Sounds good, thanks Jirka.

Gj

On Fri, 12 Jul 2024 at 19:27, Jiří Kovalský <[email protected]>
wrote:

> Since there has been no further communication on this topic, let's take
> this as a lazy consensus that signatures won't be considered at all when
> verifying plugins going forward. I have updated step 8 in the "Install
> plugin" Synergy test accordingly:
>
> https://synergy.netbeans.apache.org/#/case/6314/suite/2525/v/1
>
> Anyone please speak up if you disagree.
>
> Mani, Carlos, Geertjan - FYI
>
> Thanks,
> -Jirka
>
> Dne 14. 08. 23 v 19:44 Neil C Smith napsal(a):
> > On Sun, 13 Aug 2023 at 21:10, Matthias Bläsing
> > <[email protected]> wrote:
> >> Reasoning:
> >>
> >>     Plugin unsigned. Please sign (self-signed is ok) and re-submit for
> >>     verification
> >>
> >> This was not a problem in: 11, 12, 16 and 17.
> >>
> >> _Nothing_ changed for these plugins and I don't see why I should was
> >> resources in CI/CD systems and on maven central, just to "fix"
> >> something, that was not broken for a long time.
> >
> > Yes, anything that was previously verified should be allowed through
> > unless it's actually broken.  We have a limited RC window for people
> > to test with plugins as it is.  Making plugin authors jump through
> > unnecessary hoops doesn't help there.
> >
> >> The requirement to sign the plugins is questionable in itself without a
> >> trust anchor or revocation list, but I can live with with requiring
> >> signature for updates (this will become fun, once the signature
> >> expires, but ...)
> >
> > Agreed!  And we have SHA in the catalog which I assume are checked?!
> >
> > As you've raised this before, I would suggest you just kick off a lazy
> > consensus thread on removing the self-sign requirement.  Or on the
> > validation rules as a whole.
> >
> > Best wishes,
> >
> > Neil
>

Reply via email to