Save and load the current planned hyperjump in the save file

[JonBooth78]

The planned hyperspace route, with all stops is saved along with the save game, and restored on load.

[impaktor]

This fixes #5178 ?

[JonBooth78]

@impaktor , only half of it; I'll get onto the rest later - thanks for pointing it out

[sturnclaw]

Especially given that you're planning on serializing more sector view state, I'd recommend that you marshal the serialized data through a Serializer:Register() call in map-sector-view.lua instead of registering the serialization callback in hyperjump-planner.lua.

The hyperjump planner / sector view split is slightly unconventional / unfinished in that the hyperjump planner is not quite a re-usable class that the sector map owns an instance of (yet), but it's close enough that I'd recommend applying that same mental model.

This is not strictly required, but it will make it easier to clean up / finish encapsulating the hyperjump planner widget in the future.

[sturnclaw]

Hmm... the more I look at this code the more I think the serialization should be entirely handled in map-sector-view.lua. 99% of this code is manipulating the C++ sector view object itself, and the only thing that involves the hyperjump planner widget is a one-off call in an onGameStart event to hyperjumpPlanner.buildRouteList().

[JonBooth78]

@Web-eWorks , I think your right; I also suspect that the call to hyperjumpPlanner.buildRouteList() is also redundant as I think it gets called on entry to the screen; I just wanted to be super-safe.

I felt that as this was all to do with the route-planner, it made sense to live in here, although I agree, it really has little to do with that element, which is a UX one.

I'll have a crack at moving it to the suggested place and add the other missing bits of #5187

[JonBooth78]

@Web-eWorks , ready for re-review, all seems to be complete and working, with a bonus bug-fix thrown in at the end :-)

[JonBooth78]

oh snap, take that back found a bug

[JonBooth78]

OK, fixed that bug (doesn't look like I introduced it, oddly) and then clang formatted a file I touched on the way through. @Web-eWorks now ready to go

[sturnclaw]

Two very minor tiny nitpicks but the code looks good otherwise. Haven't had time to test this branch, so it's provisionally approved pending a playtest sometime in the next 24 hours.

[sturnclaw]

OK, fixed that bug (doesn't look like I introduced it, oddly) and then clang formatted a file I touched on the way through.

This is quite the ancient bug oddly enough (4 years old!) but was completely benign up until yesterday when I merged #5724. I had the feeling something was odd with the onEnterSystem callback signature but rationalized it away due to lack of time. Thanks for catching that!