+1
Thanks,
Shivansh
On Mon, Nov 2, 2026 at 1:00 PM Mridul Pathak via dev <
[email protected]>
wrote:
Hi all,
Following the BIRT and bi threads, where Jacopo suggested taking the
same
approach to other experimental features, I went through applications,
framework and every plugin looking for components that are unused,
unmaintained or already disabled. The evidence for each one (origin and
history, status, JIRA, mailing-list and wiki mentions, consumers, and
what
stays behind in existing databases) is on the wiki:
https://cwiki.apache.org/confluence/spaces/OFBIZ/pages/462815951/Proposal+Retiring+Experimental+and+Abandoned+Features+2026
In short, I'd like to propose removing from trunk only: exampleext (the
extension pattern it demonstrates would move to a page in the developer
manual), multiflex, msggateway, the Clojure service engine, pricat,
passport, the ldap plugin (the framework's own LDAP login stays), the
old
webtools developer pages (EOModel export, JSP entity tools, 2006 test
endpoints) and remote EntitySync push/pull (offline sync stays). For
SOAP
and RMI I'd suggest we simply act on Anil's July thread, which had
support
and no objections.
I'd also remove four larger features that are unmaintained and either no
longer work or would need a redesign: scrum (donated in 2011; its SVN
integration has been broken since 2021 and its demo page shows an
error),
cmssite (its demo pages are FreeMarker templates stored in the database,
which the CVE-2026-29207 fixes no longer render), datafile (no
definitions
ship on trunk since the Taxware/ZipSales removal) and the iCalendar
servlet
(only the demo calendar has worked since 2016, and it never supported
CalDAV).
External login and single sign-on are real needs, but passport and the
ldap
plugin are not the way to meet them: they use provider-specific APIs,
are
not wired into any login page, and LinkedIn login no longer works. I'd
rather remove them and propose, in a separate thread, a plugin built on
standard OpenID Connect that covers Google, Microsoft, GitHub and
Keycloak-style providers.
projectmgr and assetmaint have real users and stay. For myportal,
firstdatapaymentgateway and the JMS engine I found no evidence of use:
myportal stays for now because core depends on it, firstdata is our last
payment gateway integration, and JMS cache clearing is off by default.
If
you rely on any of these three, or would like to look after one, please
say
so; otherwise I'd bring them back for removal in a later round, after
decoupling core from myportal.
Each removal would get its own JIRA and PRs, smallest first, and nothing
touches release24.09. Tables and seed data from removed components stay
in
existing databases (OFBiz never drops them), and each removal's release
note will list them. If you use anything on the removal list, or think
one
should stay, please say so here.
Thanks,
Mridul Pathak
