> I don't think sandbox was intended originally to host ready modules with a 
> release process.

Yeah, it's a bit handwave-y, but my understanding of "sandbox" was
that it was for more experimental code, which doesn't seem like it's
the case for the encryption stuff at this point?

Even outside of the messaging or implications about experimental vs
production-readiness, the code would be easier to consume if it was
elsewhere.  Sandbox is "just" a code repo, so consumers need to do
their own releases, compatibility testing (for their particular Solr
version), packaging or installing into Solr, etc.  Moving it into the
main repo as a module would provide a lot of benefit to any potential
consumers.

Anyway I'm +1 on the overall idea, as I understand it at least.  But I
do have some feedback on the writeup/proposal:

- if the "ask" at the heart of the SIP is to change where the code
lives, the writeup should spend a bit more time outlining why that's a
better path forward.  What are the problems with keeping the code in
"sandbox"?  What's the benefit to a potential user of having the code
in the main repo?  (We've discussed some of those details already in
this thread, but it's still worth summarizing in the "Motivation"
section of the SIP)

- I love the architecture docs in the current "Proposed Changes"
section!  But the section spends so much time detailing the current
state of the code that it doesn't really talk much about what would
actually change.  I get that this is mostly "just" a move, but there
will be changes I imagine: does the sandbox code require any changes
or updates to fit into the Solr "module" paradigm?  is it already
building with Java 21?  since it currently builds against Solr 9.9 in
sandbox does it need any particular changes to get it working against
the "main" branch?  what sort of ref-guide documentation should be
added?  etc.

Best,

Jason

On Tue, Aug 12, 2025 at 5:56 AM Bruno Roustant <[email protected]> wrote:
>
> Thanks David.
>
> Yes, I'm proposing to discuss whether it should move to the main Solr repo.
> In the current state, solr-sandbox seems dedicated to incubating modules.
> Actually, my main point is to say the encryption module should exit the
> incubation state, whether it stays in the sandbox or not. But if it stays
> there, there should be a clear way to differentiate "incubating" or "ready"
> modules for external users.
> - If users want to have encryption for Solr (provided they have the right
> use-case, as described in the doc), they should have confidence they can
> use it.
> - To my knowledge, there is no section in the Solr doc that describes the
> modules in solr-sandbox.
> - Maybe a new module available should be announced in dev and user list.
> But it should wait for some feedback first I think.
>
> Technically, the sandbox is separate from the main repo, so many updates
> need to wait for the next Solr release: development is longer.
> And you are right, there is no test infra nor release process. I don't
> think sandbox was intended originally to host ready modules with a release
> process.

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to