SYNOPSIS
SWAT is Simple Web Application Test ( Tool )
$ swat examples/google/ google.ru
/home/vagrant/.swat/reports/google.ru/00.t ..
# start swat for google.ru/
# try num 2
ok 1 - successful response from GET google.ru/
# data file: /home/vagrant/.swat/reports/google.ru/content.GET.txt
ok 2 - GET / returns 200 OK
ok 3 - GET / returns Google
1..3
ok
All tests successful.
Files=1, Tests=3, 12 wallclock secs ( 0.00 usr 0.00 sys + 0.02 cusr 0.00 csys = 0.02 CPU)
Result: PASS
WHY
I know there are a lot of test tools and frameworks, but let me briefly
tell *why* I created swat. As devops, I update dozens of web application
weekly, sometimes I just have *no time* to sit and wait, while dev guys
or QA team ensure that deployment is fine and nothing breaks on the
road. So I need a tool to run smoke tests against web applications. Not
just a tool, but the way to create such tests from scratch in a way
that's easy and fast enough.
So this is how I came up with the idea of swat.
Key features
SWAT:
* is a very pragmatic tool, designed for the job to be done in a fast
and simple way
* has simple and yet flexible DSL with low price mastering ( see my
tutorial )
* produces TAP output
* leverages famous perl prove
and curl utilities
Install
Swat relies on curl utility to make http requests. Thus first you need
to install curl:
$ sudo apt-get install curl
Also swat client is a bash script so you need bash.
Then you install swat cpan module:
sudo cpan install swat
Install from source
# useful for contributors
perl Makefile.PL
make
make test
make install
Swat mini tutorial
For those who love to make long story short ...
Create tests
mkdir my-app/ # create a project root directory to contain tests
# define http URIs application should response to
mkdir -p my-app/hello # GET /hello
mkdir -p my-app/hello/world # GET /hello/world
# define the content to return by URIs
echo 200 OK >> my-app/hello/get.txt
echo 200 OK >> my-app/hello/world/get.txt
echo 'This is hello' >> my-app/hello/get.txt
echo 'This is hello world' >> my-app/hello/world/get.txt
Run tests
swat ./my-app http://127.0.0.1
DSL
Swat DSL consists of 2 parts. Routes and Swat Data.
Routes
Routes are http resources a tested web application should have.
Swat utilize file system to get know about routes. Let we have a
following project layout:
example/my-app/
example/my-app/hello/
example/my-app/hello/get.txt
example/my-app/hello/world/get.txt
When you give swat a run
swat example/my-app 127.0.0.1
It will find all the *directories with get.txt or post.txt or put.txt
files inside* and "create" routes:
GET hello/
GET hello/world
It is possible to run a single swat test setting a "test_file" variable:
test_file=hello/get.txt swat example/my-app 127.0.0.1
When you are done with routes you need to set swat data.
Swat data
Swat data is DSL to describe/generate validation checks you apply to
content returned from web application.
Swat data is stored in swat data files, named get.txt or post.txt or
put.txt.
The validation process looks like:
* Swat recursively find files named get.txt or post.txt or put.txt in
the project root directory to get swat data.
* Swat parse swat data file and *execute* entries found. At the end of
this process swat creates a *final check list* with "Check
Expressions".
* For every route swat makes http requests to web application and
store content into text file
* Every line of text file is validated by every item in a *final check
list*
*Objects* found in test data file are called *swat entries*. There are
*3 basic type* of swat entries:
* Check Expressions
* Comments
* Perl Expressions and Generators
Check Expressions
This is most usable type of entries you may define at swat data file.
*It's just a string should be returned* when swat request a given URI.
Here are examples:
200 OK
Hello World
Hello World
Using regexps
Regexps are check expressions with the usage of instead of plain strings checks. Everything started with
"regexp:" marker would be treated as perl regular expression.
# this is example of regexp check
regexp: App Version Number: (\d+\.\d+\.\d+)
Comments
Comments entries are lines started with "#" symbol, swat will ignore
comments when parse swat data file. Here are examples.
# this http status is expected
200 OK
Hello World # this string should be in the response
Hello World # and it should be proper html code
Matching block of text
Sometimes it is very helpful match a content *not against a single
string*, but against a "block of text", like here:
# this block of text
# consists of 5 strings should be at output:
begin:
# plain strings
this string followed by
that string followed by
another one
# regexps patterns:
regexp: with (this|that)
# and the last one in a block
at the very end
end:
This kind of check should be passed when running against for example
this block of text:
this string followed by
that string followed by
another one string
with that string
at the very end.
But won't be passed against this block of text:
that string followed by
this string followed by
another one string
with that string
at the very end.
"begin:" "end:" markers decorate `block of text` to be found at return
content.
Markers should not be followed by any text at the same line.
Also be aware if you leave "dangling" begin: marker without closing end:
somewhere else this will result in `block-of-text` mode till the end of
your test, which is probably not you want:
begin:
here we begin
and till the very end of test
we are in `block-of-text` mode
Perl Expressions
Perl expressions are just a pieces of perl code to *get evaled* by swat
when parsing test data files.
Everything started with "code:" marker would be treated by swat as perl
code to execute. There are a *lot of possibilities*! Please follow
Test::More documentation to get more info about useful function you may
call here.
code: skip('next test is skipped',1) # skip next check forever
HELLO WORLD
code: skip('next test is skipped',1) unless $ENV{'debug'} == 1 # conditionally skip this check
HELLO SWAT
Generators
Swat entries generators is the way to *create new swat entries on the
fly*. Technically speaking it's just a perl code which should return an
array reference: Generators are very close to perl expressions (
generators code is also get evaled ) with major difference:
Value returned from generator's code should be array reference. The
array is passed back to swat parser so it can create new swat entries
from it.
Generators entries start with ":generator" marker. Here is example:
# Place this in swat data file
generator: [ qw{ foo bar baz } ]
This generator will generate 3 swat entries:
foo
bar
baz
As you can guess an array returned by generator should contain *perl
strings* representing swat entries, here is another example: with
generator producing still 3 swat entities 'foo', 'bar', 'baz' :
# Place this in swat date file
generator: my %d = { 'foo' => 'foo value', 'bar' => 'bar value' }; [ map { ( "# $_", "$data{$_}" ) } keys %d ]
This generator will generate 3 swat entities:
# foo
foo value
# bar
bar value
There is no limit for you! Use any code you want with only requirement -
it should return array reference. What about to validate web application
content with sqlite database entries?
# Place this in swat data file
generator: \
use DBI; \
my $dbh = DBI->connect("dbi:SQLite:dbname=t/data/test.db","",""); \
my $sth = $dbh->prepare("SELECT name from users"); \
$sth->execute(); \
my $results = $sth->fetchall_arrayref; \
[ map { $_->[0] } @${results} ]
As an example take a loot at examples/swat-generators-sqlite3 project
Multiline expressions
Sometimes code looks more readable when you split it on separate chunks.
When swat parser meets "\" symbols it postpone entry execution and add
next line to buffer. This is repeated till no "\" found on next. Finally
swat execute *"accumulated"* swat entity.
Here are some examples:
# Place this in swat data file
generator: \
my %d = { \
'foo' => 'foo value', \
'bar' => 'bar value', \
'baz' => 'baz value' \
}; \
[ \
map { ( "# $_", "$data{$_}" ) } keys %d \
] \
# Place this in swat data file
generator: [ \
map { \
uc($_) \
} qw( foo bar baz ) \
]
code: \
if $ENV{'debug'} == 1 { # conditionally skip this check \
skip('next test is skipped',1) \
}
HELLO SWAT
Multiline expressions are only allowable for perl expressions and
generators
Generators and Perl Expressions Scope
Swat uses *perl string eval* when process generators and perl
expressions code, be aware of this. Follow
to get more on this.
PERL5LIB
Swat adds $project_root_directory/lib to PERL5LIB , so this is
convenient convenient to place here custom perl modules:
example/my-app/lib/Foo/Bar/Baz.pm
Take a look at examples/swat-generators-with-lib/ for working example
Captures
Captures is a way to *access* the data captured in an *last regexp
check*.
Perl expressions and code generators could access captures calling
"captures()" function. Captures() returns array reference holding all
captures found during *last regexp check*. Here is a couple of exmaples:
# Place this in swat data file
# check if outpiut contains digits
# calculate total amount
# it should be > then 10
regexp: (\d+)
code: \
my $total=0; \
for my $c (@{captures()}) { \
$total+=$c->[0]; \
} \
cmp_ok( $total,'>',10,"total amount is greater than 10" );
# Place this in swat data file
# check if output contains lines with date in `date: YYYY-MM-DD` format
# check if first date found is yesterday
regexp: date: (\d\d\d\d)-(\d\d)-(\d\d)
code: \
use DateTime; \
my $c = captures()->[0]; \
my $dt = DateTime->new( year => $c->[0], month => $c->[1], day => $c->[2] ); \
my $yesterday = DateTime->now->subtract( days => 1 ); \
cmp_ok( DateTime->compare($dt, $yesterday),'==',0,"first day found is - $dt and this is a yesterday" );
You also may use "capture()" function to get a *first element* of
captures array:
# Place this in swat data file
# check if outpiut contains digits
# calculate total amount
# it should be > then 10
regexp: (\d+)
code: cmp_ok( capture()->[0],'>',10,"fisrt number is greater than 10" );
Anatomy of swat
Once swat runs it goes through some steps to get job done. Here is
description of such a steps executed in orders
Run iterator over swat data files
Swat iterator look for all files named get.txt or post.txt or put.txt
under project root directory. Actually this is simple bash find loop.
Parse swat data file
For every swat data file find by iterator parsing process starts. Swat
parse data file line by line, at the end of such a process *a list of
Test::More asserts* is generated. Finally asserts list and other input
parameters are serialized as Test::More test scenario written into into
proper *.t file.
Give it a run by prove
Once swat finish parsing all the swat data files there is a whole bunch
of *.t files kept under a designated temporary directory, thus every
swat route maps into Test::More test file with the list of asserts. Now
all is ready for prove run. Internally `prove -r ` command is issued to
run tests and generate TAP report. That is it.
Below is example how this looks like
project structure
$ tree examples/anatomy/
examples/anatomy/
|----FOO
|-----|----BARs
| |---- post.txt
|--- FOOs
|--- get.txt
3 directories, 2 files
swat data files
# /FOOs
FOO
FOO2
generator: | %w{ FOO3 FOO4 }|
# /FOO/BARs
BAR
BAR2
generator: | %w{ BAR3 BAR4 }|
code: skip('skip next 2 tests',2);
BAR5
BAR6
BAR7
Test::More Asserts list
# /FOOs/0.t
SKIP {
ok($status, "successful response from GET $host/FOOs")
ok($status, "GET /FOOs returns FOO")
ok($status, "GET /FOOs returns FOO2")
ok($status, "GET /FOOs returns FOO3")
ok($status, "GET /FOOs returns FOO4")
}
# /FOO/BARs0.t
SKIP {
ok($status, "successful response from POST $host/FOO/BARs")
ok($status, "POST /FOO/BARs returns BAR")
ok($status, "POST /FOO/BARs returns BAR")
ok($status, "POST /FOO/BARs returns BAR3")
ok($status, "POST /FOO/BARs returns BAR4")
skip('skip next 2 tests',2);
ok($status, "POST /FOO/BARs returns BAR5")
ok($status, "POST /FOO/BARs returns BAR6")
ok($status, "POST /FOO/BARs returns BAR7")
}
POST/PUT requests
Name swat data file as post.txt (put.txt) to make http POST (PUT)
requests.
echo 200 OK >> my-app/hello/post.txt
echo 200 OK >> my-app/hello/world/post.txt
You may use curl_params setting ( follow "Swat Settings" section for
details ) to define post data, there are some examples:
* "-d" - Post data sending by html form submit.
# Place this in swat.ini file or sets as env variable:
curl_params='-d name=daniel -d skill=lousy'
* "--data-binary" - Post data sending as is.
# Place this in swat.ini file or sets as env variable:
curl_params=`echo -E "--data-binary '{\"name\":\"alex\",\"last_name\":\"melezhik\"}'"`
curl_params="${curl_params} -H 'Content-Type: application/json'"
Dynamic routes
There are possibilities to create a undetermined routes using ":path"
placeholders. Let say we have application confirming GET /foo/:whatever
requests where :whatever is arbitrary sting like: GET /foo/one or
/foo/two or /foo/baz. Using dynamic routes we could write an swat test
for it.
First let's create definition for `whatever` path in swat.ini file. This
is as simple as create bash variable with a random sting value:
# Place this in swat.ini file
export whatever=`cat /dev/urandom | tr -dc 'a-zA-Z0-9' | fold -w 5 | head -n 1`
Now we should inform swat to use bash variable $whatever when generating
request for /foo/whatever
$ mkdir foo/:whatever
And finally drop some check expressions for it:
$ echo 'generator [ $ENV{"whatever"} ]' > foo/:whatever/get.txt
Of course there are as many dynamic parts in http requests as you need:
# Place this in swat.ini file
export whatever=`cat /dev/urandom | tr -dc 'a-zA-Z0-9' | fold -w 5 | head -n 1`
export whenever=`date +%s`
$ mkdir -p foo/:whatever/:whenever
$ echo 'generator [ $ENV{"whatever"}, $ENV{"whenever"} ]' > foo/:whatever/:whenever/get.txt
Swat Settings
Swat comes with settings defined in two contexts:
* Environment variables ( session settings )
* swat.ini files ( home directory , project based, route based and
custom settings )
Environment variables
Following variables define a proper swat settings.
* "debug" - set to "1,2" if you want to see some debug information in
output, default value is 0
* "debug_bytes" - number of bytes of http response to be dumped out
when debug is on. default value is 500
* "swat_debug" - run swat in debug mode, default value is 0
* "ignore_http_err" - ignore http errors, if this parameters is off
(set to 1) returned *error http codes* will not result in test
fails, useful when one need to test something with response differ
from 2**,3** http codes. Default value is 0
* "try_num" - number of http requests attempts before give it up (
useless for resources with slow response ), default value is 2
* "curl_params" - additional curl parameters being add to http
requests, default value is "", follow curl documentation for variety
of values for this
* "curl_connect_timeout" - follow curl documentation
* "curl_max_time" - follow curl documentation
* "port" - http port of tested host, default value is 80
* "prove_options" - prove options, default value is "-v"
Swat.ini files
Swat checks files named "swat.ini" in the following directories
* ~/swat.ini - home directory settings
* $project_root_directory/swat.ini - project based settings
* $route_directory/swat.ini - route based settings
* $cwd/swat.my - custom settings
Here are examples of locations of swat.ini files:
~/swat.ini # home directory settings
my-app/swat.ini # project based settings
my-app/hello/get.txt
my-app/hello/swat.ini # route based settings ( route hello )
my-app/hello/world/get.txt
my-app/hello/world/swat.ini # route based settings ( route hello/world )
Once file exists at any location swat simply bash sources it to apply
settings.
Thus swat.ini file should be bash file with swat variables definitions.
Here is example:
# the content of swat.ini file:
curl_params="-H 'Content-Type: text/html'"
debug=1
try_num=3
Settings priority table
This table describes all the settings with priority levels, the settings
with higher priority are applied after settings with lower priority.
| context | location | settings type | priority level |
| ------------------------|--------------------------- | -------------------- | --------------- |
| swat.ini file | ~/swat.ini | home directory | 1 |
| swat.ini file | project root directory | project based | 2 |
| swat.my file | current working directory | custom settings | 3 |
| swat.ini file | route directory | route based | 4 |
| environment variables | --- | session | 5 |
Settings merge algorithm
Thus swat applies settings in order for every route:
* Home directory settings are applied if exist.
* Project based settings are applied if exist.
* Custom settings are applied if exist.
* Route based settings are applied if exist.
* And finally environment settings are applied if exist.
Custom Settings
Custom settings are way to cutomize settings for existed swat package.
This file should be located at current working directory, where you run
swat from. For example:
# override http port
$ echo port=8080 > swat.my
$ swat swat::nginx 127.0.0.1
Follow section "Swat Packages" to get more about portable swat tests.
Hooks
Hooks are extension points you may imppliment to hack into swat complie
/ runtime workflow. There are two types of hooks:
* Perl hooks
* Bash Hooks
Perl hooks
Perl hooks are files with perl code `required` *in the beginning/end of
a swat test*. There are four types of perl hooks:
* project based perl startup hook
File located at "$project_root_directory/hook.pm".
Project based startup hooks are `required` *in the begining* of a
swat test and applied for every route in project and thus could be
used for *project initialization* procedures.
For example one could define common generators here:
# place this in hook.pm file:
sub list1 { | %w{ foo bar baz } | }
sub list2 { | %w{ red green blue } | }
# now we could use it in swat data file
generator: list()
generator: list2()
* project based perl cleanup hook
File located at "$project_root_directory/cleanup.pm".
This hooks is similar to startup hook but `required` *in the end* of
a swat test.
* route based perl startup hooks
Files located at "$route_directory/hook.pm".
Routes based startup hooks are applied for every route in project
and thus could be used for *route initialization* procedures.
For example one could define route specific generators here:
# place this in hook.pm file:
# notices that we could tell GET from POST http methods here
# using predefined $method variable
sub list1 {
my $list;
if ($method eq 'GET') {
$list = | %w{ GET_foo GET_bar GET_baz } |
}elsif($method eq 'POST'){
$list = | %w{ POST_foo POST_bar POST_baz } |
}else{
die "method $method is not supported"
}
$list;
}
# now we could use it in swat data file
generator: list()
* route based perl cleanup hooks
Files located at "$route_directory/cleanup.pm".
This hooks is similar to route based startup hooks but `required`
*in the end* of a swat test.
Bash hooks
Similar to perl hooks bash hooks are just a bash files `sourced` *before
compilation* of a swat test.
There are 4 types of bash hooks:
* project based bash hook
File located at "$project_root_directory/hook.bash".
Project based bash hooks are applied for every route in project and
could be used for *project initialization* procedures.
* route based bash hooks
Files located at
"$project_root_directory/$route_directory/hook.bash".
Routes based bash hooks are route specific hooks and could be used
for *route initialization* procedures.
* global startup bash hook
File located at "$project_root_directory/startup.bash".
Startup hook is executed before swat tests gets compiled, at the
very begining, at could be used for *global initialization*
procedures.
* global cleanup bash hook
File located at "$project_root_directory/cleanup.bash".
Cleanup hook is executed *after swat tests are executed*, at the
very end, and could be used for *global cleanup* procedures.
It is important to note that bash hooks are executed *after swat
settings merge done* , see "Swat Settings" section to get more about
swat settings.
Predifined variables
List of variables one may rely upon when writting perl/bash hooks:
* http_url
* curl_params
* http_meth - "GET|POST|HEAD"
* route_dir
* project
Swat Compile and Runtime
- Execute *global startup bash hook*
- Start of swat compilation phase
- For every route gets compiled:
-- Merge swat settings
-- Set predifined variables
-- Execute *project based bash hook*
-- Execute *route based bash hook*
-- Compile route test
- The end of swat compilation phase
- Start of swat executation phase.
- For every route test gets executed:
-- Execute *project based perl startup hook*
-- Execute *route based perl startup hook*
-- Execute route test
-- Execute *route based perl cleanup hook*
-- Execute *project based perl cleanup hook*
- The end of swat compilation phase
- Execute *global cleanup bash hook*
TAP
Swat produces output in TAP format , that
means you may use your favorite tap parsers to bring result to another
test / reporting systems, follow TAP documentation to get more on this.
Here is example for converting swat tests into JUNIT format
swat --formatter TAP::Formatter::JUnit
See also "Prove settings" section.
Command line tool
Swat is shipped as cpan package, once it's installed ( see "Install"
section ) you have a command line tool called swat, this is usage info
on it:
swat
* host - is base url for web application you run tests against, you
also have to define swat routes, see DSL section.
* project_dir - is a project root directory
* swat_package - the name of swat package, see "Swat Packages" section
Default Host
Sometimes it is helpful to not setup host as command line parameter but
define it at $project_root/host file. For example:
# let's create a default host for foo/bar project
$ cat foo/bar/host
foo.bar.com
$ swat foo/bar/ # will run tests for foo.bar.com
Prove settings
Swat utilize prove utility to run
tests, so all the swat options *are passed as is to prove utility*.
Follow prove utility
documentation for variety of values you may set here. Default value for
prove options is "-v". Here is another examples:
* "-q -s" - run tests in random and quite mode
Swat Packages
Swat packages is portable archives of swat tests. It's easy to create
your own swat packages and share with other.
This is mini how-to on creating swat packages:
Create swat package
Swat packages are *just cpan modules*. So all you need is to create cpan
module distribution archive and upload it to CPAN.
The only requirement for installer is that swat data files should be
installed into *cpan module directory* at the end of install process.
File::ShareDir::Install
allows
you to install read-only data files from a distribution and considered
as best practice for such a things.
Here is example of Makefile.PL for swat::mongodb package
:
use inc::Module::Install;
# Define metadata
name 'swat-mongodb';
all_from 'lib/swat/mongodb.pm';
# Specific dependencies
requires 'swat' => '0.1.28';
test_requires 'Test::More' => '0';
install_share 'module' => 'swat::mongodb', 'share';
license 'perl';
WriteAll;
Here we create a swat package swat::mongodb with swat data files kept in
the project_root directory ./share and get installed into
"auto/share/module/swat-mongodb" directory.
Once we uploaded a module to CPAN repository we can use it:
$ cpan install swat::mongodb
$ swat swat::mongodb 127.0.0.1:28017
Check out existed swat packages here -
https://github.com/melezhik/swat-packages/
Misc
Debugging
set "swat_debug" environment variable to 1
Running a single test
It is possible to run a single swat test setting a "test_file" variable:
"test_file"=/relative/path/to/swat/data/file . Path should be relative
to the project root directory.
Examples
./examples directory contains examples of swat tests for different
cases. Follow README.md files for details.
AUTHOR
Aleksei Melezhik
Home Page
https://github.com/melezhik/swat
Thanks
To the authors of ( see list ) without who swat would not appear to
light
* perl
* curl
* TAP
* Test::More
* prove
COPYRIGHT
Copyright 2015 Alexey Melezhik.
This program is free software; you can redistribute it and/or modify it
under the same terms as Perl itself.