[I] Documented lowercase KahaDB values and AMQP charset names fail to parse under a Turkish default locale (broker cannot start) [activemq]

["sharanggupta (via GitHub)" <gi...@apache.org>]

sharanggupta opened a new issue, #2741:
URL: https://github.com/apache/activemq/issues/2741

   `MessageDatabase` parses its string-valued KahaDB settings with `Enum.valueOf(value.trim().toUpperCase())` and no `Locale`:
   
   - `preallocationScope` / `preallocationStrategy` (parsed in `createJournal()` during `start()`)
   - `setJournalDiskSyncStrategy(String)`
   - `setPurgeRecoveredXATransactionStrategy(String)`
   
   `String.toUpperCase()` follows the JVM default locale. On a JVM whose default locale is Turkish (`tr-TR`; `-Duser.language=tr -Duser.country=TR` reproduces it anywhere) an `i` upper-cases to a dotted `İ` (U+0130), so every documented lowercase value containing an `i` is rejected and the broker fails to start:
   
   ```
   <kahaDB journalDiskSyncStrategy="periodic" .../>   -> IllegalArgumentException: No enum constant ...Journal.JournalDiskSyncStrategy.PERİODİC
   purgeRecoveredXATransactionStrategy="commit"       -> ...PurgeRecoveredXATransactionStrategy.COMMİT
   preallocationScope="entire_journal_async"          -> ...Journal.PreallocationScope.ENTİRE_JOURNAL_ASYNC
   preallocationStrategy="sparse_file"                -> ...Journal.PreallocationStrategy.SPARSE_FİLE
   ```
   
   Reproduced against the released `activemq-kahadb-store` 6.3.2 (`always`, `never`, `none`, `zeros`, `rollback` work). The KahaDB reference page documents exactly these lowercase values (`journalDiskSyncStrategy = always | periodic | never`, `preallocationScope = entire_journal | entire_journal_async | none`, `preallocationStrategy = sparse_file | os_kernel_copy | zeros | chunked_zeros`). The code defaults are the enum names, so only a user who copies a documented value is affected.
   
   Same shape elsewhere in the code base:
   
   - `activemq-amqp` `AmqpContentTypeSupport.parseContentTypeForTextualCharset` lower-cases the type/subtype and upper-cases the `charset` parameter with the default locale, so under `tr-TR` AMQP text messages with `content-type: text/plain;charset=iso-8859-1` (also `us-ascii`, `windows-1252`) are rejected with `Illegal charset: İSO-8859-1`, and an upper-cased `TEXT/PLAIN` is no longer recognised as textual.
   - `activemq-jdbc-store` `MySqlJDBCAdapter.setStatements` compares `engineType.toUpperCase()` with `"INNODB"`, so `engineType="innodb"` turns the `LOCK TABLE` statement on under `tr-TR`.
   
   `IntrospectionSupport` already uses `Locale.ENGLISH` for the same reason (AMQ-2608).
   
   Fix: `toUpperCase(Locale.ENGLISH)` / `toLowerCase(Locale.ENGLISH)` at the six sites, plus regression tests that set a `tr-TR` default locale around `MessageDatabaseTest` (the four KahaDB setters, parsed through `createJournal()`) and `AmqpContentTypeSupportTest`. Affects 6.3.2 and `main`. I have the change ready and will open the PR against `main`.
   


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: gitbox-unsubscribe@activemq.apache.org.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: gitbox-unsubscribe@activemq.apache.org
For additional commands, e-mail: gitbox-help@activemq.apache.org
For further information, visit: https://activemq.apache.org/contact