Add StartMainPassPostProcessing node for ordering

[atlv24]

Objective

• Fixes InvalidNode(Bloom) crash if bevy_post_process is off #20923
• Supersedes Fix volumetric fog render graph crash when bevy_post_process feature … #20931

Solution

• Add a StartMainPassPostProcessing to order against

Testing

• bloom example

[IceSentry]

I feel like maybe this should be ordered before EndMainPass if it's considered to be part of the main pass? I'm just not sure where it should go. I guess this works well enough though.

[atlv24]

I don't think its considered to be part of the main pass

[IceSentry]

I was 100% basing this based on the comment that was above 😅.

[JMS55]

Needs migration guide imo

[github-actions]

It looks like your PR is a breaking change, but you didn't provide a migration guide.

Please review the instructions for writing migration guides, then expand or revise the content in the migration guides directory to reflect your changes.

[atlv24]

I dont believe this needs a migration guide, theres no breaking change here

[hukasu]

Should the Bloom node be made to be after this new StartMainPassPostProcessing?

bevy/crates/bevy_post_process/src/bloom/mod.rs

Lines 81 to 85 in ec735b9

	.add_render_graph_node::<ViewNodeRunner<BloomNode>>(Core3d, Node3d::Bloom)
	.add_render_graph_edges(
	Core3d,
	(Node3d::EndMainPass, Node3d::Bloom, Node3d::Tonemapping),
	)

[hukasu]

Controvesial: all of the post processing nodes should be moved to a PostProcess3d on the bevy_post_processing

[atlv24]

the issue with moving it out is that you then need to depend on bevy_post_processing to order relative to post processing. we really just need a better more powerful render graph abstraction, id like to land this fix and figure that out next cycle

[IceSentry]

The ordering of volumetric fog does feel a bit weird since it's supposed to be at the start of post processing but is ordered before the new label. This is more of an issue with the render graph abstraction though.