2025-03-15  1.6:  new installation method
                  script is now called textenhance instead of textEnhance

2022-06-29  1.5:  if --left-bracket is set to 'NONE', no comment processing takes place
                  if --whitespace-replacement is set to 'NONE', no replacement takes place
                  (we need these changes because sometimes we have to pass an argument but 
                  cannot pass an empty string (because of shell command line problems; this
                  enables services like 'subtitle' to call textEnhance without any masking actions.

2020-08-25  1.4:  Refactoring and robustification
                  
                  Minor changes:
                  - Arguments are passed to helper script in a more robust and
                    backwards-compatible manner, removing the requirement for
                    bash version >= 4.4
                  - The --logfile option was previously erroneously unmentioned
                    in the Readme, has now been added
                  - Trailing newline is only removed from output of pdftotext
                    and catdoc if it's an unnecessary second one
                  - Availability of helper programs is checked before calling
                    them using the check_available function, ensures meaningful
                    error messages
                  - A .gitattributes file has been added to the repository to
                    ensure that example input files have their encodings, line
                    terminators etc preserved


2020-07-21  1.3:  Eliminate Markdown artifacts
                  
                  Previously, when provided with input files in docx or odt
                  format, the program would sometimes produce output containing
                  certain Markdown-like formatting elements, namely:
                  - Underscores for emphasis
                  - All-caps for headlines and strong text
                  - Indentation for code blocks
                  This was due to the program pandoc, which textEnhance utilizes
                  for converting from docx and odt to plain text, inserting
                  these. A custom filter has now been added to remove any such
                  formatting information.


2020-07-10  1.2:  Document format support
                  
                  Major changes:
                  The program can now handle input files in the most common
                  document formats: pdf, docx, odt, doc and rtf. It will convert
                  them to plain text and then proceed as it does with plain text
                  input files.
                  
                  Minor changes:
                  - Debugging information can now be saved to a file specified
                    with the new --logfile option instead of stdout
                  - The program now filters out any ASCII control characters as
                    well as UTF-8-encoded soft hyphens before writing to output
                  
                  Fixes:
                  - Ensure CR (old Mac style) line breaks are converted


2020-06-05  1.1.1:  Temporary bug fix regarding '#'
                    
                    At the moment, the BAS WebServices cannot handle when an
                    option is passed whose value begins with a hash ('#'); the
                    likely cause is that this gets turned into a shell comment
                    somewhere down the line.
                    As a temporary workaround, textEnhance sets the value of the
                    --left-bracket option to '#', unless the user specifies
                    their own character.


2020-06-03  1.1:  API change

                  The input file is now a required option instead of a
                  positional argument, passed as -i/--infile.


2020-05-18  1.0:  First stable release
                  
                  Major changes:
                  - Restrictions on input file names: files with whitespace or
                    non-ascii characters in the name are refused, as well as
                    files with any of the following document format extensions:
                    doc, docx, xls, xlsx, odt, ods, pdf
                  - Restrictions on input file content: empty files are refused
                  - The user is informed of any changes to the input file (that
                    were not expressly requested with options such as
                    --brackets) with warnings instead of debug messages
                  
                  Minor changes:
                  - Temporary files are only created in /tmp
                  - Handle unknown encodings gracefully
                  
                  Fixes:
                  - Fix bug which resulted in comments being enclosed in '<>'
                    twice
                  - Fix iconv causing core dump with large input files


2020-05-13  0.1:  First alpha version
                  
                  Note: this has accidentally been tagged as 'v0.0' instead of
                  'v0.1' in git.