Hi team! 👋 My first contribution to this project!

This PR picks up the feature discussed in old #5608 to support Newline-Delimited JSON (NDJSON) in the existing JSON content view, following the approach suggested by @Prinzhorn and @mhils.

When inspecting Elasticsearch _bulk payloads, LLM/streaming responses, or log streams formatted as NDJSON, mitmproxy previously fell back to raw/text view because json.loads failed on multiple concatenated JSON records.

Changes

• Graceful NDJSON fallback: In JSONContentview.prettify, if parsing the whole payload as a single JSON document fails, we now attempt to parse each non-empty line as a distinct JSON object. Each record is formatted with 4-space indentation and joined with newlines.
• MIME type detection: Added application/x-ndjson and application/ndjson (as well as *ndjson subtypes) to render_priority so NDJSON streams are recognized automatically.
• Error handling: If the data contains invalid JSON on any line or is completely empty/whitespace, it raises ValueError as before, preserving existing error semantics.
• Changelog: Added a bullet point under ## Unreleased: mitmproxy next.

Checklist

• I have updated tests where applicable (test/mitmproxy/contentviews/test__view_json.py).
• I have added an entry to the CHANGELOG.
• Tests pass locally (uv run pytest test/mitmproxy/contentviews/test__view_json.py - 100% coverage on _view_json.py).
• Linting and typing pass cleanly (uv run ruff check . and uv run mypy).

Thanks for reviewing, and thank you for maintaining mitmproxy!