NAME

Test::Differences - Test strings and data structures and show differences if not ok

SYNOPSIS

use Test;    ## Or use Test::More
use Test::Differences;

eq_or_diff $got,  "a\nb\nc\n",   "testing strings";
eq_or_diff \@got, [qw( a b c )], "testing arrays";

## Passing options:
eq_or_diff $got, $expected, $name, { context => 300 };  ## options

## Using with DBI-like data structures

use DBI;

... open connection & prepare statement and @expected_... here...

eq_or_diff $sth->fetchall_arrayref, \@expected_arrays  "testing DBI arrays";
eq_or_diff $sth->fetchall_hashref,  \@expected_hashes, "testing DBI hashes";

## To force textual or data line numbering (text lines are numbered 1..):
eq_or_diff_text ...;
eq_or_diff_data ...;

EXPORT

This module exports three test functions and four diff-style functions:

  • Test functions

    • eq_or_diff

    • eq_or_diff_data

    • eq_or_diff_text

  • Diff style functions

    • table_diff (the default)

    • unified_diff

    • oldstyle_diff

    • context_diff

DESCRIPTION

When the code you're testing returns multiple lines, records or data structures and they're just plain wrong, an equivalent to the Unix diff utility may be just what's needed. Here's output from an example test script that checks two text documents and then two (trivial) data structures:

t/99example....1..3
not ok 1 - differences in text
#     Failed test ((eval 2) at line 14)
#     +---+----------------+----------------+
#     | Ln|Got             |Expected        |
#     +---+----------------+----------------+
#     |  1|this is line 1  |this is line 1  |
#     *  2|this is line 2  |this is line b  *
#     |  3|this is line 3  |this is line 3  |
#     +---+----------------+----------------+
not ok 2 - differences in whitespace
#     Failed test ((eval 2) at line 20)
#     +---+------------------+------------------+
#     | Ln|Got               |Expected          |
#     +---+------------------+------------------+
#     |  1|        indented  |        indented  |
#     *  2|        indented  |\tindented        *
#     |  3|        indented  |        indented  |
#     +---+------------------+------------------+
not ok 3
#     Failed test ((eval 2) at line 22)
#     +----+-------------------------------------+----------------------------+
#     | Elt|Got                                  |Expected                    |
#     +----+-------------------------------------+----------------------------+
#     *   0|bless( [                             |[                           *
#     *   1|  'Move along, nothing to see here'  |  'Dry, humorless message'  *
#     *   2|], 'Test::Builder' )                 |]                           *
#     +----+-------------------------------------+----------------------------+
# Looks like you failed 3 tests of 3.

eq_or_diff_...() compares two strings or (limited) data structures and either emits an ok indication or a side-by-side diff. Test::Differences is designed to be used with Test.pm and with Test::Simple, Test::More, and other Test::Builder based testing modules. As the SYNOPSIS shows, another testing module must be used as the basis for your test suite.

OPTIONS

The options to eq_or_diff give some fine-grained control over the output.

  • context

    This allows you to control the amount of context shown:

    eq_or_diff $got, $expected, $name, { context => 50000 };

    will show you lots and lots of context. Normally, eq_or_diff() uses some heuristics to determine whether to show 3 lines of context (like a normal unified diff) or 25 lines.

  • data_type

    text or data. This normally defaults to data. If, however, neither of $got or $expected is a reference then it defaults to text. You can also force one or the other by calling eq_or_diff_text or eq_or_diff_data.

    The difference is that in text mode lines are numbered from 1, but in data mode from 0 (and are refered to as 'elements' (Elt) instead of lines):

    # +---+-------+----------+
    # | Ln|Got    |Expected  |
    # +---+-------+----------+
    # *  1|'foo'  |'bar'     *
    # +---+-------+----------+
    
    # +----+---------+----+----------+
    # | Elt|Got      | Elt|Expected  |
    # +----+---------+----+----------+
    # *   0|[        *   0|'bar'     *
    # *   1|  'foo'  *    |          |
    # *   2|]        *    |          |
    # +----+---------+----+----------+

    The difference is purely cosmetic, it makes no difference to how comparisons are performed.

  • Sortkeys

    If passed, whatever value is added is used as the argument for Data::Dumper Sortkeys option. See the Data::Dumper docs to understand how you can control the Sortkeys behavior.

  • filename_a and filename_b

    The column headers to use in the output. They default to 'Got' and 'Expected'.

DIFF STYLES

For extremely long strings, a table diff can wrap on your screen and be hard to read. If you are comfortable with different diff formats, you can switch to a format more suitable for your data. These are the four formats supported by the Text::Diff module and are set with the following functions:

  • table_diff (the default)

  • unified_diff

  • oldstyle_diff

  • context_diff

You can run the following to understand the different diff output styles:

use Test::More 'no_plan';
use Test::Differences;

my $long_string = join '' => 1..40;

TODO: {
    local $TODO = 'Testing diff styles';

    # this is the default and does not need to explicitly set unless you need
    # to reset it back from another diff type
    table_diff;
    eq_or_diff $long_string, "-$long_string", 'table diff';

    unified_diff;
    eq_or_diff $long_string, "-$long_string", 'unified diff';

    context_diff;
    eq_or_diff $long_string, "-$long_string", 'context diff';

    oldstyle_diff;
    eq_or_diff $long_string, "-$long_string", 'oldstyle diff';
}

UNICODE

Generally you'll find that the following test output is disappointing.

use Test::Differences;
use utf8;

my $want = { 'Traditional Chinese' => '中國' };
my $have = { 'Traditional Chinese' => '中国' };

eq_or_diff $have, $want, 'Unicode, baby';

Here's what you get:

#   Failed test 'Unicode, baby'
#   at t/unicode.t line 12.
# +----+-----------------------+-----------------------+
# | Elt|Got                    |Expected               |
# +----+-----------------------+-----------------------+
# |   0|'Traditional Chinese'  |'Traditional Chinese'  |
# *   1|'\x{4e2d}\x{56fd}'     |'\x{4e2d}\x{570b}'     *
# +----+-----------------------+-----------------------+
# Looks like you failed 1 test of 1.
Dubious, test returned 1 (wstat 256, 0x100)
Failed 1/1 subtests

A patch to fix this would be *most* welcome.

Unknown::Values

Unknown::Values is a module which provides values which will never compare as being the same as anything else, not even the same as itself.

If code looks too hard at one of these values (and Test::Differences looks very hard indeed) that is a fatal error. This means that while we can detect the presence of these beasties, and tell you that they compare different, for Complicated Internals Reasons we can't show you much context. Sorry.

NB that the support for these is experimental and relies on an undocumented unstable interface in Unknown::Values. If that fails then Test::Differences will probably just die when it sees them instead of telling you that the comparison failed.

LIMITATIONS

Test or Test::More

This module "mixes in" with Test.pm or any of the test libraries based on Test::Builder (Test::Simple, Test::More, etc). It does this by peeking to see whether Test.pm or Test/Builder.pm is in %INC, so if you are not using one of those, it will print a warning and play dumb by not emitting test numbers (or incrementing them). If you are using one of these, it should interoperate nicely.

Exporting

Exports all 3 functions by default (and by design). Use

use Test::Differences ();

to suppress this behavior if you don't like the namespace pollution.

This module will not override functions like ok(), is(), is_deeply(), etc. If it did, then you could eval "use Test::Differences qw( is_deeply );" to get automatic upgrading to diffing behaviors without the sub my_ok shown above. Test::Differences intentionally does not provide this behavior because this would mean that Test::Differences would need to emulate every popular test module out there, which would require far more coding and maintenance that I'm willing to do. Use the eval and my_ok deployment shown above if you want some level of automation.

Booleans

Historically, true and false (whether resulting from logical-not ! or from equality checks like ==) were represented in perl as numeric 1 for true and the empty string '' for false. In perl 5.36.0 they became real Boolean values which for backward compatibility also have their old numeric/string values. In 5.38 (and in dev releases from 5.37.2 onwards) Data::Dumper belatedly learned about this, and so Test::Differences (which uses Data::Dumper under the bonnet) can also tell the difference.

This means that `eq_or_diff(1, !!1)` will behave differently depending on whether you are using perl 5.38 or later (in which case it will say they are different) or 5.36 or earlier (which will say they are the same).

Unicode

Perls before 5.6.0 don't support characters > 255 at all, and 5.6.0 seems broken. This means that you might get odd results using perl5.6.0 with unicode strings.

Data::Dumper and older Perls.

Relies on Data::Dumper (for now), which, prior to perl5.8, will not always report hashes in the same order. $Data::Dumper::Sortkeys is set to 1, so on more recent versions of Data::Dumper, this should not occur. Check CPAN to see if it's been peeled out of the main perl distribution and backported. Reported by Ilya Martynov <ilya@martynov.org>, although the Sortkeys "future perfect" workaround has been set in anticipation of a new Data::Dumper for a while. Note that the two hashes should report the same here:

not ok 5
#     Failed test (t/ctrl/05-home.t at line 51)
# +----+------------------------+----+------------------------+
# | Elt|Got                     | Elt|Expected                |
# +----+------------------------+----+------------------------+
# |   0|{                       |   0|{                       |
# |   1|  'password' => '',     |   1|  'password' => '',     |
# *   2|  'method' => 'login',  *    |                        |
# |   3|  'ctrl' => 'home',     |   2|  'ctrl' => 'home',     |
# |    |                        *   3|  'method' => 'login',  *
# |   4|  'email' => 'test'     |   4|  'email' => 'test'     |
# |   5|}                       |   5|}                       |
# +----+------------------------+----+------------------------+

Data::Dumper also overlooks the difference between

$a[0] = \$a[1];
$a[1] = \$a[0];   # $a[0] = \$a[1]

and

$x = \$y;
$y = \$x;
@a = ( $x, $y );  # $a[0] = \$y, not \$a[1]

The former involves two scalars, the latter 4: $x, $y, and @a[0,1]. This was carefully explained to me in words of two syllables or less by Yves Orton <demerphq@hotmail.com>. The plan to address this is to allow you to select Data::Denter or some other module of your choice as an option.

Code-refs

Test::Differences turns on $Data::Dumper::Deparse, so any code-refs in your data structures will be turned into text before they are examined, using B::Deparse. The precise text generated for a sub-ref might not be what you expect as it is generated from the compiled version of the code, but it should at least be consistent and spot differences correctly.

You can turn this behaviour off by setting $Test::Differences::NoDeparse.

AUTHORS

Barrie Slaymaker <barries@slaysys.com> - original author

Curtis "Ovid" Poe <ovid@cpan.org>

David Cantrell <david@cantrell.org.uk>

LICENSE

Copyright Barrie Slaymaker, Curtis "Ovid" Poe, and David Cantrell.

All Rights Reserved.

You may use, distribute and modify this software under the terms of the GNU public license, any version, or the Artistic license.