capitalone / bash_shell_mock Goto Github PK

@nitinsurana @jdorfman @particleflux @meustrus it seems I had the cart before the horse regarding CapitalOne's opensource contributor policy. All contributors are suppose to complete a CLM prior to me accepting the PR. Obviously that ship has sailed.

I have a PR to add this information within the repo for future contributors. The link is here https://developer.capitalone.com/open-source/

Would you visit the site and read the CLM and if you can agree to it? If so then please submit one. Much of the language is with regard to patent rights which I don't for see being an issue here. I have valued your inputs so far and hope to see your continued use and support. If you can't agree to it I certainly understand.

Hi,

Before I make a PR I was wondering if you would be opposed to having the official Bash logo in the readme?

Here are some examples: https://github.com/odb/official-bash-logo#sightings

Though there's fairly extensive testing of bashshellmock.sh, there is currently nothing running these in an automated fashion on our PRs. Given it may be more trouble than it's worth to integrate our enterprise Jenkins instances to interact with public GitHub, this issue entails evaluating which (hosted) continuous integration platform to use:

• Research other @capitalone repositories to see what continuous integration platform they are using
• Finalize which continuous integration platform to use, with approval from the open source office
• Created an automated script to execute and check the results of the sample-bats and test

The formatting of the readme is not right. The formatting of the table in the shellmock_expect has some sort of syntax issue.

Functionality similar to #19 that could not be merged due to a contributor that did not sign the CLA

I propose redirect stdout to stderr if status more than 0, in my humble opinion, this behavior make a sense

When adding some new example test cases:

@test "sample.sh-success-partial-mock" {

    shellmock_expect grep --status 0 --output "Mocked output for Partial Match" --type partial --match '"sample line"'

    run ./sample.sh

    shellmock_dump

    [ "$status" = "0" ]

    # Validate using lines array.
    [ "${lines[0]}" = "Mocked output for Partial Match" ]

    # Optionally since this is a single line you can use $output
    [ "$output" = "Mocked output for Partial Match" ]

}

This test failed to match even though sample.sh passes this double quoted string as one argument.

Please add these lines to the license headers in all the scripts.

SPDX-Copyright: Copyright (c) Capital One Services, LLC
SPDX-License-Identifier: Apache-2.0

node -e "require('./helper').validate()"

Above is how my function call looks. I can't seem to find a corresponding shellmock_expect --match argument. I've tried various combinations like :

shellmock_expect node --status 1 --match '-e "require('./helper').validate()"'
shellmock_expect node --status 1 --match '-e "require(\'./helper\').validate()"'
shellmock_expect node --status 1 --match '-e "require(\"./helper\").validate()"'

throws status 99

If I have the following functions file:
functions_file.sh:

  docker_user="$1"
  docker_password="$2"
  docker_registry="$3"
  docker login -p ${docker_password} -u ${docker_user} ${registry}
}

main_function(){
    helper_function user1 password1 nonexsistent.domain.com
}

And the bats file:

load 'libs/bats-assert/load'

setup() {
  . shellmock
  source functions_file.sh
  skipIfNot "$BATS_TEST_DESCRIPTION"
  shellmock_clean
}

teardown() {
  if [[ -z "$TEST_FUNCTION" ]]; then
    shellmock_clean
  fi
}

stub_functions() {
  shellmock_expect helper_function --match "user1 password1 registry.domain.com"
}

@test "main" {
  stub_functions
  run main_function
  assert_equal "$status" 0
}

I don't seem to be able to mock the function help_function

• Add a build.sh or some sort of script that runs tests and run shellcheck to find issues.
• Add documentation regarding build or pre-check in process.
• Update code based on recommendations from shellcheck as appropriate.

function toTest(){
 command -e --params
 echo $?   # here lies the bug, it returns 0, it should have returned 1 as per shellmock
}

Above is the function to test.

shellmock_expect command --status 1 --match "-e --params"

The easiest example of this is in the sample-bats directory, if you look at the "sample.sh failure" test, modifying the status of the grep command from 1 to 0 does not affect the outcome of the test. This is because shellmock cannot find the grep command to begin with. Easy way to see this is to reverse the if conditional on teardown to see the shellmock error log.

The current workaround that I've been using is to utilize the '--type partial' flag and not look for double quotes in the --match argument.

Hi there !

I'm programmatically doing something like echo 0 */2 * * * my_user my_script | crontab - to add an entry in the current user's crontab.

Can I check that the crontab really got 0 */2 * * * my_user my_script from STDIN with bash_shell_mock ?