Fix segfault in the Breadcrumbs system

[azeey]

The segfault occurred because a pointer to a component is invalidated by the creation of new components. This doesn't always happen, but if a component type had 99 items, and we take get a pointer to a component, as soon as new components of the same type are created, the underlying std::vector containing the component gets expanded invalidating our pointer. The underlying std::vector gets expanded every 100 components.

To test, run the attached world (extension is .txt because github doesn't allow .sdf files). It should crash after 2 sim seconds.

debug_breadcrumbs_crash.txt

[nkoenig]

The segfault occurred because a pointer to a component is invalidated by the creation of new components. This doesn't always happen, but if a component type had 99 items, and we take get a pointer to a component, as soon as new components of the same type are created, the underlying std::vector containing the component gets expanded invalidating our pointer. The underlying std::vector gets expanded every 100 components.

This feels like an issue we should resolve in ECS, correct? I could image this biting us and others in the future. If this is the case, could you create an issue for it?

[chapulina]

+1 to @nkoenig 's suggestion to ticket an issue.

And a mental note to always get Data() asap.

[iche033]

confirmed that this fixes the crash.

I think this exposed a bigger issue with components that needs to be fixed? I imagine there are other places in the code base that do something similar with these pointers.

Edit Just refreshed my page and saw the previous comments. Yes we should ticket an issue for this.

[azeey]

I can create an issue, but I'm not sure what can be done because this is inherent to using a contiguous data container for the components. We can move to using something like std::list, but that defeats the performance benefits of ECS.