Release 3.2.6

• Fix handling Uri with an anchor #209
• New option DefaultStyles.NumberedHeadingStyle to support an alternate heading style #210

Release 3.2.5

• Fix a crash with the new whitespace handling introduced in 3.2.3 #191
• Fix crash when the html contains 2 images with identical source path #193
• Support margin auto for table alignment #194
• Fix handling whitespace between runs #195
• Whitelist more mime-types as specified by the IANA standard #196
• Support EMF file #196
• Correct handling of figcaption (allow nested phrasings) #197
• Numbering list now supports type attribute <ol type="1|a|A|i|I"> #198
• Always restart nested numbering list #198
• Fix table borders being removed even when the specified word table style has borders #199
• Defensive code when download image stream is truncated #201
• Table inside list is constrained to not exceed page margin #202
• Table now supports width:auto for auto-fit content #202

Release 3.2.4

• Fix a crash with the new whitespace handling introduced in 3.2.3 #191
• Table inside list must be aligned with the list item #192

Release 3.2.3

• Improve support of table alignment #187
• Fix a crash if a span is empty #190
• Heading with only digits should not be considered as a numbering #189
• Fix whitespaces inserted between spans #179 and #185
• Support percentage size (typically width:100%) for img node #188

Release 3.2.2

• Supports a feature to disable heading numbering #175
• Support center image with margin auto #171
• Support deprecated align attribute for block #171
• Fix parsing the style attribute with a key that has no value
• Improve parsing of style attribute to avoid an extra call to HtmlDecode
• Extend support of nested list for non-W3C compliant html #173
• Change way to apply table 100% width
• Allow to apply percentage widths cells

Release 3.2.1

• Fix indentation of numbering list #166
• Bordered container must render its content with one bordered frame #168
• Fix serialisation of the "Harvard" style for lower-roman list
• Fix ParseHeader/Footer where input with multiple paragraphs output only the latest
• Ensure to apply default style for paragraphs, to avoid a paragraph between 2 list is mis-guessed

Release 3.2.0

Some API methods have been flagged as obsolete with a clear message of what to use instead. Those obsoletes are designed to highlight the expected output of the API and remove disambiguation between all the "Parse" methods. Documentation have been refreshed as well. No breaking changes as it maintain existing behaviour.

ParseHtml (string html) -> ParseBody
Parse(string html) -> ParseAsync

This release bring support for parsing into Header and Footer which was requested by open source project Cervantes (thanks for your support), a collaborative platform designed specifically for pentesters and red teams.

• Add new public API to allow parsing into Header and Footer #162.
• Add support for SVG format (either from img src or the SVG node tag)
• Automatically create the _top bookmark if needed
• Fix a crash when a hyperlink contains both img and figcation
• Fix a crash when li is empty #161

Release 3.1.1

• Fix respecting layout with div/p ending with line break #158
• Prevent crash when header/footer is incomplete and parsing image #159
• Fix combining 2 runs separated by a break, 2nd line should not be prefixed by a space

Release 3.1.0

• Fix table Cell borders are wrongly applied on the run #156
• Correctly handle RTL layout for text, list, table and document scope #86 #66
• Support property line-height #52
• Fallback to background style attribute as many users use this simplified attribute version
• In HtmlDomExpression.CreateFromHtmlNode, use the correct casting to IElement rather than IHtmlElement, to prevent crash if svg node is encountered

Release 3.0.1

Bug fixes:

• Ensure to count existing images from header and footer too #113
• Preserve line break pre for OSX/Windows
• Prevent a crash when the provided style is missing its type
• Defensive code to avoid 2 rowSpan+colSpan with a cell in between to crash #59