Crashing with a NaN time

[Alpora]

I am getting occasional crashes, mainly within the tracking station. I have no consistent way of reproducing it, but it happens every so often if I'm just looking around in the tracking station.

INFO.20260430-173734.7464.log

FATAL.20260430-173916.7464.log

lmk if anything else is needed.

EDIT: I should mention that I am using Sol instead of RSS.

[pleroy]

Could you post your save too?

[Alpora]

how should I do that? as a zip?

[Alpora]

space is easy.zip

[pleroy]

Decoded stack trace:

Principia/ksp_plugin/interface_planetarium.cpp

Lines 268 to 299 in 34c8d55

	int* const vertex_count) {
	journal::Method<journal::PlanetariumPlotCelestialPastTrajectory> m(
	{planetarium,
	plugin,
	celestial_index,
	max_history_length,
	vertices,
	vertices_size},
	{minimal_distance_from_camera, vertex_count});
	CHECK(plugin != nullptr);
	CHECK(planetarium != nullptr);
	*vertex_count = 0;
	// Do not plot the past when there is a target vessel as it is misleading.
	if (plugin->renderer().HasTargetVessel()) {
	*minimal_distance_from_camera = std::numeric_limits<double>::infinity();
	return m.Return();
	} else {
	auto const& celestial_trajectory =
	plugin->GetCelestial(celestial_index).trajectory();
	Instant const desired_first_time =
	plugin->CurrentTime() - max_history_length * Second;
	// Since we would want to plot starting from `desired_first_time`, ask the
	// reanimator to reconstruct the past. That may take a while, during which
	// time the history will be shorter than desired.
	plugin->RequestReanimation(desired_first_time);
	Instant const first_time =
	std::max(desired_first_time, celestial_trajectory.t_min());
	Length minimal_distance;
	planetarium->PlotMethod4(

Principia/ksp_plugin/planetarium_body.hpp

Lines 158 to 210 in 34c8d55

	Length* const minimal_distance) const {
	auto const final_time = reverse ? first_time : last_time;
	auto previous_time = reverse ? last_time : first_time;
	if (minimal_distance != nullptr) {
	*minimal_distance = Infinity<Length>;
	}
	Sign const direction = reverse ? Sign::Negative() : Sign::Positive();
	if (direction * (final_time - previous_time) <= Time{}) {
	return;
	}
	DegreesOfFreedom<Navigation> const initial_degrees_of_freedom =
	EvaluateDegreesOfFreedomInNavigation<Frame>(
	*plotting_frame_, trajectory, previous_time);
	Position<Navigation> previous_position =
	initial_degrees_of_freedom.position();
	Vector<AngularFrequency, Navigation> previous_projected_velocity =
	ProperMotion(initial_degrees_of_freedom) *
	Normalize(previous_position - perspective_.camera());
	Time Δt = final_time - previous_time;
	add_point(plotting_to_scaled_space_(previous_time, previous_position));
	int points_added = 1;
	Instant t;
	Angle rms_apparent_distance;
	Position<Navigation> position;
	Vector<AngularFrequency, Navigation> projected_velocity;
	Square<Length> minimal_squared_distance = Infinity<Square<Length>>;
	goto estimate_tan²_error;
	while (points_added < max_points &&
	direction * (previous_time - final_time) < Time{}) {
	do {
	// Consider the case where the trajectory is an arc of circle of angle 2 α
	// and the segment is the corresponding chord. The higher-order term of
	// the integral below is 4 α⁵ / 15, which yields an RMS of O(α²) after
	// division by the time interval and application of the square root.
	// Therefore, our errors are of degree 2 in time.
	// A safety factor prevents catastrophic retries.
	Δt *= 0.9 * Sqrt(parameters_.angular_resolution_ / rms_apparent_distance);
	estimate_tan²_error:
	t = previous_time + Δt;
	if (direction * (t - final_time) > Time{}) {
	t = final_time;
	Δt = t - previous_time;
	}
	auto const degrees_of_freedom =
	EvaluateDegreesOfFreedomInNavigation<Frame>(

Principia/ksp_plugin/planetarium_body.hpp

Line 49 in 34c8d55

return to_plotting_frame_at_t(trajectory.EvaluateDegreesOfFreedom(t));

Principia/physics/rigid_reference_frame_body.hpp

Lines 32 to 33 in 34c8d55

	Instant const& t) const {
	return ToThisFrameAtTime(t).template Forget<SimilarMotion>();

Principia/physics/body_centred_non_rotating_reference_frame_body.hpp

Lines 67 to 72 in 34c8d55

	ToThisFrameAtTime(Instant const& t) const {
	DegreesOfFreedom<InertialFrame> const centre_degrees_of_freedom =
	centre_trajectory_->EvaluateDegreesOfFreedom(t);
	RigidTransformation<InertialFrame, ThisFrame> const
	rigid_transformation(centre_degrees_of_freedom.position(),

Principia/geometry/affine_map_body.hpp

Line 20 in 34c8d55

to_origin_(to_origin),

Principia/physics/continuous_trajectory_body.hpp

Lines 197 to 199 in 34c8d55

	Instant const& time) const {
	absl::ReaderMutexLock l(&lock_);
	return EvaluateDegreesOfFreedomLocked(time);

Principia/physics/continuous_trajectory_body.hpp

Lines 709 to 710 in 34c8d55

	Instant const& time) const {
	CHECK_LE(t_min_locked(), time);

[Alpora]

is there anything in particular i should be seeing here?

[pleroy]

is there anything in particular i should be seeing here?

I'll just update the issue with technical information as I analyse it. It's mostly for our own use.

One thing that would be nice is if you could give us a journal, but that may not be easy since you mentioned that you have no consistent way of reproducing the crash.

[pleroy]

I can't make sense of this bug. The save appears fine, and I don't see how a NaN could be generated. I'd need to see the parameters passed to PlotMethod4.

@Alpora I will need a journal to be able to make progress.

[Alpora]

@pleroy luckily I was journaling when this one happened. Zipped it up because the file type is apparently not supported as an attachment.

JOURNAL.zip

[pleroy]

@Alpora Unfortunately, even though I can see that the journal ends with a crash in the "right" method, that crash doesn't reproduce on my machine. I suspect some kind of race condition, but I am at a loss as to where/how.

Do you by any chance have crash folders in %USERPROFILE%\AppData\Local\Temp\Squad\Kerbal Space Program\Crashes? If so, could you zip them and attach them here? They may give me a hint.

[Alpora]

@pleroy For a reason I don't know, this crash has not generated a folder in that directory. I assume it's because the crash was from Principia and not the game itself.

[Alpora]

had another crash, I doubt it was the same cause but it occured on the main menu, before I could even fully load in.

Crash_2026-05-01_195054520.zip

[fat-lobyte]

had another crash, I doubt it was the same cause but it occured on the main menu, before I could even fully load in.

Crash_2026-05-01_195054520.zip

SharpDXException: HRESULT: [0x8007000E], Module: [General], ApiCode: [E_OUTOFMEMORY/Out of memory], Message: Not enough memory resources are available to complete this operation.

[Alpora]

@fat-lobyte I have 32gb memory+16gb pagefile lmao how

[pleroy]

I have forgotten how to count that low 🙂

Seriously, you are using Sol which has (count them) 52 celestials, each with super-detailed textures. Yeah, that's going to burn memory.

(Not saying it explains the Principia bug, but if you are out of memory you'll be in a world of pain.)

[Alpora]

I'm upping my pagefile to 32gb lol