Thanks for the report. I can reproduce the issue in 1.1.77 however it looks like it was already fixed for 1.2.0 so i am moving it to "fixed for 1.2"

Thanks for the report however the crashes that happen are all caused by a corrupted save file: the electric network reports as belonging to a surface with index 2 while the nauvis (the only surface) has an index of 0. The only places where this index can change intentionally is when an ElectricNetwo...

Can you provide a save file that you are unable to load?

I was looking into that desync report again but there is literally nothing that i could identify as interesting. If you have some reproduction steps i would like to know the steps but other than that i suspect it may be a 1/0 magic.

This issue is really annoying. I tried making inserter to not go to sleep when furnace is not full. That makes one other test to fail which expects an inserter goes to sleep when trying to insert iron plate to a stone furnace before a recipe for steel is enabled because that test checks if the inser...

Ok i found and fixed the issue. It was related to technology effects reset in which case a furnace would stop crafting and would check what recipe it should be crafting based on the items in the ingredients slot. In this case because ingredients inventory was empty it would stay without a recipe but...

I just verified and its not even because of the json migration. The issue also happens when loading such save file in a prototypes changed mode.

Can the values be checked during the saving? Checked against what? The check that was throwing in this case was when loading a boolean which is basically 1 bit of data saved inside of 1 byte so the upper 7 bits have to be 0. I could check if bool being saved has those upper 7 bits cleared but thats...

The issue is not completely over. We can say: well, it is my hardware or cosmic rays, so what can I do to prevent next gamesave loss after another XXX hrs? I cannot give you any answers here. If we compute 2+2 and the processor gives answer of 5 (CPU error) then its outside of our assumptions of ho...

Thanks for the report. Issue is now fixed for 1.2.0. I do not want to back-port this fix to 1.1.x.

Fine. It will collide with rail signals starting from 1.1.78.

I just looked at this and the inserter is correct in saying that its waiting for the source item. It is the furnace's fault because it reports as not being full for inserters because it has no recipe in that point in time so it cannot check being full due to product overload. It can however answer t...

Thanks for the report. Issue is now fixed for 1.1.78. Issue here was related to goal_pressure_ratio setting which is used by one of the node heuristics. When a value of that setting was changing the heuristic would compare nodes differently and that was causing a server to have a heap that was not f...

Thanks for the report. Pipe-to-ground placement that finishes a pair is expected to remove all pipes in between if there are pipes on every tile between pipe-to-ground ends and pipes in between have no perpendicular connections. This behavior is primarily useful when you have long sections of pipes ...

I took a look at this save file and the error is correct: when loading level.dat at offset 249546864 the left line of the se-space-transport-belt on Nauvis Orbit at position {-17.5, 52.5} said hasRapidInteractions="64" where the only allowed values are 0(false) and 1(true). It does not loo...

Thanks for the report. Issue is now fixed for 1.1.78.

Thanks for the report. I was able to reproduce one issue which with some additional simple actions would lead to various crashes similar in stacktrace to those reported so i believe its the same issue as reported. Issue is now fixed for 1.1.78.

Unfortunately i cannot download the desync report anymore.

Thanks for the report. If your intention is to not select any entities you can use the "nothing" selection_mode flags.

Thanks for the report. Unfortunately i am not considering this to be a bug, this is an inherent limitation of the way we perform fluid flow and the only way to fix this would be to rework entire fluid flow logic which is not going to happen any time soon.