parse
The class method "parse" tries to determine the indentation style of the given piece of text (which must
start at a new line and can be passed in either as a string or as a reference to a scalar containing the
string).
Returns a letter followed by a number. If the letter is "s", then the text is most likely indented with
spaces. The number indicates the number of spaces used for indentation. A "t" indicates tabs. The number
after the "t" indicates the number characters each level of indentation corresponds to. A "u" indicates
that the indenation style could not be determined. Finally, an "m" followed by a number means that this
many characters are used for each indentation level, but the indentation is an arbitrary number of tabs
followed by 0-7 spaces. This can happen if your editor is stupid enough to do smart
indentation/whitespace compression. (I.e. replaces all indentations many tabs as possible but leaves the
rest as spaces.)
The function supports parsing of "vim" modelines. Those settings override the heuristics. The modeline's
options that are recognized are "sts"/"softtabstob", "et"/"noet"/"expandtabs"/"noexpandtabs", and
"ts"/"tabstop".
Similarly, parsing of "emacs" LocalVariables is somewhat supported. "parse" use explicit settings to
override the heuristics but uses style settings only as a fallback. The following options are recognized:
"tab-width", "indent-tabs-mode", "c-basic-offset", and "style".
There is one named option that you can pass to parse(): "skip_pod". When set to true, any section of POD
(see perlpod) will be ignored for indentation finding. This is because verbatim paragraphs and examples
embedded in POD or quite often indented differently from normal Perl code around the POD section.
Defaults to false. Example:
my $mode = Text::FindIndent->parse(\$text, skip_pod => 1);
to_vim_commands
A class method that converts the output of parse(\$text) into a series of vi(m) commands that will
configure vim to use the detected indentation setting. Returns zero (failure) or more lines of text that
are suitable for passing to VIM::DoCommand() one by one.
As a convenience, if the argument to "to_vim_commands" doesn't look like the output of "parse", it is
redirected to "parse" first.
To use this, you can put the following line in your .vimrc if your vim has Perl support. Suggestions on
how to do this in a more elegant way are welcome. The code should be on one line but is broken up for
displaying:
map <F5> <Esc> :perl use Text::FindIndent;VIM::DoCommand($_) for
Text::FindIndent->to_vim_commands(join "\n", $curbuf->Get(1..$curbuf->Count()));<CR>
(Patches to implement the equivalent for emacs would be welcome as well.)