从bash脚本一个接一个地执行Java Cucumber功能文件，并在第一个功能完成后继续执行功能文件

Question

I have a need to run a bunch of Java cucumber feature files, one by one, by using a bash script on MacOS High Sierra. The problem I have is, as soon as the first feature file is done executing (either success or failure), the Cucumber framework does its reporting on that feature file and my script does not execute the remaining feature files. It "returns" to the command line.

Here is part of the script that runs the cucumber files one by one:

function run_cucumber_features {
  for file in $(find ./src/test  -name '*Steps.feature' -print0 | xargs -0);
    do gradle cucumberTest -Dcucumber.options="$file" DmutationTesting="false" -DmutationParameter="$parameter";
  done
}

For example, let's say I execute my script:

./mutation_testing.sh 

The script then invokes Cucumber and the first feature the script finds execute, and then it 'exits' with the status code for that specific feature:

BUILD Success in 18s
7 actionable tasks: 5 executed, 5 succeeded

However, it then does not proceed to execute the next feature file, it returns to the command line:

username$ 

My question is whether there is a way for my script to not stop executing once Cucumber has done its reporting after every feature file that has been executed?

Answer 1

The error seems to be related to the steps inside your build.gradle file or the way you call the function in the Bash script.

A small test set-up. Assume the following structure#!/bin/bash

run-gradle-loop.sh                    # executable script
gradle                                # executable script, as simple replacement
src/test/subdir/file2_Steps.feature   # a dummy file
src/test/file1_Steps.feature          # a dummy file

run-gradle-loop.sh

#!/bin/bash

function run_cucumber_features {
  for file in $(find ./src/test  -name '*Steps.feature' -print0 | xargs -0);
  do
    gradle cucumberTest -Dcucumber.options="$file" DmutationTesting="false" -DmutationParameter="$parameter";
    echo "return code = $?"
  done
}

# to run the gradle script in the current directory
export PATH=$(pwd):${PATH}

run_cucumber_features

gradle

#!/bin/bash

echo
echo "running gradle dummy"
echo "passed parameters : $@"

# exit with an error code
exit 1

file1_Steps.feature and file2_Steps.feature are two empty files, only to fulfill the loop criteria.

The aim of the test is to check the behavior where gradle exit with an error code.

run the main Bash script

./run-gradle-loop.sh

output

running gradle dummy
passed parameters : cucumberTest -Dcucumber.options=./src/test/subdir/file2_Steps.feature DmutationTesting=false -DmutationParameter=
return code = 1

running gradle dummy
passed parameters : cucumberTest -Dcucumber.options=./src/test/file1_Steps.feature DmutationTesting=false -DmutationParameter=
return code = 1

This shows that even gradle would return with an error code the loop processes all .*Steps.feature files.

edit One way it might stop after the first error could be that in the run-gradle-loop.sh script there is a set -e statement.

man bash

-e Exit immediately if a pipeline (which may consist of a single simple command), a list, or a compound command

Which means. When the script gradle fails with an error the control is not returned to run-gradle-loop.sh . The output with above set-up would be

$ ./run-gradle-loop.sh

running gradle dummy
passed parameters : cucumberTest -Dcucumber.options=./src/test/subdir/file2_Steps.feature DmutationTesting=false -DmutationParameter=