Hi Jan,

I guess I'd be a bit uncomfortable with deprecating v1 until v2 has
had some decent dog-fooding.  That's probably why the SIP suggests
that deprecation coincide with "v2-used-internally": because the
earlier "v2-complete" doesn't mean that anything even uses the v2 API
yet.  But that's just my two cents: if there's consensus that we have
enough confidence in v2, or that it'd be valuable to get the
deprecation warning out there earlier, then I can live with that.

On your second point, I'm a little leery of tying "v2-complete",
"v2-used-internally", and "eventual removal" to any particular
releases - mostly because the timing around our releases has been
somewhat unpredictable.  I certainly agree that v1 removal has to
happen on a major-version boundary, and if 10.0 comes around quickly,
then the timeline you suggest seems pretty plausible. But 10.0 might
take much longer (9.0 took 3 years after all), in which case it seems
reasonable to me that v1 could be removed for 10.0 if all the
prerequisites happen early enough in the 9.x line.  I guess I'm
worried about a scenario where both 10.x and 11.x are long release
lines, and we end up supporting v1 forever.  Any thoughts on that?

Best,

Jason

On Sun, Nov 6, 2022 at 3:04 PM Jan Høydahl <[email protected]> wrote:
>
> Both.
>
> 9x: v2 complete, deprecate v1 (and adding some deprecation noise to logs)
> 10.0 Switch to using v2 internally
> 11.0 Remove v1
>
> Jan
>
> > 4. nov. 2022 kl. 16:25 skrev Jason Gerlowski <[email protected]>:
> >
> > Hi Jan,
> >
> > Just trying to make sure I understand your suggestion.  Are you suggesting:
> >
> > (1) that we announce v1 as deprecated at "V2-API-Complete" (instead of
> > the later "v2-API-used-internally")? Or...
> > (2) that we plan "v2-API-used-internally" to coincide with 10.0?
> > (3) Or both?
> >
> > Best,
> >
> > Jason
> >
> > On Mon, Oct 31, 2022 at 5:04 PM Jan Høydahl <[email protected]> wrote:
> >>
> >> Thanks for a thorough SIP.
> >>
> >> Wrt deprecation plan, could we not have "v2-API-Complete" in e.g. 9.5 (and 
> >> deprecate v1). Then we wait until 10.0 with "v2-API-used-internally", and 
> >> 11.0 for removing v1. Say we release 10.0 in March 2023, then the new main 
> >> will be 11.0 and we can already remove v1 in main.
> >>
> >> Looking forward to code generating SolrJ classes. And perhaps community 
> >> members actively using some other prog.language will be empowered to auto 
> >> generate 90% of such clients.
> >>
> >> Jan
> >>
> >>> 31. okt. 2022 kl. 21:23 skrev Jason Gerlowski <[email protected]>:
> >>>
> >>> Hey all,
> >>>
> >>> This morning I published SIP-16, which proposes the changes necessary
> >>> to "finish" (i.e. plug coverage gaps and polish) Solr's v2 APIs and a
> >>> path to deprecating Solr's v1 APIs.
> >>>
> >>> The SIP can be found here:
> >>> https://cwiki.apache.org/confluence/display/SOLR/SIP-16%3A+Polish+and+Prepare+v2+APIs+for+v1+Deprecation
> >>>
> >>> Please read the SIP description and come back here for discussion.  As
> >>> the discussion progresses we will update the SIP page with any
> >>> outcomes and eventually move things to a VOTE (or lazy consensus).
> >>>
> >>> Looking forward to hearing your feedback!
> >>>
> >>> Best,
> >>>
> >>> Jason
> >>>
> >>> ---------------------------------------------------------------------
> >>> To unsubscribe, e-mail: [email protected]
> >>> For additional commands, e-mail: [email protected]
> >>>
> >>
> >>
> >> ---------------------------------------------------------------------
> >> To unsubscribe, e-mail: [email protected]
> >> For additional commands, e-mail: [email protected]
> >>
> >
> > ---------------------------------------------------------------------
> > To unsubscribe, e-mail: [email protected]
> > For additional commands, e-mail: [email protected]
> >
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [email protected]
> For additional commands, e-mail: [email protected]
>

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

Reply via email to