Hi, I investigated a bit our support for UDF 2.50 and went through an oddity on the initialization of some const structures in src/add-ons/kernel/file_systems/udf/UdfStructures.cpp Here they are // entity_ids const entity_id kMetadataPartitionMapId(0, "*UDF Metadata Partition"); const entity_id kSparablePartitionMapId(0, "*UDF Sparable Partition"); const entity_id kVirtualPartitionMapId(0, "*UDF Virtual Partition"); const entity_id kImplementationId(0, "*OpenBeOS UDF", implementation_id_suffix(OS_BEOS, BEOS_GENERIC)); const entity_id kPartitionContentsId1xx(0, "+NSR02"); const entity_id kPartitionContentsId2xx(0, "+NSR03"); const entity_id kLogicalVolumeInfoId150(0, "*UDF LV Info", udf_id_suffix(0x0150, OS_BEOS, BEOS_GENERIC)); const entity_id kLogicalVolumeInfoId201(0, "*UDF LV Info", udf_id_suffix(0x0201, OS_BEOS, BEOS_GENERIC)); const entity_id kDomainId150(0, "*OSTA UDF Compliant", domain_id_suffix(0x0150, DF_HARD_WRITE_PROTECT)); const entity_id kDomainId201(0, "*OSTA UDF Compliant", domain_id_suffix(0x0201, DF_HARD_WRITE_PROTECT)); When using and dumping these structures, I noticed they are not initialized as expected. I'm wondering whether it's really supposed to work, or whether it's a regression. The fact that this code is located in a kernel addon might be important. Thoughts? For the time being, I added a function to init these structures correctly in every hook. Bye, Jérôme