Guessing tab size

bsb has asked for the wisdom of the Perl Monks concerning the following question:

Does anyone have code that tries to guess tab size of an unknown file? (be it Perl or something else)

I'd like to run a snippet of perl when I load a file in vim, that will take a good guess at the tabsize. My code's usually with ts=4 but other stuff is mixed.

I can't find any existing solutions so began writing one. Unfortunately, my 3 week holiday commences in 5 minutes so I'll just leave the very preliminary code I have in the hope that someone else will find the problem interesting enough to solve. Cheeky, I know.

Brad

#!/usr/bin/perl -w

@ARGV or @ARGV = (qw(
    /usr/share/perl/5.6.1/Shell.pm
));

no warnings 'uninitialized';
for my $f (@ARGV) {
    open(F,$f) or die "Open $f:$!";
    
    my $tabs = 0;
    my $ltabs = 0;
    while($_ = <F>) {
        $tabs += tr/\t/\t/;
        $ltabs += () = m/^(\t)/g;
        my ($ls, $first) = m/^(\s+)(\S+)/;
        my ($last) = m/(\S+)\s+$/;
        $ls =~ tr/ /_/;
        my ($four, $eight) = ($ls,$ls);
        $four =~ s/\t/====/g;
        print "$four$first|$last\n" if ($ls);
        $eight =~ s/\t/========/g;
        print " "x40,"$eight$first|$last\n" if ($ls);
    }
    print "$f: $tabs $ltabs\n";
    close F;
}
__END__
'#' should line up
  
leading spaces count 
tabs are multiple of leading spaces
( can have anomalies )
  
if for eval { go out >>
____eval
                                        ____eval
====sub
                                        ========sub
====____if
                                        ========____if
========\$Shell::capture_stderr
                                        ================\$Shell::captu
+re_stderr
====____}
                                        ========____}


} goes in <<
========____\$ret;
                                        ================____\$ret;
========}
                                        ================}
====____}
                                        ========____}
====}
                                        ========}
[download]

Comment on Guessing tab size Download Code

Replies are listed 'Best First'.
Re: Guessing tab size by jmcnamara (Monsignor) on Oct 15, 2003 at 10:17 UTC
Here is a program that a works on the assumption that a file with 4 space tabs will contain some lines that start with 4 spaces while a file with 8 space tabs will not. It skips any Pod in the file. `#!/usr/bin/perl -wn if (not /^=/ .. /^=cut/) { die $tab = "4\n" if /^ \S/; } END { print "8\n" unless $tab; }` [download] This is far from fool-proof however. In fact I'd go as far as to say that no method is fool-proof since a file may have been edited by people with different tab/space styles. -- John.	[reply] [d/l]
Re: Guessing tab size by edan (Curate) on Oct 15, 2003 at 10:44 UTC
Unfortunately, my 3 week holiday commences in 5 minutes I think you need to rethink your priorities in life here... :) -- 3dan	[reply]