Doesn’t pick up on invalid empty identifier/source

[jenstroeger]

It would be helpful if EpubCheck would info/warn that identifiers are empty (or don’t match a pattern). For example,

<dc:identifier id="id-e985e6af">urn:isbn:</dc:identifier>
<meta refines="#id-e985e6af" property="identifier-type" scheme="onix:codelist5">15</meta>
<dc:source id="source-e985e6af">urn:isbn:</dc:source>
<meta refines="#source-e985e6af" property="source-of">pagination</meta>

isn’t being flagged at all. Same goes for UUIDs instead of ISBNs. I guess it would be ok with the elements contained no text whatsoever (in which case maybe <meta> could be flagged), but a n existing urn:isbn: expresses an intent to use an identifier, but there’s none…

Also, would it make sense to flag numbers with dashes instead of hyphens? Not sure it matters, but an ISBN like 123–0-1234567–0-0 seems odd to me, but is accepted.

[rdeltour]

It isn't a spec violation, since EPUB basically allows any string as an identifier, but yes I can see how it can be useful.

There is a precedent of implementing UUID validation in #853 (so I'm surprised you say the same goes for UUIDs?).

It's low priority, but PRs are always welcome if anyone wants to fast-track an implementation 😊

[jenstroeger]

Thanks, @rdeltour!

It's low priority, but PRs are always welcome if anyone wants to fast-track an implementation 😊

If you can point me at the vicinity of where the responsible code lives, I’d take a stab at it.