Fix whitespace-only implicit and explicit slots during folding

[ghabriel25]

Problem

When Blaze folds a component with whitespace-only content:

<x-button> </x-button>

the whitespace was incorrectly treated as actual slot content. Blaze would create a default slot placeholder, causing the component to receive a non-empty ComponentSlot during rendering.

This could cause components to behave differently when Blaze was enabled.

Changes

Foldable::setupSlots() now checks the actual rendered loose content before creating an implicit default slot.

Whitespace-only content no longer creates a slot placeholder, while actual content continues to behave as before.

Result

• Whitespace-only implicit and explicit slots are treated as empty during folding.
• Components no longer see a synthetic non-empty slot for whitespace-only content.
• Actual slot content and explicit slots remain unchanged.
• No changes to Laravel's ComponentSlot semantics or normal non-Blaze rendering.

Related PR: livewire/flux#2821
Fixes: livewire/flux#2816

[github-actions]

Benchmark Result: Default

Attempt	Blade	Blaze	Change
#1	364.10ms	14.67ms	96%
#2 *	368.34ms	15.12ms	95.9%
#3	356.57ms	14.94ms	95.8%
#4	358.73ms	14.92ms	95.8%
#5 *	364.39ms	15.19ms	95.8%
#6	357.17ms	14.78ms	95.9%
#7	356.50ms	14.82ms	95.8%
#8	356.55ms	14.86ms	95.8%
#9	359.67ms	14.90ms	95.9%
#10	358.74ms	14.89ms	95.8%
Snapshot	360.77ms	14.82ms	95.9%
Result	357.95ms (~)	14.88ms (~)	95.8% (~)

_{Median of 10 attempts (* = outlier, excluded from result), 5000 iterations x 10 rounds, 47.23s total}

_{To run a specific benchmark, comment /benchmark <name>
attributes, aware, class, default, forwarding, merge, named-slots, no-attributes, slot, compilation}

[ganyicz]

Thanks @ghabriel25!

This looks good, however the same issue applies to any slot, not just the one created from loose content.

Can we check for empty whitespace whenever we're creating a placeholder for a slot?

[ghabriel25]

@ganyicz Sure but its already 2AM here so I'll push it later.

[ghabriel25]

@ganyicz Let me know if it needs additional test 👌

[ghabriel25]

/benchmark slot

[github-actions]

Benchmark Result: Slot

Attempt	Blade	Blaze	Change
#1	97.27ms	6.90ms	92.9%
#2	94.01ms	6.74ms	92.8%
#3	97.12ms	6.75ms	93%
#4	93.82ms	6.87ms	92.7%
#5	95.75ms	6.71ms	93%
#6	98.49ms	6.72ms	93.2%
#7 *	96.84ms	7.61ms	92.1%
#8	96.40ms	6.86ms	92.9%
#9	96.80ms	6.72ms	93.1%
#10	95.76ms	6.71ms	93%
Snapshot	95.22ms	6.75ms	92.9%
Result	96.40ms (~)	6.74ms (~)	93% (~)

_{Median of 10 attempts (* = outlier, excluded from result), 5000 iterations x 10 rounds, 14.04s total}

_{To run a specific benchmark, comment /benchmark <name>
attributes, aware, class, default, forwarding, merge, named-slots, no-attributes, slot, compilation}

[ganyicz]

Thanks!