Improve array handling in alias completion

[convergedtarkus]

Description

I was seeing issues where certain aliases do not complete as expected.

For example, my alias gbD='git branch -D' alias should provide completions to local branches only, and it does when run without an alias. However when running completion on the alias, I get all the branches which appears to be the git default completion response.

I tracked this down to an issue where an array was being incorrectly handled. The old handling caused this result, COMP_WORDS=[‘git’, ‘branch -D’, 'ma'] (length 3) where branch and -D were combined into a single entry. This produces the incorrect behavior. By changing the array to string handling to better maintain individual arguments the new result is COMP_WORDS=[‘git’, ‘branch’ ,‘-D’, 'ma'] (length 4) which correctly keeps branch and -D apart.

Motivation and Context

This improves alias completions to act more correctly.

How Has This Been Tested?

I have been unable to create a unit test that is able to call the alias completion function. I have tested this manually on my machine and it works as expected.

An easy way to manually test this is to add the alias biea='bash-it enable alias' and load bash-it with the bash-it completion enabled. Without my fix the complete for this alias are the base level bash-it commands such as doctor, version etc. With my change the output is the aliases that can be enabled as would be expected.

Screenshots (if appropriate):

N/A

Types of changes

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to change)

Checklist:

• My code follows the code style of this project.
• If my change requires a change to the documentation, I have updated the documentation accordingly.
• I have read the CONTRIBUTING document.
• If I have added a new file, I also added it to clean_files.txt and formatted it using lint_clean_files.sh.
• I have added tests to cover my changes, and all the new and existing tests pass.

[seefood]

is it necessary to go through this extra printf? I would think just switching [*] to [@] might have worked too? please try to do this a bit cleaner maybe, or at least make like 84 use a local var and calculate it down by the other line. the distance between the two doesn't make sense to me, unless I missed something.

[convergedtarkus]

@seefood The printf produces different output than using @ or * which is important to ensure the echo on line 85 produces the correct output string.

For example:

myArray=(one two three)
echo "* version: '${myArray[*]}'"
echo "@ version: '${myArray[@]}'"
printf "printf version: '"
printf "'%s' " "${myArray[@]}"
printf "'\n";

That outputs

* version: 'one two three'
@ version: 'one two three'
printf version:''one' 'two' 'three' '

Notice that the printf version put each array value in quotes. That is the important part, even more so if one of the array values has a space in it. In addition, the printf command needs to use "${myArray[@]}" not "${myArray[*]}" as the * version will also not quote individual elements of the array.

I can definitely look into cleaning up how this is done though!

[convergedtarkus]

The CI failure does not look like it is related to my changes. I'm seeing commits in master that have the same issues where some tests appears to not be running.
I'm not seeing an option to re-run the failing actions, maybe that's not supported here.

[convergedtarkus]

@seefood @akinomyoga Just wanted to bump on this for reviews.

[seefood]

looking good!