From: Junio C Hamano <gitster@pobox.com>
Date: Sat, 23 Feb 2008 22:04:05 GMT
Subject: Re: [PATCH] gitweb: Better chopping in commit search results
Message-ID: <7vd4qn1ga2.fsf@gitster.siamese.dyndns.org>
In-Reply-To: <200802222014.13205.jnareb@gmail.com>

Jakub Narebski <jnareb@gmail.com> writes:

> 	# regexps: ending and beginning with word part up to $add_len
> 	my $endre = qr/.{0,$len}[^ \/\-_:\.@]{0,$add_len}/;
> 	my $begre = qr/[^ \/\-_:\.@]{0,$add_len}.{0,$len}/;

I have no idea what these line noise characters inside [] are.
Did you mean something like "\w"?

I have a suspicion that it may be easier to read and could be
even more efficient to split an overlong line at word boundaries
and to remove elements from the end you are removing from until
it fits.

sub chop_whence {
	my ($line, $max, $slop, $where) = @_;

	my $len = length($line);
	if ($len < $max + $slop) {
		return $line;
	}

	# Cut at word boundaries
	my @split = split(/\b/, $line);
	my $filler = "...";

	while ((2 < @split)) {
		my $removed;
		my $splice_at;
		if ($where eq 'left') {
			$removed = shift @split;
		} elsif ($where eq 'right') {
			$removed = pop @split;
		} else {
			my $splice_at = int($#split / 2);
			$removed = splice(@split, $splice_at, 1);
		}
		$len -= length($removed) + length($filler);
		if ($len < $max + $slop) {
			if ($where eq 'left') {
				unshift @split, $filler;
			} elsif ($where eq 'right') {
				push @split, $filler;
			} else {
				my $splice_at = int($#split / 2);
				splice(@split, $splice_at, 0, $filler);
			}
			return join('', @split);
		}
	}
	# give up
	return $line;
}