pure pilot
I'm chaining two systems after each other like (place_shape, update_boundary).chain() and while update_boundary certainly runs after place_shape, it seems like the Command changes made by the first system (e.g. despawn_entity) are not yet applied when the second system runs. Is this the case, or is there some other bug in my code?
If it is the case, how do I schedule these systems so the second one runs after the commands of the first have been applied to the world?
sharp meadow
There is a system, apply_system_buffers, that applies the Commands as a hard sync point (cannot run in parallel with other systems because exclusive world access)
sharp meadow
Bevy already includes some of these in its system sets, so it's easiest to add one system to CoreSet::PreUpdate and its commands are applied by CoreSet::Update (where systems are added by default)
But you can chain (place_shape, apply_system_buffers, update_boundary).chain() too
pure pilot
Ah! I tried sticking apply_system_buffers inbetween them and it worked
Probably putting them in separate stages (is the term I was trying to find in the examples e.g. CoreSet::PreUpdate) is better
I guess since the stageless update they are not called stages anymore. Either way I couldn't find an example that dealt with this
and with how much this has changed in bevy 0.10 it's hard to find on google
thanks for the help!
sharp meadow
There is a notable difference in that chaining apply_system_buffers in different places does reduce bevy's ability to parallelize work, so separating tasks into sets with a sync point between them tends to work better IMO
pure pilot
How do I add systems to PreUpdate? .in_schedule(CoreStage::PreUpdate) doesn't seem to do it
sharp meadow
in_base_set(CoreSet::PreUpdate)
crimson nymph