[pdata] Change outcome of Value collections accessors misuse

[dmitryax]

Change outcome of pdata.Value.MapVal() and pdata.Value.SliceVal() functions misuse. In case of type mismatch, they now return an invalid zero-initialized instance instead of a valid detached collection.

This PR implements the approach that was agreed upon in #4775 as a replacement for #4909

[codecov]

Codecov Report

Merging #5034 (504fd0d) into main (d6b6251) will increase coverage by 0.03%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##             main    #5034      +/-   ##
==========================================
+ Coverage   91.03%   91.07%   +0.03%     
==========================================
  Files         180      180              
  Lines       10808    10808              
==========================================
+ Hits         9839     9843       +4     
+ Misses        750      747       -3     
+ Partials      219      218       -1     

Impacted Files	Coverage Δ
model/internal/pdata/common.go	94.26% <100.00%> (+0.71%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update d6b6251...504fd0d. Read the comment docs.

[bogdandrutu]

Failing tests :)

[dmitryax]

@bogdandrutu the failing test is checking the following input which unlikely to be caught in reality but I think we still need to handle it:

	orig := &otlpcommon.AnyValue{Value: &otlpcommon.AnyValue_KvlistValue{KvlistValue: nil}}
	m1 = Value{orig: orig}
	assert.EqualValues(t, AttributeMap{}, m1.MapVal())

To handle it, we would have to add another if check to the construction that you suggested, it'll look like this (Updated):

func (v Value) MapVal() AttributeMap {
	kvlist, ok := v.orig.GetValue().(*otlpcommon.AnyValue_KvlistValue)
	if !ok || kvlist.KvlistValue == nil {
		return AttributeMap{}
	}
	return newAttributeMap(&kvlist.KvlistValue.Values)
}

I run the benchmarks showing that original code of this PR appeared to be more performant:

Original:

BenchmarkAttributeValueMapAccessor-16    	1000000000	         0.7589 ns/op
BenchmarkAttributeValueArrayAccessor-16    	1000000000	         1.124 ns/op

Updated:

BenchmarkAttributeValueMapAccessor-16    	1000000000	         1.102 ns/op
BenchmarkAttributeValueArrayAccessor-16    	838671705	         1.350 ns/op

[dmitryax]

I've brought back the original code for map and slice value getters for now