March 19th, 2013
· Trailing comments after table rows are now preserved (example: matherr(3)).
· Support for some previously missing groff extension glyphs was added.
· Handling of .Bd/.Be in mdoc was improved.
August 1st, 2012
· Handle foojzs pages better. Interpret some cases of .rj.
· Recognize "Feature Test" as a function synopsis ender.
· Handle m, r and d troff conditionals.
· Process .ti with positive indent into around the following line.
· Support all mdoc special-character strings.
· Improved recognition of program listings.
· Brown-paper-bag bug in processing of mdoc
June 26th, 2012
· Fix a bug in command-synopsis parsing pointed out by Tom Browder Lifts 97% of 11029 pages in a full Ubuntu Precise Pangolin release.
August 23rd, 2011
· Improvement for lynxprep handling by Jon Vyse.
November 25th, 2010
· Clean up glitches revealed by pychecker.
· Fix buggy interpretation of ms .AI macro.
· Mab TBL "box" attribute to Docbook frame="border".
July 22nd, 2010
· eqn markup is now handled if the eqn -TMathml switch produces results. Bell Labs or Berkeley meaning of .P1 is dispatched to depending on whether .P2 is present. Added -w option for strict portability checking. Fedora bug 220736 fixed. All troff glyphs are now mapped (added bracket-pile characters, yogh, hooked-o, and underdot). You are now warned of sequences that look like glyphs but can't be mapped. Table handling for mdoc pages has been much improved. Tests for requests that can't be turned into structure are stricter. Appropriate cases of \o are now translated into Latin-1 and Latin-2 letters with accents. Inline ad-hoc tables made with .ta and literal tabs are now lifted. Groff extended escapes $* and $@ are now handled. Speed optimizations so it's about 30% faster, and a profiling switch on manlifter. Rudimentary DocBook V5 translation, but inclusions and character entities are iffy and untested. Lifts 94% of 11863 pages in a full Ubuntu Lucid Lynx install.