XML invalid entries in RoR

[Markus Demleitner]

Hi Paul,

On Tue, Feb 11, 2025 at 08:31:29AM +0000, Paul Harrison via registry wrote:
> there is quite a large proportion of the entries for registries in
> the RoR that are technically XML invalid
>
> e.g.
>
> % curl "http://rofr.ivoa.net/oai?verb=GetRecord&metadataPrefix=ivo_vor&identifier=ivo://src.pas/__system__/services/registry"
[...]

> is invalid as there is no declaration of the namespace of the vg:
> prefix. Although it is obvious what namespace this is referring to,
> if the output is read via an XML processor it will immediately stop
> at this error. I am not sure what the is the official way of
> getting these entries fixed, but clearly the most efficient would
> be a bulk update on the RoR content if someone has that access.

Ach... yeah, we should make more of an effort to maintain the RofR
codebase (or perhaps re-implement it, shedding some of the cruft);
I'm having "add vocabulary validation" on my todo list for a long
while.

At least in this case, I'm pretty sure the problem is that when
re-harvesting the record, the namespace declarations on the OAI-PMH
root element get lost; to see what probably happened, see:
<http://pithia.cbk.waw.pl/oai.xml?verb=GetRecord&metadataPrefix=ivo_vor&identifier=ivo://src.pas/__system__/services/registry>

Paul, in case you'd like to try your hand fixing this, the RofR
source is at <https://github.com/ivoa/rofr.ivoa.net>, although I
suspect the re-harvesting code is not in there yet (Dustin?).

Thanks,

          Markus