This is an archived snapshot of W3C's public bugzilla bug tracker, decommissioned in April 2019. Please see the home page for more details.
Bug 18521, bug 18523, bug 18528 and bug 18529 claim: > We've decided to treat type as an opaque string. However, the Blob constructor [1] and the slice method [2] still lower-case their argument and limit it to the U+0020-U+007E range. This seems inconsistent. [1] http://dev.w3.org/2006/webapi/FileAPI/#constructorBlob [2] http://dev.w3.org/2006/webapi/FileAPI/#dfn-slice
I'm pretty sure that's because we changed our mind. The question is where we should put the constraints and what conversion between code units and bytes we should make. It cannot be a code unit sequence without conversion algorithm to bytes. Making it a ByteString would make some sense...
Ms2ger: In fact, it's not an inconsistency so much as a compromise. Sticking to rigid parsing of type would bind us to specs that are themselves inconsistent (but I treat http://mimesniff.spec.whatwg.org/ as normative these days). Anne: OK, I've logged Bug 23138.
Since Bug 23138 was logged, I think this can be marked INVALID. Bug 23138 brings an IDL change, and some prose change about code points.