Separation of concerns

[rogard]

Hello,

The following use-case, which is based on the documentation, breaks separation of concern due to the present of source:

$ cat use-case.sh
# shebang
source '/path/to/bach.sh'
 
test-rm-rf() {
    # Write your test case
     
    project_log_path=/tmp/project/logs
    rm -rf "$project_log_ptah/" # Typo here!
}
test-rm-rf-assert() {
    # Verify your test case
    rm -rf /   # This is the actual command to run on your host!
               # DO NOT PANIC! By using Bach Testing Framework it won't actually run.
}

$ source use-case.sh

I tried this:

# bach.sh:
- # shebang
- trap .bach.on-exit EXIT
+ # trap .bach.on-exit EXIT
+ function bach-run-test() {
+    local test_file
+
+    test_file="${1:?Missing test file argument}"
+
+    source "$test_file"
+
+    .bach.run-tests
+ }

Such that we can now call:

 $ source bach.sh
 $ back-run-test use-case.sh

It appears to work, except the shell exits, but I haven't found the fix.

[chaifeng]

Hi @rogard

You can just run the script bash use-case.sh. There is no need to use source or any Bach internal functions.

[chaifeng]

Fix: 1d6fc94#diff-b335630551682c19a781afebcf4d07bf978fb1f8ac04c6bf87428ed5106870f5

[rogard]

Thanks for the reply, but I think we may be talking past each other.

In the current setup, test files (use-case.sh) are required to contain source /path/to/bach.sh. This breaks the separation of concerns principle. Ideally, test files should contain only test logic—not framework bootstrapping. To address this, I suggested modifying bach.sh to expose a function, say bach-run-test, that takes a test file as an argument:

source /path/to/bach.sh
bach-run-test use-case-1.sh
bach-run-test use-case-2.sh
...
bach-run-test use-case-N.sh

Alternatively, bach.sh could be made executable and accept a test file as argument:

./path/to/bach.sh use-case-1.sh
./path/to/bach.sh use-case-2.sh
...
./path/to/bach.sh use-case-N.sh