Many Branches

[dickinsonre]

I see you have a lot of branches—should I make WASM engines for all branches? I use the engine in Apps. I guess i can just use them all?

[cbuahin]

@dickinsonre, you are right. We need to do some cleaning up. I will work on things on my end.

@wiesnerfriedman, once you merge you changes, could you please delete stale branches?

[cbuahin]

@dickinsonre , the branch to use is develop or swmm6_rel and main.

[dickinsonre]

i actually vibe coded an app to run them all - i will scale back a bit,

[dickinsonre]

Making progress, or the AI tools are making progress, I now have thousands of models comparing swmm5 and SWMM6. L:et me try to write about the totality of the comparison for now.

"I'd only checked the main branch of HydroCouple/openswmm.engine, which is a dead mirror of stock SWMM 5.2.4 frozen since May 28. The real work is on swmm6_rel: C++20, version 6.0.0, 960 commits ahead of main, last commit two days ago (July 22), with genuinely serious content in the log — Anderson-accelerated Picard iteration, semi-implicit node continuity, a dynamic Preissmann slot, live 1D/2D coupling. It pulls Kokkos, SUNDIALS, and Hypre. It ships as an installable wheel (pip install openswmm, 6.0.0a2). I installed it and actually ran your 1729 benchmark set through it against stock SWMM5.2.4 — real comparison data, in the file above, not fabricated.

Direct answer to your issue: build WASM against swmm6_rel only. Four branches are live in-progress feature work worth tracking but not building yet (fix/windows-cuda-release-build, feature/uncertainty-sidecar, post/pr13-wq-uncertainty-layer, develop). The rest — swmm6_exp, five copilot/* branches, both build-v5.2.x, both dev-* branches, both ci/docs-deploy* — are dead or fully merged in already. Full breakdown with dates and divergence counts is in the file."?