Test::Timer - a test module to test/assert response times
The documentation in this module describes version 0.09 of Test::Timer
use Test::Timer;
time_ok( sub { doYourStuffButBeQuickAboutIt(); }, 1, 'threshold of one second');
time_atmost( sub { doYourStuffYouHave10Seconds(); }, 10, 'threshold of 10 seconds');
time_between( sub { doYourStuffYouHave5-10Seconds(); }, 5, 10,
'lower threshold of 5 seconds and upper threshold of 10 seconds');
#Will succeed
time_nok( sub { sleep(2); }, 1, 'threshold of one second');
time_atleast( sub { sleep(2); }, 2, 'threshold of one second');
#Will fail after 5 (threshold) + 2 seconds (default alarm)
time_ok( sub { while(1) { sleep(1); } }, 5, 'threshold of one second');
$test::Timer::alarm = 6 #default 2 seconds
#Will fail after 5 (threshold) + 6 seconds (specified alarm)
time_ok( sub { while(1) { sleep(1); } }, 5, 'threshold of one second');
Test::Timer implements a set of test primitives to test and assert test times from bodies of code.
The key features are subroutines to assert or test the following:
- that a given piece of code does not exceed a specified time limit
- that a given piece of code takes longer than a specifed time limit and does not exceed another
Test::Timer exports:
Takes the following parameters:
-
a reference to a block of code (anonymous sub)
-
a threshold specified as a integer indicating a number of seconds
-
a string specifying a test name
time_ok( sub { doYourStuffButBeQuickAboutIt(); }, 1, 'threshold of one second');
If the execution of the code exceeds the threshold the test fails
The is the inverted variant of time_ok, it passes if the threshold is exceeded and fails if the benchmark of the code is within the specified timing threshold.
The API is the same as for time_ok.
time_nok( sub { sleep(2); }, 1, 'threshold of one second');
This is syntactic sugar for time_ok
time_atmost( sub { doYourStuffButBeQuickAboutIt(); }, 1, 'threshold of one second');
time_atleast( sub { sleep(2); }, 1, 'threshold of one second');
The test succeeds if the code takes at least the number of seconds specified by the timing threshold.
Please be aware that Test::Timer, breaks the execution with an alarm specified to trigger after the specified threshold + 2 seconds, so if you expect your execution to run longer, set the alarm accordingly.
$Test::Timer::alarm = $my_alarm_in_seconds;
See also diagnostics.
This method is a more extensive variant of time_atmost and time_ok, you can specify a lower and upper threshold, the code has to execute within this interval in order for the test to succeed
time_between( sub { sleep(2); }, 5, 10,
'lower threshold of 5 seconds and upper threshold of 10 seconds');
This is a method to handle the result from _benchmark is initiates the benchmark calling benchmark and based on whether it is within the provided interval true (1) is returned and if not false (0).
This is a simpler variant of the method above, it is the author's hope that is can be refactored out at some point, due to the similarity with _runtest.
This is the method doing the actual benchmark, if a better method is located this is the place to do the handy work.
Currently Benchmark is used. An alternative could be Devel::Timer, but I do not know this module very well and Benchmark is core, so this is used for now.
The method takes two parameters:
- a code block via a code reference
- a threshold (the upper threshold, since this is added to the default alarm.
This is the method extracts the seconds from benchmarks timestring and returns it and an integer.
It takes the timestring from _benchmark (Benchmark) and returns the seconds part.
Test::Builder required import to do some import hokus-pokus for the test methods exported from Test::Timer. Please refer to the documentation in Test::Builder
All tests either fail or succeed, but a few exceptions are implemented, these are listed below.
-
Test did not exceed specified threshold, this message is diagnosis for time_atleast and time_nok tests, which do not exceed their specified threshold.
-
Test exceeded specified threshold, this message is a diagnostic for time_atmost and time_ok, if the specified threshold is surpassed.
This is the key point of the module, either your code is too slow and you should address this or your threshold is too low, in which case you can set it a bit higher and run the test again.
-
Test did not execute within specified interval, this is the diagnostic from time_between, it is the diagnosis if the execution of the code is not between the specified lower and upper thresholds.
-
Insufficient parameters, this is the message if a specified test is not provided with the sufficent number of parameters, consult this documentation and correct accordingly.
-
Execution exceeded threshold and timed out, the exception is thrown if the execution of tested code exceeds even the alarm, which is default 2 seconds, but can be set by the user or is equal to the uppertreshold + 2 seconds.
The exception results in a diagnostic for the failing test. This is a failsafe to avoid that code runs forever. If you get this diagnose either your code is too slow and you should address this or it might be error prone. If this is not the case adjust the alarm setting to suit your situation.
This module requires no special configuration or environment.
Tests are sensitive and be configured using environment and configuration files, please see the section on test and quality.
This module holds no known incompatibilities.
This module holds no known bugs.
As listed on the TODO, the current implementations only use seconds and resolutions should be higher, so the current implementation is limited to seconds as the highest resolution.
Coverage report for the release described in this documentation (see VERSION).
---------------------------- ------ ------ ------ ------ ------ ------ ------
File stmt bran cond sub pod time total
---------------------------- ------ ------ ------ ------ ------ ------ ------
blib/lib/Test/Timer.pm 91.7 93.8 66.7 88.5 100.0 99.9 90.7
...Timer/TimeoutException.pm 100.0 n/a n/a 100.0 100.0 0.1 100.0
Total 93.1 93.8 66.7 90.6 100.0 100.0 92.1
---------------------------- ------ ------ ------ ------ ------ ------ ------
The Test::Perl::Critic test runs with severity 5 (gentle) for now, please
refer to t/critic.t
and t/perlcriticrc
.
Set TEST_POD to enable Test::Pod test in t/pod.t
and Test::Pod::Coverage
test in t/pod-coverage.t
.
Set TEST_CRITIC to enable Test::Perl::Critic test in t/critic.t
This distribution uses Travis and Jenkins for continuous integration testing, the Travis reports are public available.
- Implement higher resolution for thresholds
- Factor out _runtest_atleast
- Add more tests to get a better feeling for the use and border cases requiring alarm etc.
- Rewrite POD to emphasize time_atleast over time_ok
Please report any bugs or feature requests either using rt.cpan.org or Github
I will be notified, and then you'll automatically be notified of progress on your bug as I make changes.
- RT mail interface:
bug-test-timer at rt.cpan.org
- RT web interface: http://rt.cpan.org/NoAuth/ReportBug.html?Queue=Test-Timer
- Github: https://github.com/jonasbn/testt/issues
You can find (this) documentation for this module with the perldoc
command.
perldoc Test::Timer
You can also look for information at:
- Jonas B. Nielsen (jonasbn)
<jonasbn at cpan.org>
- Gabor Szabo (GZABO), suggestion for specification of interval thresholds even though this was obsoleted by the later introduced time_between
- Paul Leonerd Evans (PEVANS), suggestions for time_atleast and time_atmost and the handling of $SIG{ALRM}.
- brian d foy (BDFOY), for patch to _run_test
Test::Timer and related modules are (C) by Jonas B. Nielsen, (jonasbn) 2007-2014
Test::Timer and related modules are released under the artistic license
The distribution is licensed under the Artistic License, as specified by the Artistic file in the standard perl distribution (http://www.perl.com/language/misc/Artistic.html).