Update to 2.2.6. From the changelog:

- make a test program to examine pandoc header elements
- if __mkd_trim_line trims the entire line, manually set the string to \0
- add in a handful of test to verify that empty pandoc headers are not
  auto-populated with %
- messed up the definition for missing attributes; should be /**/
  instead of '', because AC_DEFINE will convert an empty string into a 1
- move flags.o down to the last line in OBJS= to work around a plan9
  makefile bug
- have peek() return an int instead of an unsigned int (netbsd 8; ctype
  functions explode with unsigned int arguments?)
- fix all places where I pass flags to be DWORDs
- include config.h before any other headers
- make all instances of 'main()' be integer functions (sigh)
- clean up all references to flags; define them EVERYWHERE as mkd_flag_t
- shuffle around some of the definitions to make everything happily(?)
  compile with all instances of flags being mkd_flag_t
- when checking that the C compiler works, compile the test program as
  two modules with a function declared with differing pointer types
  (char ptr vs void ptr) so if the C compiler generates type signatures
  and doesn't think a void pointer is the universal solvent it will fail
  right there
- set the Plan9 CFLAGS to the two extensions (instead of having them be
  part of the CC macro) and wipe out any CFLAGS (like -T, which tells
  the Plan9 posix C compiler to be stupid) that already exists
- tweak the Makefile macros a bit so people can wedge in their own
  compilers (BUILD) and linkers (LINK) during a build
- add a list item for testing the build
- patch the cmakefile so that has an option to only generate a library
- push the fencedcode block handler up to the toplevel scanner so it
  will handle blank lines inside the fence
- add a fenced code block with embedded html
- don't give a code block class if it's just an empty string
- add a test case for a piece of fenced code with trailing spaces after
  the opening ~~~
- deal with cached text before handling fenced code blocks
- yank the repeated inline uncaches out and stuff them into a local function
- add two more tests (unfinished fenced code block, paragraph then
  fenced code block) to codeblock.t
- don't publish mkd_string_to_anchor
- add an anchor format callback, plus add the -x option to the markdown
  program to squash toc anchors to (close to) github compatability
- add some dle sanity checking into onlyifset()
- when dumping the parse tree, say which header is being dumped --
  h1..h6 vs just header
- tweak configure to (a) use the modern standard uintXX_t types for
  DWORD, WORD, BYTE if possible and (b) if a program is defined via an
  environment variable (like CC) try to use that variable instead of
  doing a path search in AC_PROG
- tweak how scalar headers are include in .in files
- support external code formatters
- add an external code block formatter to the markdown program
- mkd2html wasn't passing pgm to hoptusage()
- mkd2html wasn't properly handling two arguments for input & output
  files. While I was at it, add the -G option to use gfm_in() to read
  the input file in (and preserve newlines)


(schmonz)