Makefile: clean now targets libsline.*
README.md: added info on using sline directly as a module
Merging history.c into sline.c

After making sline_history_get() public, history.c sort of became a
weird module in isolation. After trying to split out other components
of sline.c into their own modules, which didn't work out at all, I
believe sline is a relatively small codebase that can easily be one
single module (this even opens the possibility of trivially embedding
it into your code without even bothering to use any of the libraries!)
strlcpy.c remains a different module because it's code that comes from
man/: fixing formatting
man/sline_history_get: documenting sline_history_get()
sline_test.c: factoring out error printing
sline_test.c: showcasing sline_history_get()
Private history_get() is now public sline_history_get()
sline_test.c: fixed typo (again)
sline_test.c: fixed missing newline
README.md: typo
Ignoring tabs in input

Tabs are currently ignored by sline in input to avoid the screen and
buffer to get out of sync due to variable tab widths. There doesn't
seem to be any benefit in trying to handle tabs, as sline doesn't
support multi-line editing, where tab-indentation might be of help.
By definining a specific VT_TAB symbol for the internal VT-100 driver,
everything is set in place for implementing some type of user-definable
tab-completion in a future version of sline.
config.mk: improved how we deal with version numbers
Makefile: removed syntax not defined by POSIX
README.md: removed section on no dyn libs
Dyn library support
sline.c: edited history entries treated now like new user input

Now if the user navigates through the input history and decides to
reuse the contents of one entry, edits those, and presses return, the
new edited version is added to the end of the history, leaving the
original unedited.