Add full CCSID via a unified getFileInfo component

[bobcozzi]

Feature Request: Unified getFileInfo Component for File Metadata with CCSID

Summary

Create a new component that retrieves comprehensive file metadata including CCSID information in a single call for both IBM i source file members and IFS streamfiles, using appropriate IBM i APIs that expose CCSID and field-level attributes.

Background

Currently, the extension retrieves file metadata through multiple separate API calls:

• Member metadata: getMemberInfo() via the GETMBRINFO.SQL component (wraps QUSRMBRD)
• CCSID information: getAttributes() using the PASE attr command for IFS files
• Variant character handling: Separate checks throughout the codebase

Important limitation: The current QUSRMBRD API used in GETMBRINFO does not return CCSID information for files or fields, which is critical for proper variant character handling.

This fragmented approach requires multiple round-trips to retrieve complete file information, and the lack of CCSID data at file-open time causes issues with variant character support.

Proposed Solution

Create a new getFileInfo component that retrieves comprehensive file metadata in a single call, using the appropriate IBM i APIs:

Recommended API: QUSLFLD (List Fields)

Primary recommendation: Use the QUSLFLD API which:

• Returns all columns (fields) for a database file
• Includes all field attributes including CCSID for each field
• Is easier to implement than QDBRTVFD
• Provides consistent field-level metadata

Alternative: QDBRTVFD (Retrieve Database File Description)

• Returns both File Format CCSID and individual field CCSID
• More complex to implement but provides additional file-level information

Implementation Approach

New SQL Component: GETFILEINFO.SQL

Create a table function or stored procedure that wraps:

• QUSLFLD API for retrieving field information with CCSID attributes
• Continue using QUSRMBRD for member-level metadata (timestamps, description)
• For IFS files: Continue using PASE attr command for CCSID

Suggested Return Type:

interface FileInfo {
  // Member/File identification
  library?: string;
  file?: string;
  name: string;
  extension: string;
  created?: Date;
  changed?: Date;
  description?: string;
  
  // File-level attributes
  ccsid: number;              // File format CCSID
  recordLength?: number;      // For members/database files
  fileType: 'member' | 'streamfile';
  isSource: boolean;
  
  // Field-level information (for database files)
  fields?: Array<{
    name: string;
    type: string;
    length: number;
    decimal?: number;
    ccsid?: number;           // Field-specific CCSID
    description?: string;
  }>;
  
  // For IFS files
  path?: string;
  size?: number;
  
  // Additional metadata
  asp?: string;
  owner?: string;
}

This feature would particularly benefit external extensions that consume the Code for IBM i API and need to properly handle variant characters in syntax highlighting, code analysis, and transformation tools. The availability of CCSID information at file-open time is crucial for proper character encoding and variant character support throughout the IBM i development ecosystem.
Let me know if you want me to compose the wrapper for the QUSLFLD API for this feature of it is is no longer needed.

[bobcozzi]

I put together this SQL statement to get the field's CCSID for any database file. Just need to replace the compare value on the schema and table name WHERE clause:

SELECT table_name,
       table_schema,
       column_name,
       system_table_name,
       system_table_schema,
       system_column_name,
       ccsid,  -- CCSID for "this" Field
       data_type,
       storage buffer_SIZE,
       length as GENERAL_LENGTH,
       charLen   AS "CHAR_LENGTH",
       precision AS "DEC_LENGTH",
       numeric_scale as "DEC_POS",
       CASE
         data_type
         WHEN 'INTEGER' THEN 'INT(' CONCAT
               CASE LENGTH
                 WHEN 1 THEN '3'
                 WHEN 2 THEN '5'
                 WHEN 4 THEN '10'
                 WHEN 8 THEN '20'
               END CONCAT ')'
         WHEN
           'DECIMAL'
           THEN
             'PACKED(' CONCAT PRECISION CONCAT ', ' CONCAT numeric_scale CONCAT ')'
         WHEN
           'NUMERIC'
           THEN
             'ZONED(' CONCAT PRECISION CONCAT ', ' CONCAT numeric_scale CONCAT ')'
         WHEN 'TIMESTMP' THEN 'TIMESTAMP'
         ELSE TRIM(data_type) CONCAT '(' CONCAT charLen CONCAT ')'
       END AS Field_Defn,
       CASE WHEN labeltext is NULL THEN COLUMN_HEADING else LABELTEXT end as "TEXT"
  FROM qsys2.syscolumns FFD
  WHERE TABLE_NAME = 'SYSCOLUMNS'  -- Use ${filename} or similar
        AND table_schema = 'QSYS2'    -- Use ${library} or similar