Name Utils underscoreSeparate method adds extra underscores to the beginning of strings when starting with underscore

[trynow-david]

Expected Behavior

underscoreSeparate("_nameOfThing") should result in "_name_of_thing" but actually results in "__name_of_thing"

Actual Behaviour

String prepended with an underscore result in double underscore prepended strings

Steps To Reproduce

1. Call underscoreSeparate with a string like "_nameOfThing"
2. Look at result, which is wrong.

Note I looked for unit tests for this method and there aren't any or I would have commented that you could add one to a test case. The private method beneath this method is the real problem specifically these lines

         private static String separateCamelCase(String name, boolean lowerCase, char separatorChar) {

...
                   if (c != separatorChar) {
                            if (last == separatorChar) {
                                newName.append(separatorChar);
                            }
                            newName.append(c);

Environment Information

No response

Example Application

No response

Version

4.2.0

[trynow-david]

To note the further issue this causes is that when you name a column on a micronaut-data entity something like @column(name = "_my_column_name")
it will not select the right column in querying and causes issues, which you can't fix without changing the naming strategy

[trynow-david]

For those who run into this issue, you can override the PhysicalNamingStrategy like this with the fix in it.

package com.trynow.micronaut

import io.micronaut.context.annotation.Primary
import io.micronaut.context.annotation.Replaces
import io.micronaut.data.hibernate.naming.DefaultPhysicalNamingStrategy
import io.micronaut.data.model.naming.NamingStrategy
import jakarta.inject.Singleton
import org.hibernate.boot.model.naming.Identifier
import org.hibernate.boot.model.naming.PhysicalNamingStrategy
import org.hibernate.engine.jdbc.env.spi.JdbcEnvironment

/***
 * This is an almost exact copy of the DefaultPhysicalNamingStrategy class with a fix for a bug
 * that occurs when an identifier start with an underscore.
 */
@Singleton
@Primary
@Replaces(DefaultPhysicalNamingStrategy::class)
class PhysicalNamingStrategyCustom : PhysicalNamingStrategy {
    override fun toPhysicalCatalogName(name: Identifier?, jdbcEnvironment: JdbcEnvironment?): Identifier? {
        return getIdentifier(name)
    }

    override fun toPhysicalSchemaName(name: Identifier?, jdbcEnvironment: JdbcEnvironment?): Identifier? {
        return getIdentifier(name)
    }

    override fun toPhysicalTableName(name: Identifier?, jdbcEnvironment: JdbcEnvironment?): Identifier? {
        return getIdentifier(name)
    }

    override fun toPhysicalSequenceName(name: Identifier?, jdbcEnvironment: JdbcEnvironment?): Identifier? {
        return getIdentifier(name)
    }

    override fun toPhysicalColumnName(name: Identifier?, jdbcEnvironment: JdbcEnvironment?): Identifier? {
        name?.let {
            // In this case the getIdentifier code causes double underscore to be prepended to a string,
            // so we just return the original name, and we will just handle that column name more explicitly
            if (name.text.startsWith("_")) {
                return name
            }
        }
        return getIdentifier(name)
    }

    private fun getIdentifier(name: Identifier?): Identifier? {
        return if (name == null) {
            null
        } else {
            Identifier(
                NamingStrategy.DEFAULT.mappedName(name.text),
                name.isQuoted
            )
        }
    }
}

[wetted]

underscoreSeparate("_nameOfThing") should result in "_name_of_thing" but actually results in "__name_of_thing"

Actually, the expected result should be _name_Of_Thing because the NameUtils.underscoreSeparate does not perform lowercase. I will overload the call to support it with underscoreSeparate(String camelCase, boolean lowercase)