Xiangce Liu

@lveyde - As you might notice, there is an existing parser [YumRepoList](https://github.com/RedHatInsights/insights-core/blob/900148134a25d4bbaed17455dd63171667dcc64d/insights/parsers/yum.py#L44) which parses the output of `yum -C --noplugins repolist` command and can achieve the same function with this...

@bfahr - Thanks for reviewing. Due to the current `filter` policy, the original content of the spec will be filtered per the `allowlist` by force when the spec is marked...

@bfahr - I added some notes for the two specs in the docstring, please have a look.

@bfahr - This change looks good to me, please help double confirm.

> The proposed impl would require users to test every value to see whether it's a list or not. I prefer to keep the existing impl and handle the duplicate...

> @csams if all values should be lists for consistency: > > ``` > >>> split_kv_pairs(lines, allow_duplicates=True) > {'keyword2': ['value2a=True, value2b=100M'], 'keyword1': ['value1', 'value2']} > ``` > > The the...

@shlao - Yeap, that's the impact of this update (applying the "allow_duplicates=Ture" to this parser). It's reasonable for me to make this parser more flexible to support these `duplicated keys`...

Since @vishwanathjadhav is in the DevConf@Bangalore, I just updated this PR as per Bob 's feedback. @bfahr please have a check again when free.

@csams, Thanks. It makes sense. I committed a new #2476 as per your suggestion. Please have a look when free.

@csams - Although it works on my test system, it seems the resolution of #2476 doesn't work well on some CU's systems, e.g. there are more than one insights-archives collected...