Jonathan Hedley

Details at jsoup.org/news/release-1… This release marks 15 years of jsoup development.

I'm happy to announce the release of #jsoup 1.22.1! This adds support for the re2j regular expression engine for regex-based CSS selectors, a configurable maximum parser depth, and brings a bunch of bug fixes and improvements.

Happy to announce that #jsoup 1.21.2 is out! Adds custom SSLContext support in HTTP/2 connections, brings DOM/fragment parsing perf gains, and fixes some edge cases in parsing, traversal, cloning, and concurrent reads. jsoup.org/news/release-1…

Happy to announce that #jsoup v1.21.1 is out now! Lots of improvements, particularly the ability to directly select nodes (like text, data) with the CSS selectors.

This release also delivers performance optimizations, new API enhancements such as flexible tag definitions via TagSet, concise CSS selectors, and parser thread-safety improvements. Big thanks to everyone who helped out. jsoup.org/news/release-1…

Very happy to announce that I've just released jsoup 1.20.1! Lots of improvements and bug fixes -- improved HTML parse rules to align with modern browsers, improved XML namespace handling, and a redesigned HTML pretty-printer for better consistency and customizability.

See the full list of changes: jsoup.org/news/release-1…

Good news everybody! I just released #jsoup v1.19.1. It adds http/2 request support, and has a bunch of other improvements and bug fixes.

@emergetools @ouraring Their response just reads like meaningless AI slop. "Write a nice response to this question: ..."

Customer: “Why is this app so big?” Company: “Very cool features & lots of them + you have no way to check what I’m saying” Not today corporate. Not today. 🧵 Why @ouraring is ~450 MB

@gamesover @grok Not snark; how did you know it was true? I've had a bunch of times where these sound real but have no underlying basis. Great when they link to a source, but I find a regular search already has that.

I've been using @grok to eliminate some of the pain points regarding tech dark patterns. ex, I wasn't sure if something was included with a service's base tier or required a paid plan. I got an immediate answer. This was so much faster than searching through the site's marketing.

The implementation uses Java's multi-release JAR feature to make requests via the HttpClient impl if it's available, or will fallback to the current HttpURLConnection. This also gives a path to http/3 support when that PEP lands in Java. github.com/jhy/jsoup/pull…

It's a drop-in update with no changes required for existing Jsoup.connect() code, other than setting a system property (jsoup.useHttpClient) to enable.

The next version of #jsoup will (finally!) support making http/2 requests, if you're running on Java 11+. It still works down to Java 8 if you need that.

Details (and graphs) at github.com/jhy/jsoup/rele…

Happy to announce the release of jsoup 1.18.2! Faster parsing with less GCs, and a bunch of bug fixes.

As a result, heap allocations (bytes/op) are down by -6% to -89%, and throughput has improved by -2% to +143% (with the biggest gains for smaller inputs). These improvements will be in the next release of jsoup, 1.18.2 (coming soon!)

I've been working on improving parse throughput and reducing memory allocations in jsoup (Java HTML parser) by recycling char[] and byte[] buffers between invocations—avoiding unnecessary heap allocations and garbage collection. Details: github.com/jhy/jsoup/pull…

I've just released jsoup 1.18.1! Lots of improvements - a StreamParser that acts like a hybrid DOM + SAX parser; URL download progress callbacks, and lots of other improvements. jsoup.org/news/release-1…

If you're interested in this, please take a look at the implementation, and try it out by installing a snapshot. It would be great to incorporate any initial feedback / bug-fixes prior to releasing it in the next version of #jsoup.

I've been working on a new feature for jsoup that I think is pretty cool: the new StreamParser lets you parse a document lazily with selectNext(query). Elements are parsed from the backing input stream on demand, and will include all their children. github.com/jhy/jsoup/pull…