Add __UNTAGGED_FILE_ENCODING environment variable

[IgorTodorovskiIBM]

I am added a new environment variable because the existing __UNTAGGED_READ_MODE is confusing to customers. We will use __UNTAGGED_FILE_ENCODING

Summary:

• Add __UNTAGGED_FILE_ENCODING variable for encoding-first untagged file handling
• Support semantic tokens: DETECT, IGNORE, WARN
• Support numeric CCSIDs (e.g., 1047, 819, 1208)
• Support encoding names via z/OS __toCcsid() (e.g., UTF-8, IBM-1047, ASCII)
• New variable takes precedence over legacy __UNTAGGED_READ_MODE
• Fully backwards compatible with existing __UNTAGGED_READ_MODE behavior

Implementation:

• Parser integrated into src/zos.cc (internal, no new library needed)
• Added __get_untagged_file_ccsid() getter for CCSID retrieval
• Updated __file_needs_conversion_init() to use explicit CCSID when specified
• Comprehensive man page documentation with examples

Testing:

• Integration tests verify parsing, precedence, and behavior mapping
• Tests confirm backwards compatibility with legacy variable

Limitations:

• Current conversion infrastructure supports EBCDIC (1047) ↔ ASCII (819)
• Other CCSIDs parsed correctly but conversion limited until iconv() integration