| HTTP-WebTest-Plugin-DateTest documentation | Contained in the HTTP-WebTest-Plugin-DateTest distribution. |
HTTP::WebTest::Plugin::DateTest - Evaluate the "age" of embedded date strings in response body
Version 1.01 - $Revision: 1.8 $
Compatible with HTTP::WebTest 2.x API
Not Applicable - see HTTP::WEBTEST
This plugin provides a test for the age of a date string
inside the response body. It supports anything Date::Parse
can parse.
There is limited support for other locales for which a
Date::Language::* module exist.
The Date::Languge and Date::Parse modules are
part of the TimeDate distribution, available from a
CPAN near you.
NOTE: The following parameters date_start, date_end
and date_maxage are lists, so they should be specified
in order for multiple date tests.
Text string which marks the start of a date string in the returned page. The date string should look like anything that Date::Parse is able to understand. Leading/trailing whitespace is no problem
Text string which marks the end of a date string in the returned page. The date string should look like anything that Date::Parse is able to understand. Leading/trailing whitespace is no problem
Maximum age of the parsed date string in seconds. This is evaluated against the current time at runtime.
Format: N [units]
where N is a (floating point-) number, followed by
one of these unit specifiers (case insensitive):
s(econds) - default m(inutes) h(ours) d(ays) w(eeks)
The default is seconds. Only the first character is relevant, any leading text is ignored. An unknown unit specifier defaults to seconds.
Global parameter for all date tests.
Specify the language in which the date string is written.
Locales are taken from Date::Language::(Locale) modules (part
of Date::Parse). The value of date_locale is normalized to
Capitalized notation, so this parameter is not case sensitive.
WARNING: this works by literally translating the date string components to their English names. This fails if the notational conventions are very different (order of day, month, year etc.)
There are languages where abbreviated day- and month names are the same, notably Mardi and Mars in French, and Maandag and Maart in Dutch.
To work around the resulting ambiguity, all non-numerical components are stripped from the left side of the date string. This works for the common case where a weekday starts the string, but doesn't in some other cases.
Example (this will work):
date_locale = 'French' "Mar 19 Mars 2002, 17:25" => "19 Mars 2002, 17:25" => "19 Mar 2002, 17:25"
Example (this will fail for languages other than English):
"Mar 19, 2002 17:25" => "19, 2002 17:25" => (not parsable)
Note: the last notation is very uncommon in Dutch, so this assumption is generally no problem for this locale. YMMV!
Copyright (c) 2002,2003 Johannes la Poutre. All rights reserved.
This module is free software. It may be used, redistributed and/or modified under the terms of the Perl Artistic License.
HTTP::WebTest::API (HTTP::WebTest::API)
HTTP::WebTest::Plugins (HTTP::WebTest::Plugins)
| HTTP-WebTest-Plugin-DateTest documentation | Contained in the HTTP-WebTest-Plugin-DateTest distribution. |
# $Id: DateTest.pm,v 1.8 2003/05/16 08:11:47 joezespak Exp $ package HTTP::WebTest::Plugin::DateTest; use strict; use Date::Parse; use Date::Language::English; use base qw(HTTP::WebTest::Plugin); use vars qw($VERSION); $VERSION = '1.01';
sub param_types { return q( date_start list date_end list date_maxage list date_locale scalar ); } sub check_response { my $self = shift; # response content my $content = $self->webtest->current_response->content; $self->validate_params(qw(date_start date_end date_maxage date_locale)); # test results my @results = (); my @ret = (); # store current time my $now = time(); # check for date strings my @tests = @{$self->test_param('date_start', [])}; my $locale = $self->test_param('date_locale', ''); for (my $i=0; $i < @tests; $i++) { my $maxage = ${$self->test_param('date_maxage', [])}[$i]; my $start = $tests[$i]; my $end = ${$self->test_param('date_end', [])}[$i]; my $pgdate; my $datestr = 'unknown'; if ($content =~ /\Q$start\E\s*(.+?)\s*\Q$end\E/) { $datestr = $1; $pgdate = &_str2time_locale($datestr, $locale); } my $age = ($pgdate) ? $now - $pgdate : 'unknown'; my ($maxsecs, $units) = &_str2seconds($maxage); my $ok = ($age ne 'unknown') && ($age < $maxsecs); push @results, $self->test_result($ok, sprintf("Wanted max %s and got %s (%s)", $maxage, &_seconds2str($age, $units), $datestr) ); } push @ret, ['Max. age of date string', @results] if @results; return @ret; } sub TIMETAB { my $units = shift || ''; my $tt = { s => 1, m => 60, h => 3600, d => 86400, w => 604800, }; return $tt->{$units}; } # look for trailing characters and interprete them as time unit sub _str2seconds { my $date = shift; my $units = ''; if ($date =~ /^\s*([\-+0-9.]+)\s*([smhdwSMHDW]).*/) { $units = lc($2); $date = $1 * &TIMETAB($units); } else { $date =~ s/^\s*([\-+0-9.]+)\s*.*$/$1/g; } $date = 0 unless ($date =~ /^[\-+0-9.]/); return wantarray ? ($date, $units) : $date; } # convert seconds into time string sub _seconds2str { my ($date, $units) = @_; return 'unknown' unless ($date =~ /^[+-]?[\d\.]+$/); if (&TIMETAB($units)) { return ($units eq 's') ? "$date s" : sprintf("%4.2f %s", $date/&TIMETAB($units), $units); } my $str = ''; my $frag = 0; if ($frag = int($date / &TIMETAB('w'))) { $str .= "${frag}w "; $date -= $frag * &TIMETAB('w'); } if ($frag = int($date / &TIMETAB('d'))) { $str .= "${frag}d "; $date -= $frag * &TIMETAB('d'); } $frag = int($date / &TIMETAB('h')); $str .= sprintf "%02d:", $frag; $date -= $frag * &TIMETAB('h'); $frag = int($date / &TIMETAB('m')); $str .= sprintf "%02d:", $frag; $date -= $frag * &TIMETAB('m'); $str .= sprintf "%02d", $date; return $str; } sub _str2time_locale { my $date = shift; my $locale = ucfirst(lc(shift)) || 'English'; # return if date is, well, empty... return if ($date =~ /^\s*$/sg); return str2time $date if ($locale eq 'English'); # normalize spaces, incl. multiple lines $date =~ s/\s+/ /sg; # NOTE: "mar" (Mardi) and "mars" (French) would break. # Assume we have a weekday name prefix if # date string does not start with digits. # Strip up to 1st digit: $date =~ s/^[^0-9]+//g; # strip leading/trailing spaces $date =~ s/^\s*(.+)\s*$/$1/g; eval "require Date::Language::$locale"; warn $@ if $@; my @MoY = eval "\@Date::Language::${locale}::MoY"; my $MoY_EN = \@Date::Language::English::MoY; my $nwdate = ''; foreach my $token (split(/(\s|-|:|\/)/, $date)) { if ($token =~ /^[0-9,-:\/\s]+$/) { $nwdate .= $token; } else { # match token with full month name. for (my $i = 0; $i < @MoY; $i++) { if ($MoY[$i] =~ /^\Q$token\E.*/i) { $nwdate .= $MoY_EN->[$i]; } } } } # explicitly return for empty or '0' date string return unless $nwdate; return str2time $nwdate; }
1;