Crash when clicking on rebase

[kwattt]

INFO.20240609-164451.17344.log
ERROR.20240609-164451.17344.log
FATAL.20240609-170324.17344.log

This happened when planning a second maneouver on a moon flyby, I clicked rebase and crashed without any alert box on screen.

[pleroy]

Thanks for reporting. Is this something that you know how to reproduce? If so could you give us a journal? I could paper over the bug, but I'd like to understand how we end up in this state.

[kwattt]

Hello, I managed to get a journal with the crash. I noticed a really high RAM spike before the crash, but never ran out of it.
This time I got the crash just by planning the first maneouver, again on the moon.

INFO.20240610-195416.2432.log
ERROR.20240610-195416.2432.log
FATAL.20240610-200338.2432.log

Here is the zipped journal, uncompressed has a size of about 1.1gb
https://drive.google.com/file/d/1RLH4E0QaIuvUPp_kcOmWL-SBEoJU-IA7/view?usp=sharing

[kwattt]

Same issue again, this time when I changed the reference to MCMF reference, i took a screenshot before the game closed if its worth, I also do have another journal, please let me know if you also need it and will upload it.
ERROR.20240628-095242.22284.log
FATAL.20240628-111934.22284.log
INFO.20240628-095242.22284.log
WARNING.20240628-095242.22284.log

[pleroy]

The latest log is more interesting as it points towards a singularity in the integrator. If you could upload the latest journal that would be great.

[kwattt]

Here it is
https://drive.google.com/file/d/13QyGFaQ7OoLu7hzPTnHlRMT7P6La-LJR/view?usp=sharing

[pleroy]

Decoded stack track:

Principia/base/push_pull_callback_body.hpp

Line 90 in 9fb3103

auto const result = task(callback_.ToStdFunction());

Principia/ksp_plugin/interface_collision.cpp

Line 45 in 9fb3103

return plugin->ComputeAndRenderFirstCollision(celestial_index,

Principia/ksp_plugin/plugin.cpp

Lines 1029 to 1054 in 9fb3103

	Angle const& longitude)> const& radius) const {
	auto const& celestial = FindOrDie(celestials_, celestial_index);
	auto const& celestial_body = *celestial->body();
	auto const& celestial_trajectory = celestial->trajectory();
	// TODO(phl): We should cache the apsides.
	DiscreteTrajectory<Barycentric> apoapsides_trajectory;
	DiscreteTrajectory<Barycentric> periapsides_trajectory;
	ComputeApsides(celestial_trajectory,
	trajectory,
	begin,
	end,
	max_points,
	apoapsides_trajectory,
	periapsides_trajectory);
	const auto intervals = ComputeCollisionIntervals(celestial_body,
	celestial_trajectory,
	trajectory,
	apoapsides_trajectory,
	periapsides_trajectory);
	VLOG(1) << "Found " << intervals.size() << " collision intervals";
	for (auto const& interval : intervals) {
	VLOG(1) << "Collision interval: " << interval;
	auto const maybe_collision = ComputeFirstCollision(celestial_body,

Principia/physics/apsides_body.hpp

Line 458 in 9fb3103

StreamingAdaptiveЧебышёвPolynomialInterpolant<max_чебышёв_degree>(

Principia/numerics/approximation_body.hpp

Line 247 in 9fb3103

StreamingAdaptiveЧебышёвPolynomialInterpolantImplementation<max_degree>(

Principia/numerics/approximation_body.hpp

Lines 127 to 132 in 9fb3103

	Difference<Value<Argument, Function>>* const error_estimate) {
	// Try to build an interpolation over the entire interval.
	Difference<Value<Argument, Function>> full_error_estimate;
	auto full_interpolant = ЧебышёвPolynomialInterpolant<max_degree>(
	f, lower_bound, upper_bound, max_error, &full_error_estimate);
	if (full_error_estimate <= max_error ||

Principia/numerics/polynomial_in_чебышёв_basis_body.hpp

Lines 305 to 307 in 9fb3103

	MayHaveRealRootsOrDie(Value const error_estimate) const {
	if constexpr (convertible_to_quantity<Value>) {
	CHECK_LE(Value{}, error_estimate);

[pleroy]

Interesting bit of the journal:

E0630 09:21:57.489698 30228 player.cpp:107] index: 58888953
[principia.journal.serialization.CollisionNewFlightPlanExecutor.extension] { in { plugin: 1719027084128 
celestial_index: 9 sun_world_position { x: 47003878981.001343 y: -2723705084.7652864 z: -141254979389.99554 } 
vessel_guid: "bae30dc4-c136-48e2-af6d-d33eb01b18ab" max_points: 64 } }
[principia.journal.serialization.CollisionNewFlightPlanExecutor.extension] { return { result: 1746217032704 } }
E0630 09:21:57.489698 30228 player.cpp:107] index: 58888954
[principia.journal.serialization.CollisionGetLatitudeLongitude.extension] { in { executor: 1746217032704 } }
[principia.journal.serialization.CollisionGetLatitudeLongitude.extension] { out { latitude_in_degrees: 
-38.866553115290941 longitude_in_degrees: -72.801754301775276 } return { result: true } }
E0630 09:21:57.489698 30228 player.cpp:107] index: 58888955
[principia.journal.serialization.CollisionSetRadius.extension] { in { executor: 1746217032704 radius: nan } }
[principia.journal.serialization.CollisionSetRadius.extension] { }
E0630 09:21:57.489698 30228 player.cpp:107] index: 58888956
[principia.journal.serialization.CollisionGetLatitudeLongitude.extension] { in { executor: 1746217032704 } }
[principia.journal.serialization.CollisionGetLatitudeLongitude.extension] { out { latitude_in_degrees: 
-49.273857134070475 longitude_in_degrees: -93.984193263656508 } return { result: true } }
E0630 09:21:57.489698 30228 player.cpp:107] index: 58889021
[principia.journal.serialization.CollisionSetRadius.extension] { in { executor: 1746217032704 radius: nan } }
[principia.journal.serialization.CollisionSetRadius.extension] { }
E0630 09:21:57.489698 30228 player.cpp:107] index: 58889023
[principia.journal.serialization.CollisionGetLatitudeLongitude.extension] { in { executor: 1746217032704 } }
[principia.journal.serialization.CollisionGetLatitudeLongitude.extension] { out { latitude_in_degrees: 
-45.734189037157428 longitude_in_degrees: -82.364727278886278 } return { result: true } }
E0630 09:21:57.489698 30228 player.cpp:107] index: 58889025
[principia.journal.serialization.CollisionSetRadius.extension] { in { executor: 1746217032704 radius: nan } }
[principia.journal.serialization.CollisionSetRadius.extension] { }

[pleroy]

What's happening is that the computation of the altitude of the terrain of the Moon at multiple pairs (latitude, longitude) returns NaN. This should not happen (obviously) and we couldn't reproduce it. It's also worth noting that no-one has reported anything similar.

Are you by any chance using a mod that changes the terrain of the Moon?

[linecrafter]

Hello, I encountered the same issue when plotting an impact course with MCMF
Here are the journals, logs and a video recording of what I was doing:
https://drive.google.com/drive/u/2/folders/1pwulWOwjQF9JMsmfNdAhsQkYRNXRdAFY

While researching I also found a similar closed issue: #3872

Are you by any chance using a mod that changes the terrain of the Moon?

Personally i have Parallax mod installed as part of RSS-Reborn pack. If anything else is necessary for a report, please let me know.

Thank you.

[pleroy]

Interesting to hear that you are using RSS-Reborn, it could be part of the problem as it touches the PQS.

Could you please give us complete information about the installation options that you picked for RSS-Reborn, including texture resolution and other options? We'll need that to try to reproduce the bug.

[linecrafter]

Hi, i'm using Luna 32k from here https://docs.google.com/document/d/1BfiIZ8_ZEh0lrx0nsIa27Lm7O2R4e2fKPli94XHga0Q
the rest are also from document and in the screenshot.

I also have parallax configs from github https://github.com/RSS-Reborn/RSS-Reborn/wiki/Installation

[pleroy]

I was finally able to install RSS-Reborn (had some trouble, the installer was not working perfectly) but I could not reproduce the problem.

I am adding a check early in the game to detect the case where the terrain system returns NaN. If that happens, it would be easier to figure out which mod is causing trouble.

[pleroy]

See https://github.com/mockingbirdnest/Principia/wiki/Installing,-reporting-bugs,-and-frequently-asked-questions#rss-reborn.