Changes in version 1.1.0, released on May 2, 2019

   --- NEW PARSER FEATURES ---
 * Handle <funcparams> inside <paramdef> inside <funcprototype>.
 * Ignore <jobtitle> and <orgdiv> elements.
   --- NEW FORMATTER FEATURES --
 * Write a comment "automatically generated" at the beginning.
 * Provide an -s command line option to override the section in .Dt.
 * Large numbers of improvements to horizontal and vertical whitespace.
 * End of sentence detection: new sentence, new line.
 * Drop trailing parentheses in <function> elements.
 * Print commas between Copyright years.

Changes in version 1.0.2, released on April 29, 2019

   --- NEW PARSER FEATURES ---
 * Handle <abstract>: better formatting of many non-<refentry> documents.
 * Handle <reference> like <section>.
 * Explicitly ignore <keycombo>.
   --- STRUCTURAL IMPROVEMENTS ---
 * Start a new module reorg.c, to edit and move around nodes.
 * Provide a new function pnode_alloc_text() to generate a text node.
 * Introduce an xmalloc.c module to automatically exit when out of memory.

Changes in version 1.0.1, released on April 24, 2019

   --- NEW PARSER FEATURES ---
 * Handle file inclusion via "<!ENTITY %".
 * Translate decimal character entities to Unicode escapes.
 * Handle <tag> like <sgmltag> and <markup>.
   --- RELIABILITY BUGFIX ---
 * Do not crash when an ignored element contains an attribute.
   --- FORMATTING IMPROVEMENTS ---
 * Correctly translate input to output line breaks in no-fill elements.
 * Improve spacing of text nodes inside macro arguments.
 * Better handle <void> as a child of <funcprototype>.
 * Better handle <info> as a child of <refentry>.
 * Correct delimiter spacing between <term> elements.
 * Insert a paragraph break after displays and after non-compact lists.
   --- THANKS TO ---
 * Stephen Gregoratto for a small patch and two bug reports.
 * Jan Stary for a bug report.

Changes in version 1.0.0, released on April 17, 2019

 * Stop using expat because it cannot handle invalid input gracefully.
 * Drop input validation functionality, make it a best-effort formatter.
 * Implement file inclusion via <xi:include> and via <!ENTITY>.
 * Implement -T lint and -T tree dump output modes.
 * Handle large numbers of additional elements.
 * Dynamic assignment of header levels.
 * Overhaul whitespace handling on the input and output sides.
 * Overhaul paragraph handling.
 * Initial MathML support.
 * Switch to mandoc(1)-style EXIT STATUS and DIAGNOSTICS.
 * Large numbers of formatting and structural improvements.
   --- THANKS TO ---
 * Stephen Gregoratto for four patches and several bug reports.
 * Anthony Bentley and Svyatoslav Mishyn for minor patches.

Changes in version 0.0.9, released on April 30, 2014

 * Add the -W command line option.
 * Handle trailing delimiters after macros.
 * Handle <fieldsynopsis>, <modifier>, <type>.

Changes in version 0.0.8, released on April 30, 2014

 * Handle <contstant>, <copyright>, <holder>, <informaltable>,
   <trademark>, <year>.
 * Ignore <xi:include>.

Changes in version 0.0.7, released on April 2, 2014

 * Handle <application>, <caution>, <note>, <quote>, <sgmltag>,
   <tip>, <warning>.

Changes in version 0.0.6, released on April 2, 2014

 * Initial support for tables.
 * Handle <envar>, <filename>, <group>, <itemizedlist>, <link>,
   <literal>, <orderedlist>, <prompt>, <sbr>, <screen>, <userinput>,
   <variablelist>, <varname>.

Changes in version 0.0.4, released on March 29, 2014

 * Support for XML element attributes.
 * Handle <structname>.
 * Commas between names in the NAME section.

Start of development in CVS and first public release:
Version 0.0.2, released on March 28, 2014.