Fix parsing of some purl types. Fix tests.

[Octogonapus]

This PR adds the latest test data from master into testdata/ instead of keeping an old version around.
This PR also stops overwriting that test data during test execution because that is confusing when the test data is already in the source tree.

Fixes I did while reading the purl spec:

• Fix hugging face version parsing
• Fix mlflow name adjustments
• Fix compose namespace and name parsing
• Fix qualifier order in mlflow test

Closes #38.

[sschuberth]

Just a general remark from someone who's not into Go, but trying to help out with a review to get things rolling again: What's making the review hard for me is that most of the commit messages do not provide a rationale why you're doing what you're doing, but you just repeat in prose what's more or less obvious from the diff.

[sschuberth]

For consistency with other test data in this file, wouldn't it make more sense to lower-case the SHA1 in the purl in line 496 instead?

[sschuberth]

Ok, probably not that important as you're changing it back in a later commit anyway.

[Octogonapus]

Yes I changed it back because that was a mistake as a huggingface version must be lowercase https://github.com/package-url/purl-spec/blob/master/PURL-TYPES.rst#huggingface

[sschuberth]

I'd appreciate a review from @shibumi or @pombredanne.

[Octogonapus]

What's making the review hard for me is that most of the commit messages do not provide a rationale why you're doing what you're doing, but you just repeat in prose what's more or less obvious from the diff.

Sorry, when I wrote that in the commit messages, I meant that the logic must be that way because of what the spec states.
Hopefully a better explanation:

• huggingface versions must be lowercase https://github.com/package-url/purl-spec/blob/master/PURL-TYPES.rst#huggingface
• the case of mlflow names depends on their registry https://github.com/package-url/purl-spec/blob/master/PURL-TYPES.rst#mlflow
• composer namespaces and names must be lowercase https://github.com/package-url/purl-spec/blob/master/PURL-TYPES.rst#composer
• the order of qualifiers was adjusted in the test to conform to the example in https://github.com/package-url/purl-spec/blob/master/PURL-TYPES.rst#mlflow

[sschuberth]

Hopefully a better explanation:

Thanks! Would you mind amending the commit messages with these, and force-pushing the branch for this PR again?

[Octogonapus]

Done

[sschuberth]

Done

Thanks again. I'm not sure about the rules in this repo, but usually the Git commit message title is limited to something like 72 chars (and shouldn't contain a trailing dot; think of it like the title for a book), and the "real" explanation should go as a separate paragraph in the body of the commit message.

Anyway, if there's no negative response from @shibumi or @pombredanne by today, I plan to merge this to move things forward.

[sschuberth]

Anyway, if there's no negative response from @shibumi or @pombredanne by today, I plan to merge this to move things forward.

Ok, let's just merge this to move things forward.