Fast path in merge() for single class default

[SanderMuller]

The most common usage of merge() in Blade components is $attributes->merge(['class' => 'some classes']), a single string default for the class key. This adds an early return for that exact pattern, bypassing the general-purpose loop that separates appendable/non-appendable attributes, iterates defaults, and calls array_merge().

What changed

A fast path at the top of BlazeAttributeBag::merge() detects when:

1. Exactly one default is passed
2. The key is class
3. The value is a string

When all three hold, it merges the class directly and returns, skipping the escape loop, the appendable/non-appendable separation, and the final array_merge().

Why this works

The general path iterates $this->attributes to classify each key as appendable or not, runs a second loop to resolve appendable values against defaults (with AppendableAttributeValue instanceof checks), and finishes with array_merge(). The fast path skips one e() call, one array copy, one string concatenation.

[SanderMuller]

/benchmark merge

[github-actions]

Benchmark Result: Merge

Attempt	Blade	Blaze	Improvement
#1	233.96ms	13.95ms	94%
#2	238.00ms	14.08ms	94.1%
#3	236.28ms	13.91ms	94.1%
#4	237.18ms	14.51ms	93.9%
#5	233.56ms	14.24ms	93.9%
#6	237.98ms	13.94ms	94.1%
#7	238.56ms	14.18ms	94.1%
#8	231.87ms	14.16ms	93.9%
#9	235.71ms	13.91ms	94.1%
#10	230.51ms	14.40ms	93.8%
Snapshot	235.47ms	16.39ms	93%
Result	236.00ms (~)	14.12ms (-13.8%)	94% (+1%)

_{Median of 10 attempts, 5000 iterations x 10 rounds, 31.98s total}

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

[github-actions]

Benchmark Result: Default

Attempt	Blade	Blaze	Improvement
#1	349.46ms	15.95ms	95.4%
#2	343.77ms	15.68ms	95.4%
#3	346.25ms	15.89ms	95.4%
#4	349.80ms	16.27ms	95.3%
#5	348.56ms	16.42ms	95.3%
#6	351.51ms	16.14ms	95.4%
#7	347.67ms	15.84ms	95.4%
#8	351.74ms	16.09ms	95.4%
#9	352.71ms	15.88ms	95.5%
#10	346.73ms	16.03ms	95.4%
Snapshot	345.45ms	15.95ms	95.4%
Result	349.01ms (~)	15.99ms (~)	95.4% (~)

_{Median of 10 attempts, 5000 iterations x 10 rounds, 45.99s total}

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