putting tests at next level

[franiglesias]

Yesterday I was thinking in a sort of set up (prepare scenario) and tear down (remove scenario) for test cases, and just see pr #4 where you implement in part this idea.

In fact, what I am thinking of is a sort of "callback" in the testAll.sh script that looks for and execute a global set up (and tear down after all tests are finished) and even one test case based.

With this you can isolate the scenario creation from the test itself so they will become easier to understand.

[pkristian]

hm, could you visualize schema you have in mind? pseudocode will be anough. because i am afraid i do not fully understand how it should look like.
maybe #4 is place where it could fall into?

[franiglesias]

testAll.sh is sort of a generic test runner. Sometimes you need to create a scenario for the test to run (in this case you need to create repositories, directories, etc). Sometimes this scenario is common for all the tests, and sometimes is different.

The idea is something like:

#!/usr/bin/env bash

set -e

cd "$(dirname "$0")"

if [ -e "./setUp.sh" ];then
    bash "./setUp.sh"
fi

countTests=0
countFailed=0

for testCase in $(ls -d */)
do
{
    countTests=$((countTests + 1))

    expected="$(cat $testCase/expected)"

    # execute if present bash $testCase/setUp.sh
    actual="$(bash $testCase/command.sh)"
    # execute if present bash $testCase/tearDown.sh

   # all other stuff 
}
done

if [ -e "./tearDown.sh" ];then
    bash "./tearDown.sh"
fi

# all other stuff

The setUp.sh scripts are optional and used to prepare scenarios
The tearDown.sh scripts are optional and used to remove scenarios, so you leave the system clean

[pkristian]

oh, thank you! i like this, so i should create TestAll.sh make more generic, aye? copy :)

[pkristian]

extended duplicate of #4

[pkristian]

i decidet this should be sole PR... :) thx @franiglesias ! :)

[pkristian]

• setUp and TearDown
• comprimed test remote
• failing test diff