ikreymer
commented
2 months ago - additional fixes to make surt conversion and indexing closer to original python implementation in cdxj-indexer and surt.
- sort query args by splitting on '&', instead of parsing via URLSearchParams to preserve empty query args
- SURT contains non urlencoded query args when possible, including POST request body
- warc record: parse URL to ensure trailing slash included in http/https URLs
- fixes #70