Fixed tags parsing for metadata

[Splamy]

Summary of the changes (in less than 80 chars)

• Fix parsing tags from feeds which had them in a single space-separated string
• Fix parsing tags from nuspec from [ "tag1 tag2" ] to ["tag1", "tag2"]

Addresses https://github.com/bagetter/BaGetter/issues/89

Annotations

It felt weird in the unit test to check the deserialized value of the tags to be [ "json bson serializer" ], but I don't think tag normalizing should happen in the JsonConveter. The tags will be properly split up in the V3UpstreamClient.ToPackage method.

I had to add .Replace("\r\n", "\n") in the unit test json writer otherwise all related tests would fail. It seems that the strings in the tests are declared with \n even though the .editorconfig defines \r\n as the project line endings. The github actions will probably be default run linux, but i'm on a windows machine so it apparently wrote the json with crlf. You could maybe consider checking json for structural equality instead of string-wise like this https://github.com/Splamy/JsonBinMin/blob/main/JsonBinMin.Tests/AssertUtil.cs#L17-L57 or if you find a good library.

[FroggieFrog]

Is this the same as PR #85 and PR #90 ?

[Splamy]

Huh, yeah, seems like you were faster than me 😅

The only thing I have additionally is the NormalizeTags method, since otherwise you would return [ "tag1 tag2" ] from an upstream mirror like github Feel free to close this if it conflicts and just pick the snippet you need.

[FroggieFrog]

Update: Everything should be included in PR #90.

I will merge your additions/extensions (incl. tests) into PR #90 . Thank you!

[Splamy]

Thanks! And sorry for the duplication trouble, shoud've checked the open pr's, I really just didn't except you to tackle my issue so quickly 😉

[seriouz]

@FroggieFrog @Splamy Thank you, you two :)