Hi all, Please find the PR with the revised changes - https://github.com/apache/incubator-resilientdb/pull/183 Passing workflow checks - https://github.com/apache/incubator-resilientdb/pull/183/checks
Changelog: - Added .bazelignore to ignore the ecosystem directory from being included in the parent Bazel build. - Removed nested NOTICE files to rely on top top-level NOTICE file Regarding attribution, a couple of points: 1. This looks like a GitHub visibility issue as outlined earlier. GitHub, the service, is not able to consolidate history for these subfolders with great accuracy. This does not mean we are losing out on attribution because tools like GitLens can provide the exact commit to which a line of code belongs. 2. We are doing a "Squash and Merge" currently. This is the reason why all commits for the ecosystem folder are found under a parent commit. This is a cleaner strategy given our monorepo changes contain 700+ individual commits. 3. In conclusion, attribution is maintained and is visible using the right tool. We could use your help if you believe there are better strategies or tools to do the monorepo changes we are looking for. On Fri, Aug 22, 2025 at 1:56 PM Harish Gokul <harishgoku...@gmail.com> wrote: > Hi Junchao, > > Thanks for pointing it out. I did not know it was just a top-level > requirement. > > On Fri, Aug 22, 2025 at 1:08 AM Junchao Chen <jucc...@ucdavis.edu.invalid> > wrote: > >> hi Harish, >> >> Thanks for raising this task. >> >> As all the repos are merged into one, we don't need the NOTICE >> < >> https://github.com/harish876/incubator-resilientdb/blob/monorepo-setup/ecosystem/cache/resilient-node-cache/NOTICE >> > >> files anymore, only rely on the top one. >> >> >> Junchao >> >> On Fri, Aug 22, 2025 at 12:42 AM Harish Gokul <khgo...@ucdavis.edu.invalid >> > >> wrote: >> >> > Hi all, >> > >> > As part of improving our release process and making it easier for the >> > Apache Incubator committee to review and validate our releases, I would >> > like to propose a shift towards a *monorepo setup* for ResilientDB. >> > >> > Specifically, we suggest adding a new top-level *ecosystem/* folder >> inside >> > the existing *incubator-resilientdb >> > <https://github.com/apache/incubator-resilientdb>* repository. All our >> > ecosystem projects (currently spread across multiple repos) would be >> > consolidated here. >> > Motivation >> > >> > - >> > >> > *Simplified Release Process:* A single repository allows us to cut >> and >> > vote on releases more easily, with clear visibility for the IPMC. >> > - >> > >> > *Improved Organization:* Contributors, reviewers, and downstream >> users >> > will find everything related to ResilientDB in one place, with >> > consistent >> > structure and governance. >> > - >> > >> > *Attribution & History Retained:* By using *git subtree*, we will >> > preserve commit history and attribution when migrating existing >> > projects. >> > - >> > >> > *Flexible Consumption:* Users who do not need the full repository can >> > leverage *git sparse-checkout* to pull only the ecosystem subfolders >> > they are interested in. >> > >> > >> > >> > >> > >> > Implementation Plan >> > >> > 1. >> > >> > Introduce a new ecosystem/ directory in the incubator-resilientdb >> repo. >> > 2. >> > >> > Gradually migrate ecosystem projects into this directory using *git >> > subtree add*, ensuring history and attribution are retained. >> > 3. >> > >> > Configure existing github actions to not include the ecosystem >> folder, >> > and introduce CI/CD checks for these folders gradually. >> > 4. >> > >> > Document usage of git sparse-checkout for contributors/users who >> prefer >> > working with a subset of the repository. >> > >> > This structure keeps our repository aligned with ASF release guidelines >> > while improving both contributor and reviewer experience. >> > >> > Progress Plan >> > >> > - An initial commit for this plan can be found here. This is >> obviously >> > subject to change based on your feedback. - >> > https://github.com/apache/incubator-resilientdb/pull/182 >> > >> > There is an Open PR to correct the current github workflow actions. >> > https://github.com/apache/incubator-resilientdb/pull/183 >> > >> > Documentation on usage can be found here and here >> > >> https://github.com/apache/incubator-resilientdb/tree/master/ecosystem >> > >> > >> > Issues >> > - There is a github attribution visibility error, where the github >> > author names pop up, but their commits are linked under a single merge >> > commit. This might mostly be due to a bug in the way we added git >> subtrees, >> > but this might need some double checking. Regardless, if you have an >> > approach better than git subtrees or ways to rectify this bug , we >> would be >> > happy to discuss. >> > >> > Looking forward to hearing your thoughts and feedback. >> > >> > Best, >> > Harish Krishnakumar >> > PPMC Apache ResilientDB >> > >> >
