Overview of changes:

 *  Added ability to reload modules

    `shellfu reload` now reloads module that was once loaded.  This
    enables applications re-init modules if they have changed envvars
    that affect them, for example the dumb plugin system in pretty.sh
    can be reinitialized from outside if application wants to change
    messaging mode.

 *  pretty.sh now won't enforce debuging and verbosity modes on import

    Ie. it abandons the habit of setting PRETTY_VERBOSE and PRETTY_DEBUG
    on import; instead properly deals with them missing.  Turns out this
    costs nothing and avoids a mild security risk by relying on value
    comparison instead of running (supposedly!) `true`/`false` builtins.

Useful for modules like pretty.sh so that they can benefit from safety and
simplicity of having default but app can still reload it.

Note that in most cases using sfpi is cleaner and does not need
reloading, but pretty.sh is using dumb way of loading.

Also don't rely on running the value; ie. sec risk.

Overview of changes:

 *  Fix issue with sfpi__key() not ever returning anything

    sfpi__key() would not return anything unless plugin was already
    loaded.  Now the plugin is loaded (in a subshell) before aking for
    the key value.

Overview of changes:

 *  Fixed typo in usage error template

 *  Updated mkusage() docstring about usage error templates

Overview of changes:

 *  Added extra shorthand arguments for common usage errors

    Things like missing positional arguments or unknown arguments need
    to be dealt with so often it warrants these shorthands.

 *  Fortified command, file and pipe delimiters

    Often it's helpful to see more clearly where output begins and where
    it ends.  Note that this is a small step from pretty.sh's total
    minimalism, on the other hand these file and command printers are
    not exactly minimalistic already; they are not intended for normal
    use but rather for debugging.

Often it helps to see better where output begins or end exactly.

Overview of changes:

 *  Added arr.sh, couple of utilities for array manipulation

 *  Codebase maintenance

Overview of changes:

 *  Added 'forcecolor' pretty.sh module

    Just as 'color' but does not disable color in scripts et al.

Same as 'color' but does not turn off the colors on non-terminal tty.

Overview of changes:

 *  Updated MKit to v0.0.37

Overview of changes:

 *  Fixed coerce__noctl() to actually do something

    Turns out that the function could not ever work properly since
    Perl's tr/// operator does not support POSIX character classes like
    traditional tr(1) does.  The expression was replaced with s///
    variant, preserving the same character range.

    Also it will now *replace* characters $COERCE__REPCHAR, instead of
    removing them.  (Removing behavior can be achieved by setting the
    variable to empty string.)

Make it more consistent with other similar functions.  Note that one can
always get the old behavior by setting $COERCE__REPCHAR to empty string.

Turns out tr/// *does not support POSIX classes* at all.  This command
could not have ever worked.  (schroedinbug).

Overview of changes:

 *  Enabled shellfu-sh-coerce to avoid Term::ANSIColor dependency on RHEL-6

    Turns out this module is not easily available on RHEL-6, so by adding
    the dependency in recent version, we broke shellfu-sh-coerce and its
    reverse dependencies there.

Overview of changes:

 *  Fixed debian packaging

debian/control does not allow creating dependency lists by repeating
Depends: field, like RPM .spec files do.  I did not realize it when
adding coerce.sh

Overview of changes:

 *  Fixed few minor packaging problems

Overview of changes:

 *  Added module coerce.sh for character set coercion

 *  sfdoc will jump to first object mention in object mode (`-O`)

 *  Minor fixes and cleanups

For situations when there are constraints on what characters are
allowed, this module provides several non-reversible conversion
functions.

Overview of changes:

 *  Improved an updated tab completion for sfdoc

    The tab completion now offers more recent argument and most
    importantly, module and/or object names.

 *  Fixed global variable '$modname' leaking from shellfu()

 *  Fixed --which treated as if no value was needed

 *  Mitigated risk of grep ambiguities

 *  Smaller cleanups and updates

Define it as local and keep it.

Where regular expression is read from a variable and -e argument is not
provided, the resulting call may end up being interpreted as grep
argument,  This can be pretty dangerous, eg. if the variable expands to
`--help`, grep help is shown and grep exits with zero, which would be
interprered as match.  Another example is when the variable expands to
a valid grep parameter; this would mean that next argument would be
interprered by grep as the regex, and if the argument after that would
be missing, grep would read stdin, resulting in data messup or grep
waiting
indefinitely.

See also:

    https://github.com/koalaman/shellcheck/issues/1342

It's been followed for years in number of projects (granted, all being
exclusively written by author of this style), so calling it "draft"
seems just weasely.

Overview of changes:

 *  Added SFPI functions to find and dereference plugin variables

 *  Renamed SFPI name-access functions to be less confusing

 *  Fixed broken sfpi__import()

 *  Finished style guide on naming, capitalization and namespacing

 *  Minor fixes and cleanups

I was about to write something about blank lines between declarations
but that does not seem to be necessary.