ROBODoc tidy

[Dave Offiler]

While testing ROBODoc on RHEL6 by running on a local copy of 5.1_prototype on eld999 with make html, several f90 files threw up warnings about duplicate headers and other inconsistencies.

All modules threw errors about a missing file macros.html.

Also, ropp_pp/doc/Makefile doesn't have an html target in the Makefile, so this needs adding.

HTML generation isn't normally done (make refman makes LaTeX for the Reference Manuals, but the duplicate headers, etc, above will still occur) so these targets probably haven't been tested for several releases. But since this target is much less verbose than refman, the messages don't disappear off the screen and the operation is much quicker.

Not a show-stopper - part of the v5.1 general tidy.

[Dave Offiler]

Doc tidy across all modules, including:

• Updated Makefile.am - found at least one subdir in each module not being scanned by Robodoc; modified 'clean' list
• macros.html is not a generated file, so removed the post-processing of this file
• Updated several source files to clean-up duplicate Robodoc headers or missing trailers
• Removed refman PDF fies (an in one case, the dvi file) from SVN

General code tidy also dealt with the following (principally source code in each modules' tools/ directory):

• Consolidated Robodoc headers to document all command line options consistently
• Added new files ropp_<mod>/common/ropp_<mod>_version.f90 which implement functions ropp_<mod>_version() to return that module's version ID string in a consistent manner. (TODO: ideally the string should be generated automatically via an m4 macro. Also these functions need to be documented in the User Guides). Currently set to "v5.1 1-Sep-2011" in all modules.
• Made all help and version output to stdout consistent both in text output and in programming style; use ropp_<mod>version() to output the tool version ID instead of this string having to be edited in each and every tool's source code
• Ran all code through nagfor with warnings enabled; removed many cases of 'unused variables' and set up dummy use of several cases of 'unused dummy argument'. The latter cases should be re-reviewed and unless there is a reason to retain them, unused arguments should be deleted from the routine and from the calling level. (I've flagged such cases with a "!! fix..." comment)

Checked in by module: UTILS [3008], IO [3009], PP [3010], FM [3011], 1DVAR [3012].