wget/tests/Test-idn-robots.px

#!/usr/bin/env -S perl -I .

use strict;
use warnings;

use WgetFeature qw(iri);
use HTTPTest;

# " Kon'nichiwa <dot> Japan
my $euc_jp_hostname = "\272\243\306\374\244\317.\306\374\313\334";
my $punycoded_hostname = 'xn--v9ju72g90p.xn--wgv71a';
my $escaped_hostname = "%ba%a3%c6%fc%a4%cf.%c6%fc%cb%dc";

###############################################################################

my $starter_file = <<EOF;
<a href="http://$euc_jp_hostname/foo.txt">The link</a>
<a href="http://$punycoded_hostname/foo2.txt">The second link</a>
<a href="http://$escaped_hostname/foo3.txt">The third link</a>
EOF

my $result_file = <<EOF;
Found me!
EOF

# code, msg, headers, content
my %urls = (
    "http://$punycoded_hostname/index.html" => {
        code => "200",
        msg => "Yes, please",
        headers => {
            'Content-Type' => 'text/html; charset=EUC-JP',
        },
        content => $starter_file,
    },
    "http://$punycoded_hostname/foo.txt" => {
        code => "200",
        msg => "Uh-huh",
        headers => {
            'Content-Type' => 'text/plain',
        },
        content => $result_file,
    },
    "http://$punycoded_hostname/foo2.txt" => {
        code => "200",
        msg => "Uh-huh2",
        headers => {
            'Content-Type' => 'text/plain',
        },
        content => $result_file,
    },
    "http://$punycoded_hostname/foo3.txt" => {
        code => "200",
        msg => "Uh-huh3",
        headers => {
            'Content-Type' => 'text/plain',
        },
        content => $result_file,
    },
    "http://$punycoded_hostname/robots.txt" => {
        code => "200",
        msg => "Uh-huh",
        headers => {
            'Content-Type' => 'text/plain',
        },
        content => '',
    },
);

my $cmdline = $WgetTest::WGETPATH . " --iri -r"
    . " -e http_proxy=localhost:{{port}} --local-encoding=EUC-JP"
    . " http://$euc_jp_hostname/";

my $expected_error_code = 0;

my %expected_downloaded_files = (
    "$punycoded_hostname/index.html" => {
        content => $starter_file,
    },
    "$punycoded_hostname/foo.txt" => {
        content => $result_file,
    },
    "$punycoded_hostname/foo2.txt" => {
        content => $result_file,
    },
    "$punycoded_hostname/foo3.txt" => {
        content => $result_file,
    },
    "$punycoded_hostname/robots.txt" => {
        content => '',
    },
);

###############################################################################

my $the_test = HTTPTest->new (input => \%urls,
                              cmdline => $cmdline,
                              errcode => $expected_error_code,
                              output => \%expected_downloaded_files);
exit $the_test->run();

# vim: et ts=4 sw=4