2010-10-07:
TODO:
* make sure that parse errors for .Rt files result in meaningful output -- try putting <TESTONLY> on a line

2010-09-05:
TODO:
* make evalCapture code handle truncation of long warnings correctly (cf options("warning.length"))

2010-07-13:
TODO:
* think about cleaning up/removing the self-check stuff (this is the stuff that adds calls to the end of tests that go back and see if the tests ran correctly)
* remove function checkTestOutput() -- this is used by the self-check stuff, but nowhere else

2009-08-19:

DONE:
* make reading of Description Collate depend on system
* make function that creates .R and .Rout.save files from .Rt to enable use of the standard R CMD check testing without the ScripTests package
* make it possible to run own-process tests in a new copy of the test dir
* get the interactive tests running
  * optional copy of test directory
      runtests(pattern="plus", "ScripTestsExample", full=T, clobber=T, debug=T)
  * optionally run inside R
* update documentation, esp ScripTests.design.Rd

DONE:
* fix library() output to use package name, not directory name
* use library.dynam() to load DLLs in source.pkg()
* incorporate code from Hadley Wickham's source_package() function http://gist.github.com/180883
* prune obselete functions
* add file= argument to runtests
* pass path argument from runtests to runTestsHereFast
* pass subst arg from runtests to runPackageTests
* add path= argument to runtests()
* put standard entries from CONFIG as defaults (but not in .runPackageTests()
* get a shell script version of tests running - nah, don't need that -- just run from R
* have everything running under "R CMD check" (YAY!)


---------------

Follow plan for dealing with R-2.9.0 testing framework as outlined in man/RtTests.design.Rd

source.pkg("RtTests")
testWrapper("tmp", "tmp/tests")

working in rttests.R: checkTestOutput()
  change read.table to scan() to avoid problems with duplicate row names

these commands in an R session:

source.pkg("RtTests")
setwd("tmp/tests")
# setupTests(T)
setwd("../..")
checkTestOutput("allperfect.Rt", "allperfect.Rt.save", TRUE)
checkTestOutput("plus.Rt", "plus.Rt.save", TRUE)

> setwd("../..")
> source.pkg("RtTests")
Reading 5 .R files into env at pos 2: 'pkgcode:RtTests'
Sourcing RtTests/R/createRfromRt.R
Sourcing RtTests/R/interactive.R
Sourcing RtTests/R/plus.R
Sourcing RtTests/R/rttests.R
Sourcing RtTests/R/testing.R
list()
> setwd("tmp/tests")
> setupTests(T)

(old)
* Find a way of showing versions/paths of all libraries in test output
* Make runtests() print a summary of all the tests run when multiple test files are run
* Make runtests() remove package::: from tests when source.pkg() has been used
* Make runtests() not execute options(error=function() NULL) (or restore options after running?)
