Perf/misc ds improvements

[lanarimarco]

Description

Heavy improvements for this kind of DSs:

     D                 DS
     D SMEP                         200    DIM(10000)
     D  SMEP_G_P                     02  0 OVERLAY(SMEP:01)
     D  SMEP_G_A                     02  0 OVERLAY(SMEP:*NEXT)
     D  SMEP_P01                     07  0 OVERLAY(SMEP:*NEXT)
     D  SMEP_P02                     07  0 OVERLAY(SMEP:*NEXT)
     D  SMEP_TVA                     20    OVERLAY(SMEP:*NEXT)
     D  SMEP_VAL                    100    OVERLAY(SMEP:*NEXT)

This was achieved through the implementation of the new DataStructValueBuilder interface, responsible for all operations that can change the state of DSs.
Currently, the logic for choosing the best implementation of DataStructValueBuilder is based on the DataStructureType and is provided by the new callback JarikoCallback.createDataStructValueBuilder.
The use case demonstrating the effectiveness of this approach can be viewed here (see the final assert of executeDSPERF02CompareIndexedStringBuilderVsStringBuilder function) where we prove that performance improvement can be up to 10 times.

During this work, we also noted that the decodeFromDS function, which is involved when transforming a String into a BigDecimal, performed a lot of logic not always necessary.

Before this implementation:

fun decodeFromDS(value: String, digits: Int, scale: Int): BigDecimal {
    <convert value into number if value contains not numeric chars>
    return try {
        value.toBigDecimal()
    } catch (e: Exception) {
        number.toBigDecimal()
    }
}

After this implementation:

fun decodeFromDS(value: String, digits: Int, scale: Int): BigDecimal {
    return try {
        value.toBigDecimal()
    } catch (e: Exception) {
        <convert value into number if value contains not numeric chars>
        number.toBigDecimal()
    }
}

Related to # (issue)

Checklist:

• If this feature involves RPGLE fixes or improvements, they are well-described in the summary.
• There are tests for this feature.
• RPGLE code used for tests is easily understandable and includes comments that clarify the purpose of this feature.
• The code follows Kotlin conventions (run ./gradlew ktlintCheck).
• The code passes all tests (run ./gradlew check).
• Relevant documentation is included in the docs directory.