Test results

• All 5971 tests pass
• New test covers the regression case
• Ruff and pyright clean

Disclosure

Written with AI assistance, reviewed and verified by me.

Fixes #1933

[github-actions]

This PR has been closed automatically. This repo only keeps pull requests open when they come from a maintainer, or from a contributor a maintainer has assigned to the linked issue, and you aren't currently assigned to #1933.

If a maintainer assigns you to #1933, this PR reopens on its own and there's nothing more you need to do here. Assignment is a maintainer call based on capacity; comments that only ask to be assigned don't factor in. What does help is engaging on the issue itself by confirming the repro, explaining why it matters for your use case, or describing the approach you'd take.

You're welcome to keep pushing commits here (just avoid force-pushing, since GitHub can't reopen a rewritten branch), but that on its own won't get the PR reviewed or the issue assigned, and realistically most auto-closed PRs stay closed. There's no need to open a new PR either way.

CONTRIBUTING.md has the full reasoning, but in short:

• We're a small team with very little capacity to review community PRs right now.
• Many recent PRs are AI-generated with little human review, and reviewing one carefully still costs a maintainer as much time as it ever did. A well-described issue is usually more useful to us than the code.

Maintainers: reopen, remove missing-issue-link, or add bypass-issue-check to override.