bats-file
is a helper library providing common filesystem related
assertions and helpers for Bats.
Assertions are functions that perform a test and output relevant
information on failure to help debugging. They return 1 on failure and 0
otherwise. Output, formatted for readability, is
sent to the standard error to make assertions usable outside of @test
blocks too.
Features:
Dependencies:
bats-support
- output formatting, function call restriction
See the shared documentation to learn how to install and load this library.
Fail if the given file or directory does not exist.
@test 'assert_exists()' {
assert_exists /path/to/non-existent-file-or-dir
}
On failure, the path is displayed.
-- file or directory does not exist --
path : /path/to/non-existent-file-or-dir
--
Fail if the given file or directory does exist.
@test 'assert_not_exists()' {
assert_not_exists /path/to/existent-file-or-dir
}
On failure, the path is displayed.
-- file or directory exists, but it was expected to be absent --
path : /path/to/existent-file-or-dir
--
Fail if the given file does not exist.
@test 'assert_file_exists()' {
assert_file_exists /path/to/non-existent-file
}
On failure, the path is displayed.
-- file does not exist --
path : /path/to/non-existent-file
--
Fail if the given file or directory exists.
@test 'assert_file_not_exists() {
assert_file_not_exists /path/to/existing-file
}
On failure, the path is displayed.
-- file or directory exists, but it was expected to be absent --
path : /path/to/existing-file
--
Fail if the given directory does not exist.
@test 'assert_dir_exists()' {
assert_dir_exists /path/to/non-existent-directory
}
On failure, the path is displayed.
-- directory does not exist --
path : /path/to/non-existent-directory
--
Fail if the given directory exists.
@test 'assert_dir_not_exists() {
assert_dir_not_exists /path/to/existing-directory
}
On failure, the path is displayed.
-- directory exists, but it was expected to be absent --
path : /path/to/existing-directory
--
Fail if the given symbolic link does not exist.
@test 'assert_link_exists()' {
assert_file_exists /path/to/non-existent-link-file
}
On failure, the path is displayed.
-- symbolic link does not exist --
path : /path/to/non-existent-link-file
--
Fail if the given symbolic link exists.
@test 'assert_link_not_exists() {
assert_file_not_exists /path/to/existing-link-file
}
On failure, the path is displayed.
-- symbolic link exists, but it was expected to be absent --
path : /path/to/existing-link-file
--
Fail if the given block special file does not exist.
@test 'assert_block_exists()' {
assert_file_exists /path/to/non-existent-block-file
}
On failure, the path is displayed.
-- block special file does not exist --
path : /path/to/non-existent-block-file
--
Fail if the given block special file exists.
@test 'assert_block_not_exists() {
assert_file_not_exists /path/to/existing-block-file
}
On failure, the path is displayed.
-- block special file exists, but it was expected to be absent --
path : /path/to/existing-block-file
--
Fail if the given character special file does not exist.
@test 'assert_character_exists()' {
assert_file_exists /path/to/non-existent-character-file
}
On failure, the path is displayed.
-- character special file does not exist --
path : /path/to/non-existent-character-file
--
Fail if the given character special file exists.
@test 'assert_character_not_exists() {
assert_file_not_exists /path/to/existing-character-file
}
On failure, the path is displayed.
-- character special file exists, but it was expected to be absent --
path : /path/to/existing-character-file
--
Fail if the given socket does not exist.
@test 'assert_socket_exists()' {
assert_file_exists /path/to/non-existent-socket
}
On failure, the path is displayed.
-- socket does not exist --
path : /path/to/non-existent-socket
--
Fail if the given socket exists.
@test 'assert_socket_not_exists() {
assert_file_not_exists /path/to/existing-socket
}
On failure, the path is displayed.
-- socket exists, but it was expected to be absent --
path : /path/to/existing-socket
--
Fail if the given named pipe does not exist.
@test 'assert_fifo_exists()' {
assert_file_exists /path/to/non-existent-fifo-file
}
On failure, the path is displayed.
-- fifo does not exist --
path : /path/to/non-existent-fifo-file
--
Fail if the given named pipe exists.
@test 'assert_fifo_not_exists()' {
assert_file_not_exists /path/to/existing-fifo-file
}
On failure, the path is displayed.
-- named pipe exists, but it was expected to be absent --
path : /path/to/existing-fifo-file
--
Fail if the given file is not executable.
@test 'assert_file_executable()' {
assert_file_executable /path/to/executable-file
}
On failure, the path is displayed.
-- file is not executable --
path : /path/to/executable-file
--
Fail if the given file is executable.
@test 'assert_file_not_executable()' {
assert_file_not_executable /path/to/executable-file
}
On failure, the path is displayed.
-- file is executable, but it was expected to be not executable --
path : /path/to/executable-file
--
Fail if given user is not the owner of the given file.
@test 'assert_file_owner() {
assert_file_owner $owner /path/to/owner
}
On failure, the path is displayed.
-- user is not the owner of the file --
path : /path/to/notowner
owner : user
--
Fail if given user is the owner of the given file.
@test 'assert_not_file_owner() {
assert_not_file_owner $owner /path/to/notowner
}
On failure, the path is displayed.
-- Fail if given user is the owner, but it was expected not to be --
path : /path/to/owner
owner : $owner
--
Fail if given file does not have given permission.
@test 'assert_file_permission() {
assert_file_permission $permission /path/to/permission
}
On failure, the path is displayed.
-- given file does not have permissions 777 --
path : /path/to/nopermission
permission: $permission
--
Fail if given file has given permission.
@test 'assert_not_file_permission() {
assert_not_file_permission $permission /path/to/nopermission
}
On failure, the path is displayed.
-- given file has permissions 777, but it was expected not to have --
path : /path/to/permission
permission : $permission
--
Fail if the given file size does not match the input.
@test 'assert_file_size_equals() {
assert_file_size_equals /path/to/non-empty-file bytecount
}
On failure, the path and expected bytecount are displayed.
Fail if file is not zero byte.
@test 'assert_size_zero() {
assert_size_zero /path/to/zerobyte
}
On failure, the path is displayed.
-- file is not zero byte --
path : /path/to/notzerobyte
--
Fail if file size is zero byte.
@test 'assert_size_not_zero() {
assert_size_not_zero /path/to/notzerobyte
}
On failure, the path is displayed.
-- file is 0 byte, but it was expected not to be --
path : /path/to/zerobyte
--
Fail if group id is not set.
@test 'assert_file_group_id_set() {
assert_file_group_id_set /path/to/groupidset
}
On failure, the path is displayed.
-- group id is not set --
path : /path/to/groupidnotset
--
Fail if group id is set.
@test 'assert_file_not_group_id_set() {
assert_file_not_group_id_set /path/to/groupidnotset
}
On failure, the path is displayed.
-- group id is set, but it was expected not to be --
path : /path/to/groupdidset
--
Fail if user id is not set.
@test 'assert_file_user_id_set() {
assert_file_user_id_set /path/to/useridset
}
On failure, the path is displayed.
-- user id is not set --
path : /path/to/useridnotset
--
Fail if user id is set.
@test 'assert_file_not_user_id_set() {
assert_file_not_user_id_set /path/to/groupidnotset
}
On failure, the path is displayed.
-- user id is set, but it was expected not to be --
path : /path/to/userdidset
--
Fail if stickybit is not set.
@test 'assert_sticky_bit() {
assert_sticky_bit /path/to/stickybit
}
On failure, the path is displayed.
-- stickybit is not set --
path : /path/to/notstickybit
--
Fail if stickybit is set.
@test 'assert_not_sticky_bit() {
assert_not_sticky_bit /path/to/notstickybit
}
On failure, the path is displayed.
-- stickybit is set, but it was expected not to be --
path : /path/to/stickybit
--
Fail if the given file or directory is not empty.
@test 'assert_file_empty()' {
assert_file_empty /path/to/empty-file
}
On failure, the path and the content of the file is displayed.
-- file is not empty --
path : /path/to/empty-file
output (2 lines) : content-line-1
content-line-2
--
Fail if the given file or directory empty.
@test 'assert_file_not_empty() {
assert_file_not_empty /path/to/non-empty-file
}
On failure, the path is displayed.
-- file empty, but it was expected to contain something --
path : /path/to/non-empty-file
--
Fail if the given file does not contain the regex.
@test 'assert_file_contains() {
assert_file_contains /path/to/non-empty-file regex engine
}
engine
is optional and can be one of grep
, egrep
or pcregrep
. The specified engine must be available on the system running the tests.
On failure, the path and expected regex are displayed.
Fail if the given file contains the regex or if the file does not exist.
@test 'assert_file_not_contains() {
assert_file_not_contains /path/to/non-empty-file regex
}
On failure, the path and regex are displayed.
Fail if the given file is not a symbolic to a defined target.
@test 'assert_symlink_to() {
assert_symlink_to /path/to/source-file /path/to/symlink
}
On failure, the path is displayed.
-- symbolic link does not have the correct target --
path : /path/to/symlink
--
Fail if the given file is a symbolic to a defined target.
@test 'assert_not_symlink_to() {
assert_not_symlink_to /path/to/source-file /path/to/symlink
}
On failure, the path is displayed.
-- file is a symbolic link --
path : /path/to/symlink
--
-- symbolic link does have the correct target --
path : /path/to/symlink
--
When testing code that manipulates the filesystem, it is good practice to run tests in clean, throw-away environments to ensure correctness and reproducibility. Therefore, this library includes convenient functions to create and destroy temporary directories.
Create a temporary directory for the current test in BATS_TMPDIR
. The
directory is guaranteed to be unique and its name is derived from the
test's filename and number for easy identification.
<test-filename>-<test-number>-<random-string>
This information is only available in setup
, @test
and teardown
,
thus the function must be called from one of these locations.
The path of the directory is displayed on the standard output and is meant to be captured into a variable.
setup() {
TEST_TEMP_DIR="$(temp_make)"
}
For example, for the first test in sample.bats
, this snippet creates a
directory named sample.bats-1-XXXXXXXXXX
, where each trailing X
is a
random alphanumeric character.
If the directory cannot be created, the function fails and displays an error message on the standard error.
-- ERROR: temp_make --
mktemp: failed to create directory via template ‘/etc/samle.bats-1-XXXXXXXXXX’: Permission denied
--
The directory name can be prefixed with an arbitrary string using the --prefix <prefix>
option (-p <prefix>
for short).
setup() {
TEST_TEMP_DIR="$(temp_make --prefix 'myapp-')"
}
Following the previous example, this will create a directory named
myapp-sample.bats-1-XXXXXXXXXX
. This can be used to group temporary
directories.
Generally speaking, the directory name is of the following form.
<prefix><test-filename>-<test-number>-<random-string>
Delete a temporary directory, typically created with temp_make
.
teardown() {
temp_del "$TEST_TEMP_DIR"
}
If the directory cannot be deleted, the function fails and displays an error message on the standard error.
-- ERROR: temp_del --
rm: cannot remove '/etc/samle.bats-1-04RUVmBP7x': No such file or directory
--
Note: Actually, this function can be used to delete any file or
directory. However, it is most useful in deleting temporary directories
created with temp_make
, hence the naming.
During development, it is useful to peak into temporary directories post-mortem to see what the tested code has done.
When BATSLIB_TEMP_PRESERVE
is set to 1, the function succeeds but the
directory is not deleted.
$ BATSLIB_TEMP_PRESERVE=1 bats sample.bats
During debugging, it is useful to preserve the temporary directories of failing tests.
When BATSLIB_TEMP_PRESERVE_ON_FAILURE
is set to 1, the function
succeeds but the directory is not deleted if the test has failed.
$ BATSLIB_TEMP_PRESERVE_ON_FAILURE=1 bats sample.bats
The outcome of a test is only known in teardown
, therefore this
feature can be used only when temp_del
is called from that location.
Otherwise and error is displayed on the standard error.
Sometimes paths can be long and tiresome to parse to the human eye. To help focus on the interesting bits, all functions support hiding part of the displayed paths by replacing it with an arbitrary string.
A single pattern substitution is performed on the path before displaying it.
${path/$BATSLIB_FILE_PATH_REM/$BATSLIB_FILE_PATH_ADD}
The longest match of the pattern BATSLIB_FILE_PATH_REM
is replaced
with BATSLIB_FILE_PATH_ADD
. To anchor the pattern to the beginning or
the end, prepend #
or %
, respectively.
For example, the following example hides the path of the temporary directory where the test takes place.
setup {
TEST_TEMP_DIR="$(temp_make)"
BATSLIB_FILE_PATH_REM="#${TEST_TEMP_DIR}"
BATSLIB_FILE_PATH_ADD='<temp>'
}
@test 'assert_file_exists()' {
assert_file_exists "${TEST_TEMP_DIR}/path/to/non-existent-file"
}
teardown() {
temp_del "$TEST_TEMP_DIR"
}
On failure, only the relevant part of the path is shown.
-- file does not exist --
path : <temp>/path/to/non-existent-file
--
No one would want to develop piece of bash dependant libraries on their laptops due to single mistake (globbing for instance) can cause a disaster. In order to prevent this there is a Vagrantfile that you can use.
In order to start development environment, you have to take two steps;
user@localhost:~/bats-file$ vagrant up
The line above spins up a brand new virtualbox image and provisions with prerequisites.
However, as the tests require not to be on a network share due to running commands eg: mknod
, the files are shared into the VM by rsync
module. Rsync in vagrant only runs initialy and stops. During the active development, you regularly change files and might want to see the impact. To achive that, you have to use auto rsync.
auto-rsync
is a long running command. It means that it has to run on terminal screen as long as the VM is up and running. So, you have to keep this in a dedicated terminal screen.
After bringing up the VM, you can simply run following command;
user@localhost:~/bats-file$ vagrant rsync-auto
WARNING! WARNING! WARNING!: There will be a small delay between file save and sync. When you save a file please keep eye on the terminal window/pane and sync is triggered and finished. Simply run your next test with 5s-10s delay.
The repo files can be found under /home/vagrant/bats-file
and you can run tests with
user@localhost:~/bats-file$ bats test
# or
user@localhost:~/bats-file$ bats test/on-particular-test.bats
Once you're done with development, you can simply turn the VM off with
user@localhost:~/bats-file$ vagrant halt
- Why Vagrant, not Docker
- This replicates the 100% CI
- Why EoL Xenial image
- This replicates the 100% CI. Also, one step at a time. This wasn't in place at all. However, this change will bring up other concerns. So, that problems should be solved at another time.