[epub-rs-34] missing definition on DOMString parameter matching

[himorin]

Section

Appendix B. epubReadingSystem object

Describe the problem

epubReadingSystem object defines its interface as

boolean hasFeature(DOMString feature, optional DOMString version);

No description provided in appendix B. 3.1 for matching DOMString feature nor DOMString version, former might be ascii case insensitive match and could have restriction to ASCII characters as similar to other enum-ish specifications, and latter would be comparison after numerical (incl. decimal into floating point) converted.

Describe the fix or new feature you propose

Add clear definition of string matching, using something like is (exact / case sensitive match) or ASCII case-insensitive match from INFRA spec.

[mattgarrish]

It used to be that matching was case insensitive when RSO was first introduced in 3.0.1:

Feature names are case-insensitive.

But then that text was removed in 3.1/3.2 when the object underwent a bunch of revisions. I can't find any discussion in those about changing to case-sensitive matching, so my guess is the text got lost in one of the edits.

I'm fine either way as the RSO is largely under-implemented anyway; it's always bordered on outdated/deprecated status. The authoring spec barely even acknowledges it anymore, with only one reference to it in the security and privacy section and an unmentioned and unneeded example of it in an appendix example on scripted contexts.

[iherman]

Because, in theory, this interface can be used by a programming language where case-sensitive is the norm, let us say that it is case sensitive. We already refer to infra normatively, so referring to https://infra.spec.whatwg.org/#string-is is ok to me...

But, as @mattgarrish says, it is, sort of, a flipping-a-coin issue...

[mattgarrish]

It looks like only Apple, Thorium and VitalSource have implemented the object going by the 3.3 test results, so maybe we should find out how they've done their implementations and match up to that. I don't expect we'll get many new implementations going forward.

@danielweck @llemeurfr @davemanhall @rickj (or whoever might know) can you confirm if hasFeature is case-sensitive or not?