Fix IndexError when empty string for credential.helper

[SID262000]

This PR addresses #1858

When credential.helper= we need to just strip off any whitespaces and return the empty string, as compared to the other scenario credential.helper=osykeychain or similar values where we return the first value only considering the fact that credential.helper is multivalued.

[Wauplin]

Thanks for opening a PR @SID262000! I reviewed and left some comments. Two main things:

• maybe a regex would be easier here to handle all tricky possibilities
• we should still return multiple values => it is possible to configure multiple git credential helpers on a machine

[Wauplin]

We must return a set of all git credential helpers configured on the machine. So here we would need to return a value for each line with "credential.helper=" in it.

[SID262000]

I see, thanks for pointing it out.

[Wauplin]

Here I think the parsing will break if line = credential.helper= since line.split("=")[-1] will evaluate to True but line.split("=")[-1].split()[0] will fail.

Would you like to try to fix this once and for all using a regex? I started something here that in theory should work for all the cases we want to handle (still has to be tested for real).

[SID262000]

Sure, I'll re-evaluate the parsing.

Regarding the usage of Regex,
It sounds more robust approach, many thanks for putting down the sample regex and sharing it with me. I'll test it for other potential use cases and incorporate the same in my ongoing PR

[SID262000]

I modified the above Regex slightly from "^\s*credential\.helper\s*=\s*(\w+)\s*" to "^\s*credential\.helper\s*=\s*(\w*)\s*$" to capture empty strings as well.
Now the above change captures all the values for credential.helper= and fixes it all once and for all using Regex.

[HuggingFaceDocBuilderDev]

The documentation is not available anymore as the PR was closed or merged.

[Wauplin]

Using re.findall is good. However, you don't need to run it on each line one by one. You can run on the whole output at once and use re.MULTILINE as flag. Here is how it looks like when you apply it only once:

Suggested change

	for line in output.split("\n"):
	l_cred_helper = sorted( # Sort for nice printing
	# Might have some duplicates
	re.findall(r"^\s*credential\.helper\s*=\s*(\w*)\s*$", line)
	)
	return l_cred_helper
	return sorted( # Sort for nice printing
	set( # Might have some duplicates
	re.findall(
	r"^\s*credential\.helper\s*=\s*(\w+)\s*$",
	output,
	flags=re.MULTILINE | re.IGNORECASE,
	)
	)
	)

[SID262000]

Thanks @Wauplin for the feedback. I've incorporated the above changes as indicated.

[Wauplin]

Thanks for making the changes @SID262000! It's good to merge 🔥