======================================================================
Version 1.0.1, by Aleksey Cheusov, Thu, 14 Aug 2014 01:44:39 +0300

  Unbuffered (broken) stdout seen on Darwin was fixes.
  Who knows, this may happen on other systems too.

======================================================================
Version 1.0.0, by Aleksey Cheusov, Sun, 13 Jul 2014 19:10:48 +0300

  mk-configure>=0.27.0 is required for build.

  Option -X was introduced to paexec(1) for ignoring
  calculator's stdout.

  PAEXEC_ENV environment variable sets a list of variables passed to
  the calculator.

  PAEXEC_TRANSPORT environment variable sets the transport unless
  option -t was applied.

  PAEXEC_NODES environment variable sets the nodes unless
  option -n was applied.

  pareorder(1) is a synonym for paexec_reorder(1)

======================================================================
Version 0.19.1, by Aleksey Cheusov, Mon, 26 Aug 2013 19:05:53 +0300

  paexec: Fix for segfault seen on Linux

Version 0.19.0, by Aleksey Cheusov, Sun, 25 Aug 2013 17:21:03 +0300

  This release of paexec was successfully tested on the following
  platforms: NetBSD-6.1/amd64, OpenBSD-5.3/i386, FreeBSD-8.3/i386,
  Solaris-10/sparc, Solaris-11/amd64 and diverse Linux/{i386,amd64}.

  paexec is now selfcontained, libmaa is not needed anymore.

  Presentation paexec.pdf was added to presentation/. I hope it will
  help easier understand how "paexec" works. It is installed to
  ${DOCDIR}.

  paexec:
    - POSIX-2008 getline(3) is used for reading lines
      instead of home-made function.
    - t '' is equivalent to "no transport", spaces are trimmed.
    - "-n +NNN" has higher priority than -t, i.e. if they both are
      specified, transport is ignored.
    - Fix for -W1.
    - Environment variable PAEXEC_EOT was introduced.
    - Option -y was added to paexec(1) and paexec_reorder(1).
    - Option -C was added to paexec(1).

  paexec_reorder:
    - Option -x was added to paexec_reorder(1).

  A number of fixes and improvements in regression tests.

======================================================================
Version 0.18.0, by Aleksey Cheusov, Thu,  7 Mar 2013 15:17:33 +0300

  paexec:
    - fixed: NULL dereference when the first line given on input is
      empty.  Thanks to Sergey Revyako for bug report!
    - fixed: entire command passed tp ssh-like transport should be
      shquoted. In particular this fixes -x that didn't work with -t.
      Thanks to Sergey Revyako for bug report!
    - 'paexec -g' accepts empty strings as tasks.
    - More regression tests were added

  paexec_reorder.1:
    - Mistype fix

======================================================================
Version 0.17.0, by Aleksey Cheusov, Sun,  9 Sep 2012 02:07:34 +0300

  paexec:
    - Option -x was added. With its help paexec can run one command
      per task.  If -g is also specified, command's exit status is
      analysed. Appropriate task and dependants are marked as "failed"
      if it is non-zero.
    - First character of -n argument must be alphanumeric, `+', `_',
      `:' or `/'.  Other symbols are reserved for future extentions.
    - With '-n :filename' paexec reads a list of nodes from the
      specified file.
    - With a help of new option '-m t=<eot>' end of task string
      may be specified, which is an empty line by default.
    - Option -md=<delim> was added that overrides the default
      delimiter (space character) between tasks in graph mode (-g).
    - Output line that contains failed dependants no longer ends with
      unnecessary space.
    - Long options were completely removed.

  paexec_reorder:
    - Fix. "paexec_reorder -g" now handles correctly failed tasks'
      output.  One extra line after "fatal" is expected.
    - Options -m was added. It does the same things as paexec's -m.

  More examples of use and regression tests.

  Documentation update, clean-ups and improvements.

  Regression tests:
    - Signals handling was fixed in.
    - LC_ALL is always set to C in regression tests, this fixes some
      problems in internationalized environment.

  mk-configure>=0.23.0 is required at build time

======================================================================
Version 0.16.1, by Aleksey Cheusov, Wed, 23 Mar 2011 00:14:15 +0200

  paexec.1 and paexec_reorder.1 are included to paexec-0.16.1.tgz.
  Just like in paexec<=0.15.1 pod2man is not needed
  for building paexec.

======================================================================
Version 0.16, by Aleksey Cheusov, Fri, 11 Mar 2011 11:58:31 +0200

  Project's structure has been reorganized. Now the top-level Makefile
  uses mkc.subprj.mk. This adds a lot of flexibility in building the
  project and development.
  See doc/INSTALL for updated installation instructions.

  New modes for reordering tasks were added: -W0 and -W2
  See the manual page for details about option -W.

  Long options are considered deprecated. They are still supported but
  will be removed in the future. Please use POSIX short options. At
  the moment use of them produces warning message on stderr.

  Fix the compilation bug on old versions of OpenBSD (at least <=3.8)
  and probably other systems where intptr_t is declared in inttypes.h
  but not in stdint.h

  Fix for 'mkcmake test' failure on Solaris and HP-UX because their
  /usr/bin/awk sucks. On these platforms now it is possible to run the
  following command.
    env AWK=/full/path/to/gawk mkcmake all test

======================================================================
Version 0.15.0, by Aleksey Cheusov, Sun, 10 Oct 2010 18:06:31 +0300

  After some thoughts I decided to switch from plain mk-files to
  mk-configure build system. It provides very cool features and makes
  development drammatically easier. It also makes Makefiles much
  cleaner and easier. Installation instructions has been updated in
  README file.

  PAEXEC reads the tasks from stdin and distributes them to all
  available hosts or CPU. If some tasks are easy and can be made
  quickly while others require much more CPU time it makes sense to
  reorder tasks calcultion in order to reduce total calculation time
  and reduce computers/CPUs idle time. For this purpose -W<num> option
  is added to paexec(1) using weights assigned to each tasks. At the
  moment only two values are allowed for <num>: 0 -- do not use
  weights at all, 1 -- run heavier tasks first as soon as possible.

  New tool paexec_reorder(1) for reordering sliced output of
  paexec(1).  It is written in runawk. So, you'll need it at run time.

  FIXED: 1 byte buffer overflow if -d option is applied.

  Documentation update.  Tons of new regression tests.  Regression
  tests framework has been significantly reworked.  Clean-ups.

  README: notes about my Debian/Lenny/x86 repository.

  Minor fixes (warning messages) for different compilers.

======================================================================
Version 0.14.0, by Aleksey Cheusov, Sun,  3 Jan 2010 12:38:54 +0200

  fixed: "paexec -z" without -s never worked.  Now it is fine.

  New options -g|--graph are synonyms for -s|--pos.

  New option -w for waiting for failed nodes if they ALL failed.
  By default (with -Z) if ALL nodes fail paexec exits with error.

  New option -m for setting alternative strings for
  success/failure/fatal.

  Target for installing directories has been renamed from
  install-dirs to installdirs.

  Other minor fixes and code clean-ups.

======================================================================
Version 0.13.0, by Aleksey Cheusov, Sat,  7 Mar 2009 19:17:03 +0200

  FIXED: When 'paexec -s' retreives 10000 tasks, it allocates
  10000*10000*sizeof(int) bytes for detecting cycles, i.e. ~400Mb on
  32-bit system or 800Mb on 64-bit system. This is absolutely
  inacceptable. This bad algorithm is replaced with new one which
  doesn't need quadratic matrix and works much faster.

  ADDED: -Z<timeout> option
     When I<-z> applied, if a I<command> fails, appropriate node is
     marked as broken and is excluded from the following task
     distribution. But if B<-Z> applied, every I<timeout> seconds an
     attempt to rerun a comand on a failed node is made. I<-Z> implies
     I<-z>. This option makes possible to organize clusters over
     unreliable networks/hardware.

  No EINTR wrappers anymore (iread, xread etc.). SIGCHLD and SIGALRM
  are blocked most of the time. They are unblocked before select(2)
  and blocked just after it. SA_RESTART is not used anymore.

  Minor clean-ups in Makefile.

======================================================================
Version 0.12.1, by Aleksey Cheusov, Tue, 23 Dec 2008 20:33:20 +0200

  FIX: support for -z appeared in paexec-0.12.0 was actually
  incomplete :-/ Now everything should be fixed. More regression
  tests.

  Makefile: adapted for FreeBSD make which doesn't support .PARSEDIR

======================================================================
Version 0.12.0, by Aleksey Cheusov, Mon, 22 Dec 2008 21:48:01 +0200

  ADDED: -z option. If applied, read/write(2) operations from/to nodes
  becomes not critical. In case paexec has lost connection to the
  node, it will reassign failed task to another node and, if -s option
  applied, will output string "fatal" to stdout.  This makes paexec
  resistant to the I/O errors, as a result you can create paexec
  clusters even over network consisting of unreliable hosts
  (Internet?). Failed hosts are marked as such and will not be used
  during the current run of paexec. NOTE: "success", "failure" and
  "fatal" strings should be used to process the output of 'paexec -s'.

  select(2) do not listen stdin (fd=0) anymore. Blocking read(2) is
  used to read tasks from stdin.

  Makefile: CPPFLAGS -> CFLAGS, FreeBSD make Mk scripts
  don't use CPPFLAGS.

======================================================================
Version 0.11.0, by Aleksey Cheusov, Sat, 25 Oct 2008 17:49:47 +0300

   paexec -s: before beginning actual work an input tasks graph is
   checked for cycles. If they are detected, paexec exits with error.

   minor fix in man page

======================================================================
Version 0.10.1, by Aleksey Cheusov, Sat, 25 Oct 2008 02:00:55 +0300

   This version of paexec was sucessfully tested on the following
   platforms:
       NetBSD/x86
       NetBSD/Alpha(64bit)
       Linux/x86
       Linux/x86-64
       Interix-3.5/x86
       FreeBSD/x86
       Solaris-10/x86

   Minor fix for rhomb-like dependencies (paexec -s).
       Suppose A depends on (B1 and B2), (B1 and B2) depends on C.
       If C fails, earlier paexec versions produced
       the following output
                failure
                C B1 A B2 A
       That is, A was output twice. Now it is fixed and A
       is printed only once. New regression test for this.

   By default(!) getopt_long(3) is enabled on the following platforms
   (macro): __NetBSD__, __FreeBSD__, __OpenBSD__, __DragonFly__,
   __linux__ and __APPLE__.  On others only short options provided by
   getopt(3) are used. This means that long option may to not work
   on your platform.

   fix in manual page: accept(2) -> select(2)

   code clean-ups

   .sinclude removed from Makefile

   tests/test.sh:
     diff -U ---> standard diff -C10
     gawk appeared by mistake removed

======================================================================
Version 0.10.0, by Aleksey Cheusov, Sun, 31 Aug 2008 21:37:59 +0300

   Lots of new regression tests

   README file: 'make test' is documented

   ADDED: -s option
       Partially ordered set of tasks are read from stdin.

       Instead of autonomous tasks, graph of the tasks is read from
       stdin.  In this mode every task can either FAIL or SUCCEED.  As
       always an empty line output by command means end of task.  The
       line before it shows an EXIT STATUS of the task.  The word
       "failure" means failure, "success" - success.
       See examples/1_div_x/1_div_X_cmd for the sample.  An input line
       (paexec's stdin) should contain either single task without
       spaces inside or two tasks separated by single space character,
       e.g.  task1<SPC>task2. task1<SPC>task2 line means that task1
       must be done before task2 and it is mandatory, that is if task1
       fail all dependent tasks (including task2) are also failed
       recursively.  Tasks having dependencies are started only after
       all dependencies are succeeded. When a task succeeds paexec
       outputs "success" word just before end_of_task marker (see -e
       or -E), otherwise "failure" word is output followed by a list
       of tasks failed because of it.

              Samples:

                tasks (examples/make_package/make_package_tasks file)

                  textproc/dictem
                  devel/autoconf wip/libmaa
                  devel/gmake wip/libmaa
                  wip/libmaa wip/dict-server
                  wip/libmaa wip/dict-client
                  devel/m4 wip/dict-server
                  devel/byacc wip/dict-server
                  devel/byacc wip/dict-client
                  devel/flex wip/dict-server
                  devel/flex wip/dict-client
                  devel/glib2
                  devel/libjudy

                command (examples/make_package/make_package_cmd__flex)

                  #!/usr/bin/awk -f
                  {
                     print $0 # print a package name

                     if ($0 == "devel/flex")
                        print "failure" # cannot build flex ;-)
                     else
                        print "success" # all other packages are ok

                     print ""       # end of task marker
                     fflush()
                  }

                output of "paexec -s -l -c make_package_cmd__flex -n +10 \
                           < make_package_tasks"

                  3 devel/autoconf
                  3 success
                  4 devel/gmake
                  4 success
                  7 devel/m4
                  7 success
                  8 devel/byacc
                  8 success
                  9 devel/flex
                  9 failure
                  9 devel/flex wip/dict-server wip/dict-client
                  10 devel/glib2
                  10 success
                  11 devel/libjudy
                  11 success
                  1 textproc/dictem
                  1 success
                  2 wip/libmaa
                  2 success

======================================================================
Version 0.9.0, by Aleksey Cheusov, Sun, 15 Jun 2008 10:23:52 +0300

   -t '' means "no transport". This significantly simplifies writing
   shell scripts with paexec. Added: tests for this case.

   paexec has no limited internal buffers anymore. All they are
   resized automatically as it is needed. PAEXEC_BUFSIZE environment
   variable sets an *initial* buffer size, not *maximum* one.

   README: note about non-standard function getopt_long, and advice
   how to build paexec on platforms with no getopt_long support
   (HP-UX, Solaris etc.).

   More regressions tests

   paexec.1: minor corrections.

   'make test' fix: In case regression test fails, 'make test' exits
   with non-zero exit status.

   paexec can be built with ancient version pmake-1.45 (found in som
   Linux distributions).

   paexec -h|--help outputs messages to stderr - my new religion :-)

======================================================================
Version 0.8.0, by Aleksey Cheusov, Tue,  4 Mar 2008 00:32:56 +0200

   New options implemented: -E, -i and -I
   See manual page for details.

   Fixes and minor improvements in the documentation

======================================================================
Version 0.7.0, by Aleksey Cheusov, Tue, 26 Feb 2008 23:57:15 +0200

   new -e|--eot option implemented.
   It prints the end-of-task marker (an empty line) to stdout.
   See manual page.

   WARNS=4 to see compilation warnings.
   several gcc warnings fixed.

   fixed: '-n +0' and '-n ""' now fails with error message.

   minor fix in man page

   minor clean-ups

======================================================================
Version 0.6.0, by Aleksey Cheusov, Thu, 24 Jan 2008 21:31:26 +0200

    First publicly available release

======================================================================