Thanks for the engagement. This thread started over 72 hours ago and I've only seen support. Per our process and ASF norms 72 hrs is sufficient time for a discussion of this nature.
I'll write the FSIP now and we'll hold a VOTE. Dhroov - i will be adding tickets ; this is a long term need and it will take both people writing tests and people reviewing tests. As a "test" we can start w one area as you've mentioned but not sure that's the best one. Sent from Gmail Mobile On Mon, Nov 2, 2026 at 4:11 AM Dhroov Sankla <[email protected]> wrote: > Hi team, > > +1 for this FSIP initiative. > > Looking at SonarQube, core modules such as fineract-charge,fineract-accounting > and other modules currently show 0% coverage. I would like to contribute > by writing Cucumber E2E and integration tests for *fineract-charge* to > help establish its baseline. > > Is there a specific Jira epic or tracking issue logged for this FSIP > effort where sub-tasks are being organized? > > Regards, > > Dhroov > > > On Wed, Oct 28, 2026 at 12:18 PM Aman Mittal <[email protected]> > wrote: > >> Hi James >> >> +1 for this FSIP, however i want to know how we will track the coverage >> for Fineract. One thing we can rely on is to use sonarqube >> https://sonarcloud.io/project/overview?id=apache_fineract (While it >> needs configuring) it would be a good way to track the maintainability and >> coverage overtime. Can we mandate coverage in the e2e coverage report or >> perhaps utilize Codecov for new coverage? >> >> Regards >> Aman >> >> On Wed, Oct 28, 2026 at 1:13 AM James Dailey <[email protected]> wrote: >> >>> Devs - >>> >>> As recently discussed on matrix >>> https://matrix.to/#/#apache-fineract-dev:matrix.org >>> >>> We have a LOT of tests. Adam S. did much of the work to enable them to >>> run in parallel on VMs. [1] [2], I believe alongside Kristoff. This >>> enables them to run without bogging down. >>> >>> Yet, despite having something in the range of >1500 tests, we do not >>> have good test coverages for large swaths of the project. It is HEAVILY >>> skewed toward the Progressive Loan module. >>> >>> I would like to propose that a new FSIP (fineract significant >>> improvement project) is discussed now, formulated, then voted upon in a >>> week or two, then worked on by a range of folks over the next year. i.e. >>> GSOC students, other volunteers, new devs, etc. We can certainly get some >>> help from the AI tooling per previous discussions. Brainstorming time. >>> >>> AND, as a corollary, if we don't want to support an area of >>> functionality, we should prune it from the code release. >>> >>> To put an arbitrary number to this: we need at least 40-50% coverage of >>> an area of code to allow it to continue to be in the project. We want 70% >>> coverage for some level of stability. We ramp this up on a quarterly >>> reporting basis: >>> >>> 2027: >>> * January 10th - what test coverage do we have today across the modules, >>> assumed to be 0% for some modules, and up to 70% for others. Baseline >>> published. >>> * April 10th - at least 10% cucumber test coverage for all components >>> that we intend to keep. Plans to prune begin. >>> * July 10th - at least 20% cucumber test coverage for all components >>> that we intend to keep. >>> * Oct 10 - at least 40% . Pruning. >>> * Jan 2028 - stability >>> >>> To work on something like this, you need to dig into the topic, examine >>> the code, pay close attention to the existing cucumber tests and >>> setup, write domain specific tests, and incrementally increase the test >>> coverage. It must be coordinated and someone must own it - I would want an >>> eager volunteer for this, just for the coordination. >>> >>> [1] >>> https://github.com/apache/fineract/tree/develop/fineract-e2e-tests-runner/src/test/resources/features >>> [2] (.github/workflows/build-e2e-tests.yml, scripts/split-features.sh, >>> >>> Comments, concerns, feedback? >>> >>> -James >>> >>>
