Improvements

• Optimized the throughput and memory use throughout the input read and parse flows, with heap allocations and GC down between -6% and -89%, and throughput improved up to +143% for small inputs. Most inputs sizes will see throughput increases of ~ 20%. These performance improvements come through recycling the backing byte[] and char[] arrays used to read and parse the input. 2186
• Speed optimized html() and Entities.escape() when the input contains UTF characters in a supplementary plane, by around 49%. 2183
• The form associated elements returned by FormElement.elements() now reflect changes made to the DOM, subsequently to the original parse. 2140
• In the TreeBuilder, the onNodeInserted() and onNodeClosed() events are now also fired for the outermost / root Document node. This enables source position tracking on the Document node (which was previously unset). And it also enables the node traversor to see the outer Document node. 2182
• Selected Elements can now be position swapped inline using Elements#set(). 2212

Bug Fixes

• Element.cssSelector() would fail if the element's class contained a * character. 2169
• When tracking source ranges, a text node following an invalid self-closing element may be left untracked. 2175
• When a document has no doctype, or a doctype not named html, it should be parsed in Quirks Mode. 2197
• With a selector like div:has(span + a), the has() component was not working correctly, as the inner combining query caused the evaluator to match those against the outer's siblings, not children. 2187
• A selector query that included multiple :has() components in a nested :has() might incorrectly execute. 2131
• When cookie names in a response are duplicated, the simple view of cookies available via Connection.Response#cookies() will provide the last one set. Generally it is better to use the Jsoup.newSession method to maintain a cookie jar, as that applies appropriate path selection on cookies when making requests. 1831
• When parsing named HTML entities, base entities should resolve if they are a prefix of the input token (and not in an attribute). 2207
• Fixed incorrect tracking of source ranges for attributes merged from late-occurring elements that were implicitly created (html or body). 2204
• Follow the current HTML specification in the tokenizer to allow < as part of a tag name, instead of emitting it as a character node. 2230
• Similarly, allow a < as the start of an attribute name, vs creating a new element. The previous behavior was intended to parse closer to what we anticipated the author's intent to be, but that does not align to the spec or to how browsers behave. 1483

https://jsoup.org/news/release-1.18.1

Improvements

• Stream Parser: A StreamParser provides a progressive parse of its input. As each Element is completed, it is
  emitted via a Stream or Iterator interface. Elements returned will be complete with all their children, and an
  (empty) next sibling, if applicable. Elements (or their children) may be removed from the DOM during the parse,
  for e.g. to conserve memory, providing a mechanism to parse an input document that would otherwise be too large to fit
  into memory, yet still providing a DOM interface to the document and its elements. Additionally, the parser provides
  a selectFirst(String query) / selectNext(String query), which will run the parser until a hit is found, at which
  point the parse is suspended. It can be resumed via another select() call, or via the stream() or iterator()
  methods. 2096
• Download Progress: added a Response Progress event interface, which reports progress and URLs are downloaded (and
  parsed). Supported on both a session and a single connection
  level. 2164, 656
• Added Path accepting parse methods: Jsoup.parse(Path), Jsoup.parse(path, charsetName, baseUri, parser),
  etc. 2055
• Updated the button tag configuration to include a space between multiple button elements in the Element.text()
  method. 2105
• Added support for the ns|* all elements in namespace Selector. 1811
• When normalising attribute names during serialization, invalid characters are now replaced with _, vs being
  stripped. This should make the process clearer, and generally prevent an invalid attribute name being coerced
  unexpectedly. 2143

Changes

• Removed previously deprecated internal classes and methods. 2094
• Build change: the built jar's OSGi manifest no longer imports itself. 2158

Bug Fixes

• When tracking source positions, if the first node was a TextNode, its position was incorrectly set
  to -1. 2106
• When connecting (or redirecting) to URLs with characters such as {, } in the path, a Malformed URL exception would
  be thrown (if in development), or the URL might otherwise not be escaped correctly (if in
  production). The URL encoding process has been improved to handle these characters
  correctly. 2142
• When using W3CDom with a custom output Document, a Null Pointer Exception would be
  thrown. 2114
• The :has() selector did not match correctly when using sibling combinators (like
  e.g.: h1:has(+h2)). 2137
• The :empty selector incorrectly matched elements that started with a blank text node and were followed by
  non-empty nodes, due to an incorrect short-circuit. 2130
• Element.cssSelector() would fail with "Did not find balanced marker" when building a selector for elements that had
  a ( or [ in their class names. And selectors with those characters escaped would not match as
  expected. 2146
• Updated Entities.escape(string) to make the escaped text suitable for both text nodes and attributes (previously was
  only for text nodes). This does not impact the output of Element.html() which correctly applies a minimal escape
  depending on if the use will be for text data or in a quoted
  attribute. 1278
• Fuzz: a Stack Overflow exception could occur when resolving a crafted <base href> URL, in the normalizing regex.
  2165

Improvements

• Attribute object accessors: Added Element.attribute(String) and Attributes.attribute(String) to more simply
  obtain an Attribute object. 2069
• Attribute source tracking: If source tracking is on, and an Attribute's key is changed (
  via Attribute.setKey(String)), the source range is now still tracked
  in Attribute.sourceRange(). 2070
• Wildcard attribute selector: Added support for the [*] element with any attribute selector. And also restored
  support for selecting by an empty attribute name prefix ([^]). 2079

Bug Fixes

• Mixed-cased source position: When tracking the source position of attributes, if the source attribute name was
  mix-cased but the parser was lower-case normalizing attribute names, the source position for that attribute was not
  tracked correctly. 2067
• Source position NPE: When tracking the source position of a body fragment parse, a null pointer
  exception was thrown. 2068
• Multi-point emoji entity: A multi-point encoded emoji entity may be incorrectly decoded to the replacement
  character. 2074
• Selector sub-expressions: (Regression) in a selector like parent [attr=va], other, the , OR was binding
  to [attr=va] instead of parent [attr=va], causing incorrect selections. The fix includes a EvaluatorDebug class
  that generates a sexpr to represent the query, allowing simpler and more thorough query parse
  tests. 2073
• XML CData output: When generating XML-syntax output from parsed HTML, script nodes containing (pseudo) CData
  sections would have an extraneous CData section added, causing script execution errors. Now, the data content is
  emitted in a HTML/XML/XHTML polyglot format, if the data is not already within a CData
  section. 2078
• Thread safety: The :has evaluator held a non-thread-safe Iterator, and so if an Evaluator object was
  shared across multiple concurrent threads, a NoSuchElement exception may be thrown, and the selected results may be
  incorrect. Now, the iterator object is a thread-local. 2088

jsoup 1.17.1 is out now with support for request-level authentication, attribute name & value source ranges, stream() iterable support, and a bunch of other improvements and bug fixes.

Many thanks to everyone who contributed to this release!

Improvements

• Request-Level Authentication: Added support for request-level authentication in Jsoup.connect(), enabling authentication to proxies and servers. More.

• Elements DOM Mutators: In the Elements list, added direct support for Elements#set(int, Element), Elements#remove(int), Elements#remove(Object), Elements#clear(), Elements#removeAll(), Elements#retainAll(), Elements#removeIf(), Elements#replaceAll(). These methods update the original DOM, as well as the Elements list. More.

• Stream Interface: Introduced the NodeIterator class for efficient node tree traversal using the Iterator interface. Added Stream Element#stream() and Node#nodeStream() methods for fluent composable stream pipelines of node traversals. More.

• XML OutputSettings: Automatically sets the xhtml EscapeMode as default when changing the OutputSettings syntax to XML.

• is() Selector: Added the :is(selector list) pseudo-selector to find elements that match any selectors in the selector list. This enhances readability for large ORed selectors. More.

• JPMS Module Support: Repackaged the library with native JPMS module support. More.

• Source Position Fidelity: Improved fidelity of source positions when tracking is enabled. Implicitly created or closed elements are now trackable via Range.isImplicit(). More.

• Attribute Source Positions: Enabled source position for attribute names and values when source tracking is on. Attribute#sourceRange() provides the ranges. More.

• Virtual Threads: Enhanced performance under Java 21+ Virtual Threads by replacing the internal ConstrainableInputStream with ControllableInputStream. More.

• XML Mimetype Support: Extended XML mimetype support in Jsoup.connect() to include any XML mimetype. More.

Bug Fixes

• XML Data Nodes: Fixed a bug where HTML elements parsed as data nodes were not correctly emitted as CDATA nodes when outputting with XML syntax. More.

• Immediate Parent Selector: Corrected a bug where the Immediate Parent selector > could match elements above the root context element. More.

• Sub-Query Parsing: Resolved a bug where combinators following the , Or combinator in a sub-query were incorrectly skipped. More.

• Empty Doctype: Fixed a bug in W3CDom where the conversion would fail if the jsoup input document contained an empty doctype. The doctype is now discarded, and the conversion continues.

• SVG Elements Cleaning: Fixed incorrect nesting when cleaning a document containing SVG elements or other foreign elements with preserved-case names. More.

• Unknown Self-Closing Tags: Preserved the output style of unknown self-closing tags from the input when cleaning a document. More.

Build Improvements

• Local Test Proxy: Added a local test proxy implementation for proxy integration tests. More.

• HTTPS Request Tests: Added tests for HTTPS request support using a local self-signed certificate. Includes proxy tests. More.

Changes

• Response BodyStream: The InputStream returned in Connection.Response.bodyStream() is now a plain BufferedInputStream. More.

Improvements

• Optimized the performance of complex CSS selectors, by adding a cost-based query planner. Evaluators are sorted by their relative execution cost, and executed in order of lower to higher cost. This speeds the matching process by ensuring that simpler evaluations (such as a tag name match) are conducted prior to more complex evaluations (such as an attribute regex, or a deep child scan with a :has).

• Added support for <svg> and <math> tags (and their children). This includes tag namespaces and case preservation on applicable tags and attributes. #2008

• When converting jsoup Documents to W3C Documents in W3CDom, HTML documents will be placed in the http://www.w3.org/1999/xhtml namespace by default, per the HTML5 spec. This can be controlled by setting W3CDom#namespaceAware(boolean false). #1848

• Speed optimized the Structural Evaluators by memoizing previous evaluations. Particularly the ~ (any preceding sibling) and :nth-of-type selectors are improved. #1956

• Tweaked the performance of the Element nextElementSibling, previousElementSibling, firstElementSibling, lastElementSibling, firstElementChild, and `lastElementChild. They now inplace filter/skip in the child-node list, vs having to allocate and scan a complete Element filtered list.

• Optimized internal methods that previously called Element.children() to use filter/skip child-node list accessors instead, reducing new Element List allocations.

• Tweaked the performance of parsing :pseudo selectors.

• When using the :empty pseudo-selector, blank textnodes are now considered empty. Previously, an element containing any whitespace was not considered empty. #1976

• In forms, <input type="image"> should be excluded from Element.formData() (and hence from form submissions). #2010

• In Safelist, made isSafeTag() and isSafeAttribute() public methods, for extensibility. #1780

Bug Fixes

• Bugfix: form elements and empty elements (such as img) did not have their attributes de-duplicated. #1950

• If Document.OutputSettings was cloned from a clone, an NPE would be thrown when used. #1964

• In Jsoup.connect(String url), URL paths containing a %2B were incorrectly recoded to a '+', or a '+' was recoded to a ' '. Fixed by reverting to the previous behavior of not encoding supplied paths, other than normalizing to ASCII. #1952

• In Jsoup.connect(String url), strings containing supplemental characters (e.g. emoji) were not URL escaped correctly.

• In Jsoup.connect(String url), the ConstrainableInputStream would clear Thread interrupts when reading the body. This precluded callers from spawning a thread, running a number of requests for a length of time, then joining that thread after interrupting it. #1991

• When tracking HTML source positions, the closing tags for H1...H6 elements were not tracked correctly. #1987

• In Jsoup.connect(), a DELETE method request did not support a request body. #1972

• When calling Element.cssSelector() on an extremely deeply nested element, a StackOverflowError could occur. Further, a StackOverflowError may occur when running the query. #2001

• Appending a node back to its original Element after empty() would throw an Index out of bounds exception. Also, now the child nodes that were removed have their parent node cleared, fully detaching them from the original parent. #2013

• In Connection when adding headers, the value may have been assumed to be an incorrectly decoded ISO_8859_1 string, and re-encoded as UTF-8. The value is now left as-is.

Changes

• Removed previously deprecated methods Document.normalise(), Element.forEach(org.jsoup.helper.Consumer<>), Node.forEach(org.jsoup.helper.Consumer<>), and the org.jsoup.helper.Consumer interface; the latter being a previously required compatibility shim prior to Android's de-sugaring support.

• The previous compatibility shim org.jsoup.UncheckedIOException is deprecated in favor of the now supported java.io.UncheckedIOException. If you are catching the former, modify your code to catch the latter instead. #1989

• Blocked noscript tags from being added to Safelists, due to incompatibilities between parsers with and without script-mode enabled.

jsoup Java HTML Parser release 1.16.1

Improvements

• In Jsoup.connect(String url), natively support URLs with Unicode characters in the path or query string, without having to be escaped by the caller. #1914

• Calling Node.remove() on a node with no parent is now a no-op, vs a validation error. #1898

Bug Fixes

• Aligned the HTML Tree Builder processing steps for AfterBody and AfterAfterBody to the updated WHATWG standard, to not pop the stack to close <body> or <html> elements. This prevents an errant </html> closing the preceding structure. Also added appropriate error message outputs in this case. #1851

• Corrected support for ruby elements (<ruby>, <rp>, <rt>, and <rtc>) to current spec. #1294

• When using Node.before(Node) or Node.after(Node), if the incoming node was a sibling of the context node, the incoming node may be inserted into the wrong relative location. #1898

• In Jsoup.connect(String url), if the input URL had components that were already % escaped, they would be escaped again, causing errors when fetched. #1902

• When tracking input source positions, text in tables that was fostered had invalid positions. #1927

• If the Document.OutputSettings class was initialized, and then Entities.escape(String) called, an NPE may be thrown due to a class loading circular dependency. #1910

• When pretty-printing, the first inline Element or Comment in a block would not be wrap-indented if it were preceded by a blank text node. #1906

• When pretty-printing a <pre> containing block tags, those tags were incorrectly indented. #1891

• When pretty-printing nested inlineable blocks (such as a <p> in a <td>), the inner element should be indented. #1926

• <br> tags should be wrap-indented when in block tags (and not when in inline tags). #1911

• The contents of a sufficiently large <textarea> with un-escaped HTML closing tags may be incorrectly parsed to an empty node. #1929

jsoup Java HTML Parser release 1.15.4

jsoup 1.15.4 is out now, and includes a bunch of improvements, particularly when pretty-printing HTML, and bug fixes.

jsoup is a Java library for working with real-world HTML. It provides a very convenient API for extracting and manipulating data, using the best of HTML5 DOM methods and CSS selectors.

Download jsoup now.

Improvements

• Added the ability to escape CSS selectors (tags, IDs, classes) to match elements that don't follow regular CSS syntax. For example, to match by classname <p class="one.two">, use document.select("p.one\\.two"); #838

• When pretty-printing, wrap text that follows a <br> tag. #1858

• When pretty-printing, normalize newlines that follow self-closing tags in custom tags. #1852

• When pretty-printing, collapse non-significant whitespace between a block and an inline tag. #1802

• In Element.forEach() and Node.forEachNode(), use java.util.function.Consumer instead of the previous Android compatibility shim org.jsoup.helper.Consumer. Subsequently, the latter has been deprecated. #1870

• Added a new method Document.forms(), to conveniently retrieve a List<FormElement> containing the <form> elements in a document.

• Added a new method Document.expectForm(), to find the first matching FormElement, or blow up trying.

Bug Fixes

• URLs containing characters such as and <code> were not escaped correctly, and would throw a MalformedURLException when fetched. #1873

• Element.cssSelector() would create invalid selectors for elements where the tag name, ID, or classnames needed to be escaped (e.g. if a class name contained a : or .). #1742

• Element.text() should have a space between a block and an inline element. #1877

• If a Node or an Element was replaced with itself, that node would incorrectly be orphaned. #1843

• Form data on a previous request was copied to a new request in newRequest(), resulting in an accumulation of form data when executing multi-step form submissions, or data sent to later requests incorrectly. Now, newRequest() only copies session related settings (cookies, proxy settings, user-agent, etc) but not the request data nor the body. #1778

• Fixed an issue in Safelist.removeAttributes() which could throw a ConcurrentModificationException when using the :all pseudo-attribute.

• Given extremely deeply nested HTML, a number of methods in Element could throw a StackOverflowError due to excessive recursion. Namely: #data(), #hasText(), #parents(), and #wrap(html). #1864

Changes

• Deprecated the unused Document.normalise() method. Normalization occurs during the HTML tree construction, and no longer as a distinct phase.

My sincere thanks to everyone who contributed patches, suggestions, and bug reports. If you have any suggestions for the next release, I would love to hear them; please get in touch with me directly.

You can also follow me (@[email protected]) on Mastodon / Fediverse to receive occasional notes about jsoup releases.

jsoup 1.15.3 is out now, and includes a security fix for potential XSS attacks, along with other bug fixes and improvements, including more descriptive validation error messages.

Details:

• Security advisory
• Release notes
• Download

jsoup 1.15.2 is out now with a bunch of improvements and bug fixes.

jsoup 1.15.1 is out now with a bunch of improvements and bug fixes.