Each chunk gets converted to bytes, we find and remove and CR and CRLF's, and then they decode back to a string. replacements for CR and CRLF.
In an effort to change as little as possible, I avoided removing decode_unicode=True which would remove the requirement of converting the chunks to bytes. Also there might not be a need to do the bytes conversion.
Description of change
Each chunk
gets converted to bytes, we find and remove and CR and CRLF's, and then they decode back to a string.replacements for CR and CRLF.In an effort to change as little as possible, I avoided removing
decode_unicode=Truewhich would remove the requirement of converting the chunks to bytes. Also there might not be a need to do the bytes conversion.Manual QA steps
Risks
Rollback steps