File Coverage

blib/lib/Locale/Maketext/Utils/Phrase/Norm/Ellipsis.pm

Criterion	Covered	Total	%
statement	66	73	90.4
branch	27	28	96.4
condition	3	3	100.0
subroutine	3	3	100.0
pod	0	1	0.0
total	99	108	91.6

package Locale::Maketext::Utils::Phrase::Norm::Ellipsis;

use strict;
use warnings;

sub normalize_maketext_string {
    my ($filter) = @_;

    my $string_sr = $filter->get_string_sr();

    # 1. placeholder for BN w/ empty string args: ',,'
    while ( ${$string_sr} =~ m/(\[.*?\])/g ) {    # see note about this regex in Consider.pm
        my $bn_match = $1;
        if ( $bn_match =~ m/[,]{2,}/ ) {
            my $bn_match_tmp = $bn_match;
            $bn_match_tmp =~ s/([,]{2,})/my $n=CORE::length("$1");"MULTI_COMMA_IN_BN_$n"/ge;
            ${$string_sr} =~ s/\Q$bn_match\E/$bn_match_tmp/;
        }
    }

    # 2. look for multi's
    if ( ${$string_sr} =~ s/(?:[.]{2,}|[,]{2,})/…/g ) {
        $filter->add_warning('multiple period/comma instead of ellipsis character');
    }

    # 3. restore placeholder
    ${$string_sr} =~ s/MULTI_COMMA_IN_BN_([0-9]+)/"," x "$1"/eg;

    # TODO: output,latin so this occurance is more rare:
    # if ( ${$string_sr} =~ s/([,.]{2,})/\[comment,should “$1” here be an ellipsis?\]/g ) {
    #     $filter->add_warning('multiple concurrent period and comma');
    # }

    if ( ${$string_sr} =~ s/^(|\xc2\xa0|\[output\,nbsp\])…/ …/ ) {
        $filter->add_warning('initial ellipsis should be preceded by a normal space');
    }

    # 1. placeholders for legit ones
    my %l;
    my $copy = ${$string_sr};
    if ( ${$string_sr} =~ s/((?:\x20|\xc2\xa0|\[output\,nbsp\])…[\!\?\.\:])$/ELLIPSIS_END/ ) {    # final
        $l{'ELLIPSIS_END'} = $1;
    }

    if ( ${$string_sr} =~ s/^( …(?:\x20|\xc2\xa0|\[output\,nbsp\]))/ELLIPSIS_START/ ) {           # initial
        $l{'ELLIPSIS_START'} = $1;
    }

    while ( ${$string_sr} =~ m/(\(|\x20|\xc2\xa0|\[output\,nbsp\])…(\)|\x20|\xc2\xa0|\[output\,nbsp\])/g ) {
        ${$string_sr} =~ s/(\(|\x20|\xc2\xa0|\[output\,nbsp\])…(\)|\x20|\xc2\xa0|\[output\,nbsp\])/ELLIPSIS_MEDIAL/;
        push @{ $l{'ELLIPSIS_MEDIAL'} }, [ $1, $2 ];
    }

    # 2. mark any remaining ones (that are not legit)
    if ( ${$string_sr} =~ s/\A …(?!\x20|\xc2\xa0|\[output\,nbsp\])/ … / ) {
        $filter->add_warning('initial ellipsis should be followed by a normal space or a non-break-space (in bracket notation or character form)');
    }

    if ( ${$string_sr} =~ s/…(?:\x20|\xc2\xa0|\[output\,nbsp\]|\s)+\z/…/ ) {
        $filter->add_warning('final ellipsis should be followed by a valid punctuation mark or nothing');
    }

    if ( ${$string_sr} =~ m/…\z/ && ${$string_sr} !~ m/(?:\x20|\xc2\xa0|\[output\,nbsp\])…\z/ ) {
        ${$string_sr} =~ s/…$/ …/;
        $filter->add_warning('final ellipsis should be preceded by a normal space or a non-break-space (in bracket notation or character form)');
    }

    my $medial_prob = 0;
    if ( ${$string_sr} =~ s/(.{1})((?:(?
        $medial_prob++;
    }

    if ( ${$string_sr} =~ s/(.{2})…(?!\x20|\xc2\xa0|\[output\,nbsp\]|\z)(.{1})/$1… $2/g ) {
        $medial_prob++;
    }

    if ($medial_prob) {
        $filter->add_warning('medial ellipsis should be surrounded on each side by a parenthesis or normal space or a non-break-space (in bracket notation or character form)');
    }

    # 3. reconstruct the valid ones
    ${$string_sr} =~ s/ELLIPSIS_END/$l{'ELLIPSIS_END'}/     if exists $l{'ELLIPSIS_END'};
    ${$string_sr} =~ s/ELLIPSIS_START/$l{'ELLIPSIS_START'}/ if exists $l{'ELLIPSIS_START'};
    if ( exists $l{'ELLIPSIS_MEDIAL'} ) {
        for my $medial ( @{ $l{'ELLIPSIS_MEDIAL'} } ) {
            ${$string_sr} =~ s/ELLIPSIS_MEDIAL/$medial->[0]…$medial->[1]/;
        }
    }

    return $filter->return_value;
}

1;

__END__