Highlights:

• Better Windows support
• Python 3.7 compatibility
• Big documentation improvements, including a switch from .extract_first() + .extract() API to .get() + .getall() API
• Feed exports, FilePipeline and MediaPipeline improvements
• Better extensibility: item_error and request_reached_downloader signals; from_crawler support for feed exporters, feed storages and dupefilters.
• scrapy.contracts fixes and new features
• Telnet console security improvements, first released as a backport in Scrapy 1.5.2 (2019-01-22)
• Clean-up of the deprecated code
• Various bug fixes, small new features and usability improvements across the codebase.

Full changelog is in the docs.