Releases: hannorein/rebound
Releases · hannorein/rebound
4.6.0
4.5.1
- Added leapfrog integrators of order 4, 6, and 8. Order can be set with
r->ri_leapfrog->order. - Heartbeat function is now also called when using
reb_simulation_steps(). - OrbitPlot supports more colours.
- Fixed an alignment issue when comparing binary snapshots which could have triggered a segfault.
4.5.0
- Added support for (Frequency) Modified Fourier Transforms. Heavily based on David Nesvorny's code. The new functions are
reb_frequency_analysis()in C andrebound.frequency_analysis()in python. For usage, see C examplessecular_frequenciesandfrequency_analysisas well as the iPython notebookFrequencyAnalysis. - Support for Jacobi coordinates added when using WHFast with OpenMP.
- Allow negative periods when initializing hyperbolic orbits.
4.4.11
- The collision resolve function now returns a type
enum REB_COLLISION_RESOLVE_OUTCOME. The actual integer values remain unchanged. - Bug in TRACE was fixed.
- Convergence check for M_to_E function.
- New API example that shows how to use the Kepler solver without a REBOUND simulation. Updated other examples.
4.4.10
4.4.9
4.4.8
4.4.7
- Added option to disable SSL checks for Horizon queries with
rebound.horizons.SSL_CONTEXT = 'unverified'. - Added unit tests.
- Added barycentric coordinates for WHFast.
- Bug fix for when MEGNO is used with adaptive timestepping.
- Added more error messages.
- Fixed various issues in documentation.
4.4.6
- When initializing particles with "uniform" in python, REBOUND now uses its own
reb_random_uniform()function. This avoids importing the "random" library and makes results reproducible as the random seed of the simulation is used when generating random numbers. - More cracefull interrupt handling. REBOUND now stop the integration after the next timestep when CTRL-C is pressed the first time. If CTRL-C s pressed twice, then long loops (during gravity, collision calculations) are terminated immediately. Continuing an integration after one CTRL-C press should be easier with this change as the simulation does not get corrupted.
- Fixed typos in documentation.