I made great progress on the roadmap. I completed the first chapter of the tutorial and drafted the rest of it, finished and internationalized the documentation (moving from YAML files to Sphinx), added many new food templates and support for more food databases, reworked the guidelines so they match the reference data (including configurable authorities, energy-rate limits, pregnancy support, and extra vitamin C for smokers), and implemented a long list of new features such as menu tags, weight tracking, menu export and import, unit conversion, meal notes, and menu durations.

It was an experiment: I used Godot to try to create an action RPG based on The Battle for Wesnoth. I have already decided to abandon the project.

In addition to reviewing many pull requests, I opened 45 pull requests, of which 27 have already been merged. In CI and tooling, I set up CodSpeed, switched to installing dependencies with uv, and reduced the test job matrix while maintaining coverage.

I improved the test coverage of scrapy.utils, scrapy.commands, scrapy.extensions, scrapy.pipelines, scrapy.cmdline, and the spiders.

I added scrapy.utils.asyncio.sleep(), support for CONCURRENT_REQUESTS = 0 for unlimited concurrency, RobotParser.crawl_delay() and a robots_parsed signal, exporting item fields in declaration order, AWS_MAX_POOL_CONNECTIONS, and treating broken cache records as cache misses. I also fixed a port issue with cached DNS.

I documented urlparse_cached, added documentation sections for callbacks and errbacks, added a stats reference, and converted several documentation pages to autodoc (contracts, trackref, stats, exceptions, and spiders).

I also proposed deprecating the download_delay spider attribute, limiting response headers size like browsers, supporting HTTPS proxies with Twisted and aioboto3, supporting feed modes, and browser-like bad header handling for the default download handler. I kept working on the new throttling API and solved conflicts in my older pull requests.

I reviewed pull requests, including the release notes for version 1.5.0.

I proposed only detecting JSON objects and arrays as JSON, caching CSS-to-XPath translations per instance, and an interactive XPath tutorial. I also reviewed pull requests.

I fixed the JSON Schema pattern flags and anchoring, proposed using $defs/$ref for reused and recursive item classes, and reviewed a pull request.

I added a minimal CI environment and reviewed pull requests.

I reviewed a series of URL-parsing pull requests and solved conflicts in my can_ada pull request.

I added support for running without a reactor, which makes it possible to disable the Twisted reactor, and proposed supporting discarding requests from user code, making stats linkable, and a ZyteApiParams page input.