[bug#48437] [PATCH] lint: archival: Lookup content in Disarchive database.

["Ludovic Courtès" <ludo@gnu.org>]

Hi!

Timothy Sample <samplet@ngyro.com> skribis:

> Ludovic Courtès <ludo@gnu.org> writes:
>
>> This patch makes the ‘archival’ checker check the Disarchive database(s)
>> when SWH ‘lookup-content’ returns #f.  [...]
>>
>> It assumes that the swhid found in the Disarchive metadata is valid, a
>> reasonable assumption IMO.
>>
>> Thoughts?
>
> One thing to consider is that just because Disarchive has captured an
> archive’s metadata and computed the SWHID of its contents doesn’t mean
> that the contents are actually in the SWH archive.  (Maybe that’s what
> you meant when you wrote about valid IDs above.)

Yes, I thought we could assume the contents were necessarily in the
archive.

> It would be neat if the lint check looked up the SWHID to see if it
> exists.  Unfortunately, Disarchive doesn’t make getting the underlying
> SWHID easy at the moment.  One option would be to pass a resolver to
> “disarchive-assemble” that exfiltrates the ID using “set!”.  Another
> one would be to “read” the specification and search for a form like
> ‘(swhid "swh:1:dir:...")’.  Neither is particularly lovely....

Hmm yeah.  There’s no API to deserialize the (disarchive …) sexp as a
record, right?

> Other than that, the code looks good and everything seems to work.  :)

Maybe we can assume (with a comment) that the SWHID points to a valid
content, and when creating the Disarchive database we actually make sure
this is the case?

More generally, we need to talk about that database, how to create it
and maintain it.  :-)

Thanks,
Ludo’.