command-line parsing code, splitting it up into several methods (new
methods: '_parse_command_opts()', '_show_help()') and making it put options
into the 'command_options' dictionary rather than instantiating command
objects and putting them there.

Lots of other little changes:
  * merged 'find_command_class()' and 'create_command_obj()' and
    called the result 'get_command_class()'
  * renamed 'find_command_obj()' to 'get_command_obj()', and added
    command object creation and maintenance of the command object cache to
    its responsibilities (taken over from 'create_command_obj()')
  * parse config files one-at-a-time, so we can keep track of the
    filename for later error reporting
  * tweaked some help messages
  * fixed up many obsolete comments and docstrings

destination path, otherwise show just the directory.

  objects, it now has method names.
Added three methods, 'has_lib()', 'has_scripts()', and 'has_data()'
  to determine if we need to run each of the three possible sub-commands.
Added 'get_sub_commands()' to take care of finding the methods named
  in 'sub_commands', running them, and interpreting the results to
  build a list of sub-commands that actually have to be run.

have any Python modules to install (rather than bomb when we try to copy
a non-existent directory).

implemented by subclasses, since they are needed by 'gen_lib_options()'.

and 'link_shared_lib()'.  In MSVCCompiler, this is meaningful: it adds
/EXPORT: options to the linker command line.  In UnixCCompiler, it
is ignored.

Added some guiding comments.

(Completely uninspected and untested by me, this is just to
get the code into CVS!)

have to worry about "or []" in 'get_outputs()'.

rather than 'data'.

'get_inputs()'.

new flexibility, specifically the 'root' option.  Now, we just use
"install" to do a fake installation into a temporary directory
(the 'bdist_dir' option, which derives from the 'bdist_base' option of
"bdist"), and then tar/zip up that directory.  This means that dumb
built distributions are now relative to the root directory, rather than
the prefix or exec-prefix; this is probably a feature, but does make
them slightly less flexible.

"bdist" command rather than "build".

file to write the list of installed files to.

the list returned by 'get_outputs()', thanks to changes in the
"install_lib" command.

to ensure that compiled bytecode files are considered part of the output
of the "install_lib" command.

last (after writing list of installed files) -- that way, the warning
is more visible.

installed files to INSTALLED_FILES.

top-level temporary directory for creating built distributions.  (Won't
work yet, since the "build" command doesn't yet have a 'build_bdist'
option, and none of the "bdist" commands support it yet.)

variable.

(Is this really needed?  Can we drop it when the config file mechanism
allows users to set compiler flags in setup.cfg?)

the 'no_format_option' class attribute.

correctly ensure that it's 'dirname' exists.

already been set (eg. by a command-line option).

have to wade through all the 'build' output when testing installation.

Caught up with renaming in 'install_misc' base class.
Changed 'run()' to chmod installed scripts under Unix.

  - renamed '_copydata()' to 'copy_files()'
  - changed it to record complete output filenames
  - dropped '_outputdata()' in favour of much simpler 'get_outputs()'

adds the 'install_data' and 'install_scripts' commands; these two
are trivial thanks to the 'install_misc' base class in cmd.py.
(Minor tweaks and commentary by me; the code is untested so far.)

Tweaked error-generating code.

  - added 'find_config_files()' and 'parse_config_files()' methods
  - added 'command_options' attribute
Comment/docstring updates.

added a global '_environ_checked' so we know if it's already been
called.

Also added creation of 'implib_dir', a temporary directory specific to
  MSVC++ -- but I checked in two ways of fixing it (Lyle's and mine),
  because I'm not sure which is right.