What about if ! check_for_equality "$1" "$2"; then ?

That is shorter.
I am not a bash guru 😁 but the first is a bit more clear to me that we are checking the return value of the function as the function itself is returning the full stdout of the compare - which is the reason the if failed to begin with as the comparison wasn't comparing the things that were expected.
Result-wise they are the same. Having said that: I did make a mistake when adding and pushing the check_return 😊 It should read check_return=$?. So this needs an update, either way.

Do not mention bash :)
This should not be a bash-script but an sh POSIX-compatible script.
Well, it is shorter and it is what b2dc4ac does, which will eventually get merged.
So I think we should use that.

I propose that your PR be merged for 2.10.3 and it will be overwritten by b2dc4ac when merging for 2.11.0

You make a lot of ground 😁
It seems we are patching stuff that is already patched. Ah well... I fixed my my fix.

Not directly related but with testing it bites me that the version has not been updated (e.g. testing with the incorrect executable). Mind if we bump the version to 2.10.3 so it becomes clear what we are working on?
Can make a separate PR for that for clarity, obviously.

It seems we are patching stuff that is already patched.

Yes, the idea was to use specifically if ! check_for_equality "$1" "$2"; then in your PR so that line will not conflict when merging later.

Not directly related but with testing it bites me that the version has not been updated (e.g. testing with the incorrect executable).

Do you get a warning of some sort ?

line will not conflict when merging later.

Ah.. that makes sense 👍

Do you get a warning of some sort ?

Nothing to worry about, just me not working with the exec I just build (and therefore getting confusing results)

Nothing to worry about, just me not working with the exec I just build (and therefore getting confusing results)

When you run RM_TS_PRINT_CMD=1 RM_TS_KEEP_TESTDIR=1 nosetests -s tests/test_formatters/test_sh.py to test this PR, it compiles and run rmlint from the repository, not from the system path

😁 I was not running the tests but trying to replay an issue report with something I modified to see whether my idea was sound. which told me enough but I was still second-guessing myself. Either way... having a new version after a release is always sound in my book. The current develop/master issue is adding some confusion too, which makes me a bit more cautious than if I were free hacking away on develop. Thanks for #694