For example, in 'fc20', these will match 'fc' and '20', respectively.

Things have grown and some parts of layout now make more sense being in different module

Colorized output should bear roughly the same info as plain text even
after stripping color chars (e.g. by c/p)

Now it's getting serious ;)

Due to immature way how mkpretty loads pretifier, `ffood import` is
called with every output, which makes any other debugging output
useless.

Hot fix is to use different switch.  Should the mkpretty mechanism be
improved, we can merge the switches back.

Mkpretty is now always on and defined globally as name of the pretifier,
which is "plain" by default.

"plain" prettifier only adds "debug" and caller name prefix to the debug
output.

usage is supposed to be implemented in user script by call to usage_is
(so that the actual patterns can be defined only once and on the lowest
indent level).

So actually when we see usage_is, we want to skip one FUNCNAME in stack.
(This will not work properly if the suggested usage is not adhered to.)

Use common backend, add -f and -c, add arbitrary pretty printers support

Do not rely on IFS; it does not work with quotes, i.e. spaces in paths.

ffood now employs simpler algorithm to choose file to import and also
is able to list module contents.

I.e. do not use global variable

iniread will now read either specified files or try to find them in
FFOOD_INIPATH, which is empty by default

All other files have been moved to *eecc* and whole qecc
history (~300 commits) has been squashed into this commit.