sf -update wikidata downloads Archivematica extensions as well

[ross-spencer]

I haven't explored why, but sf -update wikidata is downloading Archivematica extensions as well. This in and of itself is not particularly harmful, but does impede testing of the Wikidata sig.

spencer@rvarchivespencer2:/tmp$ sf -update wikidata
... downloading latest signature file ...
... writing /home/spencer/siegfried/default.sig ...
Your signature file has been updated

spencer@rvarchivespencer2:/tmp$ sf -version
siegfried 1.9.4
/home/spencer/siegfried/default.sig (2022-11-06T17:46:49+01:00)
identifiers:
  - archivematica: wikidata-definitions-3.0.0 (2022-11-06); extensions: archivematica-fmt2.xml, archivematica-fmt3.xml, archivematica-fmt4.xml, archivematica-fmt5.xml

[ross-spencer]

NB. connected to https://github.com/richardlehane/siegfried/pull/178 which should be the first place to look i.e. in gen.go.

[richardlehane]

this is the fault of my crappy config.go package: if you build signatures sequentially in the same session, previous config options can easily pollute later builds.

In this case the archivematica.sig is being built before the wikidata one and the extend config option is still set. There is a config.Clear() command that you can pass to clear earlier options.

Probably the cleanest way to fix this would be to add config.Clear() at the head of the wikidataOpts slice i.e.:

wikidataOpts := []config.Option{
  config.Clear(),
  config.SetWikidataNamespace(),
  config.SetWikidataNoPRONOM(),
}

config.Clear() doesn't currently wipe the extend option so that would need to be updated as well. config.Clear() is defined in /pkg/config/identifier.go and should be updated like:

func Clear() func() private {
    return func() private {
        identifier.name = ""
        identifier.extend = nil
                loc.fdd = ""
        mimeinfo.mi = ""
        return private{}
    }
}

[richardlehane]

fixed with v1.10.0