[sharanggupta]

ParsingUtils.split(source, toLower) lower-cases each camel-case part with String.toLowerCase(), which follows the JVM's default locale. Under a Turkish default locale (tr-TR, the default on any Turkish OS or user account) an uppercase I becomes a dotless ı (U+0131), so every consumer of splitCamelCaseToLower / reconcatenateCamelCase produces identifiers that do not exist:

// -Duser.language=tr -Duser.country=TR, spring-data-commons 4.1.1
ParsingUtils.reconcatenateCamelCase("InvoiceItem", "_");   // "ınvoice_ıtem"
ParsingUtils.reconcatenateCamelCase("invoiceId", "_");     // "invoice_ıd"
DefaultNamingStrategy.INSTANCE.getTableName(InvoiceItem.class); // "ınvoice_ıtem"  (Spring Data JDBC / R2DBC)

Affected callers: NamingStrategy.getTableName/getColumnName in Spring Data Relational (default SQL table and column names), SnakeCaseNamingStrategy in Spring Data Cassandra, CamelCaseSplittingFieldNamingStrategy (document field names) and XmlRepositoryConfigurationSource. The camel-case split itself uses \p{Lu} and is locale-independent; only the lower-casing is affected.

This uses Locale.ROOT, as AotContext, RepositoryBeanDefinitionBuilder and RepositoryConfigurationExtension already do in this module, and adds a unit test that switches the default locale to tr-TR around the conversion.

• You have read the Spring Data contribution guidelines.
• You use the code formatters provided here and have them applied to your changes. Don’t submit any formatting related changes.
• You submit test cases (unit or integration tests) that back your changes.
• You added yourself as author in the headers of the classes you touched. Amend the date range in the Apache license header if needed. For new types, add the license header (copy from another file and set the current year only).

[christophstrobl]

@sharanggupta thank you for reporting that issue!

[christophstrobl]

From a quick glimpse across the codebase, ParsingUtils is not alone in calling lowercase without an argument.

[christophstrobl]

Regarding the suggested code changes, I think we need a more differentiated approach.
We cannot simply change behaviour that has been this way for more than a decade.

[sharanggupta]

Thanks, @christophstrobl. Understood that behaviour this old should not simply flip. I am happy to rework this along whatever direction the team settles on, for example one case-folding helper that the other call sites you spotted could share, or to close this in favour of a team-driven change.