Issues

The #201 regression test is ~10% flaky on main: a second ExprTk race between clone() and a GIL-free evaluate(), not compile()

#257

· wshlavacek opened

on Aug 10, 2026

a parameter carrying an SBML assignmentRule is still listed as a knob: primary_param_names keeps it, and set_param accepts a write the rule discards

#256

· wshlavacek opened

on Aug 10, 2026

compute_all_sensitivities: whether the solve succeeds depends on chunk_size, a documented performance knob

#243

· wshlavacek opened

on Aug 9, 2026

codegen: the ^ → pow() rewrite splits a scientific-notation literal at its exponent sign, emitting invalid C

#240

· wshlavacek opened

on Aug 9, 2026

`not(...)` around a compound condition is declined where the `!` spelling of the same condition is admitted

#234

· wshlavacek opened

on Aug 9, 2026

A time threshold inside a called functionDefinition registers no discontinuity root (the GH #72 twin of #194's funcDef descent)

#231

· wshlavacek opened

on Aug 8, 2026

scripts/rebuild_editable.py cannot configure in a uv venv: pybind11 is a build-system requirement, so no extra puts it in .venv — and it is the remedy the staleness guard names

#229

· wshlavacek opened

on Aug 8, 2026

primary_param_names is wrong in both directions on a .net model: it omits literal-valued constants and lists every function name

#227

· wshlavacek opened

on Aug 8, 2026

A per-species atol built from initial values still misses a species that decays below it: expose CVodeWFtolerances (the over-time half of #196)

#213

· wshlavacek opened

on Aug 7, 2026

The codegen .so cache grows without bound and has no supported way to inspect, prune, or clear it (2.0 GB / 14,377 entries on one machine)

#205

· wshlavacek opened

on Aug 7, 2026

NFsim fires a rule at 2x when its reactant pattern is symmetric and its rate is a function

#195

· wshlavacek opened

on Aug 6, 2026

BNGSIM_KLU_AUTOBUILD cannot complete on a bare Linux host: SuiteSparse_config requires a BLAS, so REQUIRE_KLU=ON makes a source install a hard configure failure

bug

#178

· wshlavacek opened

on Aug 5, 2026

Navigation Menu

Uh oh!

All issues

The #201 regression test is ~10% flaky on main: a second ExprTk race between clone() and a GIL-free evaluate(), not compile()

a parameter carrying an SBML assignmentRule is still listed as a knob: primary_param_names keeps it, and set_param accepts a write the rule discards

compute_all_sensitivities: whether the solve succeeds depends on chunk_size, a documented performance knob

codegen: the ^ → pow() rewrite splits a scientific-notation literal at its exponent sign, emitting invalid C

`not(...)` around a compound condition is declined where the `!` spelling of the same condition is admitted

A time threshold inside a called functionDefinition registers no discontinuity root (the GH #72 twin of #194's funcDef descent)

scripts/rebuild_editable.py cannot configure in a uv venv: pybind11 is a build-system requirement, so no extra puts it in .venv — and it is the remedy the staleness guard names

primary_param_names is wrong in both directions on a .net model: it omits literal-valued constants and lists every function name

A per-species atol built from initial values still misses a species that decays below it: expose CVodeWFtolerances (the over-time half of #196)

The codegen .so cache grows without bound and has no supported way to inspect, prune, or clear it (2.0 GB / 14,377 entries on one machine)

NFsim fires a rule at 2x when its reactant pattern is symmetric and its rate is a function

BNGSIM_KLU_AUTOBUILD cannot complete on a bare Linux host: SuiteSparse_config requires a BLAS, so REQUIRE_KLU=ON makes a source install a hard configure failure

Uh oh!

Search results