Matthias Volk

Given your last example, I would be in favor of making the simplification a bit more transparent such that at least some traceability is given. Ideally we decouple a general...

#613 adds this workaround. To fix it, we might want to look into https://github.com/trolando/sylvan/issues/42.

Thanks for the report. While the call should indeed only use the right variable type, giving a segfault is definitely not good. I added assertions in Storm, but I believe...

Related to https://github.com/moves-rwth/stormpy/issues/95 After a quick look, I found two alternatives on the pybind repos: - https://github.com/pybind/cmake_example - https://github.com/pybind/scikit_build_example The CMake one is probably closer to our current setup but...

I tried building with mold, see [this branch](https://github.com/moves-rwth/storm/compare/master...volkm:storm:mold) and [this action](https://github.com/volkm/storm/actions/runs/9452779651). Some first insights: - mold is available as package for Ubuntu >= 22.04 and Debian >= 12. - Older...

There is definitely some room for improvement by combing commands in one layer. Another possible option could be to simply install Storm via CMake and then remove all the build...

I agree that installing within the Docker container is the best option. This issue then depends on #527.

I would be in favor of using `pathlib.Path`. My approach would be twofold: we first use `pathlib.Path` but allow strings as a fall-back (and give a deprecation warning). Later, we...

Both the `SparseSimulator` and the `PrismSimulator` currently only support discrete-time models. The simulators need an extension to continuous-time models.

depends on https://github.com/moves-rwth/storm/issues/816