###############################################################################
#                                                                             #
# EXACT Release 1.1.0 Changelog                                               #
#                                                                             #
###############################################################################

-------------------------------------------------------------------------------
EXACT 1.1.0 (2007/09/20)
-------------------------------------------------------------------------------

- Changes to allow arbitrary information from parser to be passed to
the analysis to be printed on failures, and embedded into the results.xml

- Fixes to baselines

- Changed defaults for base class to warn if necessary functions are not defined

- Should handle zombies now

- Create static data used for XML-ization just once

- Adding a default timelimit of 10min for all experiments.  This can be
overriden with the EXACT_DEFAULT_TIMELIMIT environmental variable, or
with the explicit 'timelimit' XML attribute.

- Fixing errors in the generation of *in files when you
have newlines in the factor value.

- Adding the ability to process 'infinity' values for measurements.  These are set to
very positivie/negative values, since Python does not uniformly handle
Infinity in a standard way.

- Adding R wrapper utilites for statistical tests.

- Translating 'bad' characters that could make the *log.xml an invalid
XML file.

- Revisions to handle --foo (without the ...=bar) factors and options

- Misc fix for option/value management.  Needed to add a __iter__
method to the OptionValueDictionary.

- Adding an analysis mechanism that applies 'lcov' to generate
code coverage web pages.

- Resolved portability issues on windows.

- Revisions of EXACT to remove the --preserve option and add
the --force option (or the -f option).  This changes the
execution semantics such that if an experiment directory or
analysis file is found, then the experiment or analysis is
only run if it is forced.  This has the impact of
not allowing the specification of experiments or analyses with
the same names (which was actually detected in example12.study.xml).

- Consolidated the kill mechanism to ensure consistent behavior
on Linux and MSWindows platforms.  Unfortunately, this doesn't fix the
issue with killing MSWindows applications under Cygwin.


-------------------------------------------------------------------------------
EXACT 1.0.0
-------------------------------------------------------------------------------

- Initial release.
