Fix document type detection (limited-quirks mode)

[squidfunk]

This PR fixes a bug with DOCTYPE detection, where the lowercased publicId will never match the limited-quirks document types with mixed case.

[inikulin]

@squidfunk Good catch, thanks! Though, seems like linting in CI is unhappy with line endings.

[squidfunk]

Fixed, sorry for that.

[squidfunk]

BTW, I don't know if the spec changed since you implemented quirks-mode detection, but I think the prefixes are not entirely correct as they specify //en at the end. See this section of the spec which doesn't specify them. Also my testing with document.compatMode denotes that the language must not necessarily be en for correct quirks-mode detection.

[inikulin]

@squidfunk Thanks for yet another great catch. I've just looked through the spec commit history and it seems like it always was this way and it's a mistake on my side. I'm surprised it wasn't captured by html5lib-tests.

May I ask you to fix that in context of this PR as well and open a ticket in https://github.com/html5lib/html5lib-tests to add test cases for that?

[squidfunk]

May I ask you to fix that in context of this PR as well and open a ticket in https://github.com/html5lib/html5lib-tests to add test cases for that?

Will do.

[squidfunk]

I could combine some of the prefixes and add two missing - should now be spec-compliant.

[squidfunk]

It would be really awesome if you could issue a bugfix release, as my downstream project pretty much depends on this fix :-)

[inikulin]

@squidfunk Great work, thank you!

It would be really awesome if you could issue a bugfix release, as my downstream project pretty much depends on this fix :-)

Will do =)

[inikulin]

@squidfunk Published as v5.1.1

[squidfunk]

Perfect, thanks! Happy to contribute.