Thanks for the feedback! I believe (5) can be re-proposed to accommodate… we heard other questions about this in the Town Hall, too, for what it’s worth. I believe marking the DASHBOARD_NATIVE_FILTERS flag as deprecated in the subsequent major version, along with relative code paths, and enabling your migration workflow (thank you!!!) would suffice for this 3.0 release. We can then remove the flag, force the migration, and clear other code paths as a change in 4.0.
I’ll hereby DROP the prior proposal, and hereby submit the above proposal for lazy consensus. I’ll update the task on the Github Project accordingly. This may preclude removing the react-select component proposed in #7, but I think we can let that proposal stand as-is. Thanks, Evan Rusackas Apache Superset PMC On Mar 17, 2023 at 11:37 AM -0600, John Bodley <john.bod...@airbnb.com.invalid>, wrote: > Hi Evan, > > Thanks for sharing. > > Regarding (5) is it possible to delay this? We (Airbnb) are currently > working on a revised migration workflow from legacy filters (filter-box > charts, filter scopes, et al.) to native filters (see SIP-64 for details). > It seems potentially prudent to have this logic persist for a period to > allow organizations sufficient time to manually migrate their legacy > filters prior to fully deprecating/removing filter-box charts. > > -John > > On Sat, Mar 18, 2023 at 4:43 AM Evan Rusackas <e...@preset.io.invalid> > wrote: > > > Hi all, > > > > As proposed last week, we’re going to be sending waves of (mostly > > breaking) change proposals for lazy consensus. This is the first round, but > > there will likely be more. Again, these are seeking permission to merge > > these changes in while the breaking change window is open on the repo. They > > might not all make it in, and those that are not completed before the > > release is cut will be carried over as “pre-approved” for the subsequent > > major release (4.0). > > > > Each of these is seeking individual lazy consensus. If you object to any > > of these changes or wish for further clarification on any of them, please > > let us know which you would like to stop or discuss, and we can hash it out > > in a here on the list, on Slack, in Town Hall, etc. > > > > 1) Migrate RLS server side page to client side React - this will re-merge > > a previously reverted PR that converts this view to a React page in our > > Single Page App > > > > 2) Deprecate and remove the UX_BETA feature flag - these features have > > been stable/tested long enough! > > > > 3) Make all API v1 resource names plural. Ex: /api/v1/dashboard/ -> > > /api/v1/dashboards/ > > > > 4) Remove support for legacy short URLs ( /r URLs) > > > > 5) Deprecate/Remove FilterBox component, > > > > 6) Remove drag and drop feature flags (ENABLE_EXPLORE_DRAG_AND_DROP & > > ENABLE_DND_WITH_CLICK_UX), and remove legacy code paths > > > > 7) Remove react-select package dependency and clear out or refactor > > remaining usage accordingly > > > > 8) Remove Pivot Table in favor of Pivot Table V2 > > > > 9) Remove all superset/* REST APIs marked as deprecated > > > > 10) Set GENERIC_CHART_AXES Feature flag to True by default > > > > 11) [Cleanup] Follow through on removal of Legacy Druid Connector > > > > 12) convert ENABLE_BROAD_ACTIVITY_ACCESS and MENU_HIDE_USER_INFO into > > feature flags > > > > 13) Remove/Migrate Multiple Line Charts in favor of ECharts replacement > > > > > > Thank you all for your consideration. The next wave (dare I say tranche?) > > will be coming soon. > > > > Evan Rusackas > > Preset | preset.io > > Apache Superset PMC > >
