0.1   First try.  Only does sign test and associated confidence interval.

 0.2   Now does Mann-Whitney-Wilcoxon (two sample) test and associate c. i.

 0.3   Finally!!!  Now does Wilcoxon (one sample) test and associate c. i.

 0.3-1   Removed test under WinDoze that makes it segfault under WinDoze.

 0.3-2   CHAR macro in Rinternals.h now returns const char * (or gcc now
         warns about).  Anyway need to make corresponding changes in my code.

         Also wording of blather from various bits of R had changed so
         updated .Rout.save in tests directory accordingly

         Also to typos had been corrected in inst/doc/design.tex but
         were not in the the previous version.

 0.3-3   Made R CMD check (2.15.1) not barf

 0.3-4   Licence issues

 0.3-5   Extensive additions to design doc (inst/doc/design.tex) to go with
         submission of paper

 0.3-6   explicitly import from stats, graphics as required by R-3.3.0

 0.3-7   got rid of .Platform$OS.type == "unix" in tests because of
             complaint from CRAN

 0.3-8   modernize registration of native routines

 0.3-9   add DOI to reference in DESCRIPTION file
         also comment out one test with confusing result in tests/sign-ci.R

0.4      call C functions wilcox_free and signrank_free as per
         Writing R Extensions