APPEND Statement

Overview

Append RAW Alpha or Group or Literal to RAW Alpha or Group

• This is the current implementation – no changes needed here.

Append UNICODE Alpha or NATIONAL Alpha to UNICODE Alpha or NATIONAL Alpha

• The trailing UTF-32 space characters in the destination field are replaced by the UTF-32 characters from the UTF-32 source field
• If necessary, the destination field is padded with UTF-32 space characters after appending the source field.
• If necessary, the result is truncated to the length of the destination field when appending the source field.

Append Raw Alpha to UNICODE Alpha or NATIONAL Alpha

• The RAW Alpha operand must be transcoded from an 8859-1 encoding to a UTF-32 encoding before the append can be performed
• The operation is performed by appending the resulting UTF-32 characters to the destination field using the same logic as above where a UNICODE Alpha is appended to a UNICODE Alpha.

Append Literal to UNICODE Alpha or NATIONAL Alpha

• A Literal operand is consider to be a RAW Alpha and must therefore be transcoded from an 8859-1 encoding to a UTF-32 encoding before the operation be performed.
• If a literal operand contains embedded Unicode escape sequences, the corresponding UTF-32 characters will replace the Unicode escape sequences in the resulting UTF-32 literal
• The operation is performed by appending the resulting UTF-32 characters to the destination field using the same logic as above where a UNICODE Alpha is appended to a UNICODE Alpha.

Append UNICODE Alpha or NATIONAL Alpha to RAW Alpha

• Each UTF-32 character in the source field is transcoded to the equivalent 8859-1 encoded character
• If a UTF-32 character is encountered that cannot be transcoded to an equivalent 8859-1 character, a data exception error occurs.
• The append operation is performed by appending the resulting 8859-1 encoded string to the destination field using the same logic as appending a Raw Alpha to a Raw Alpha.

Append Group to UNICODE Alpha or NATIONAL Alpha

• This combination of operands will not be allowed, at least not initially.
• The process compiler should report an error
• Since we are going to redefine many Alpha PDF fields and TEMP fields to be UNICODE Alpha fields, this decision will break any application that appends a Group field to any of these fields.

Append UNICODE Alpha or NATIONAL Alpha to Group

• This combination of operands will not be allowed, at least not initially.
• The process compiler should report an error
• Since we are going to redefine many Alpha PDF fields and TEMP fields to be UNICODE Alpha fields, this decision will break any application that any of these fields to a Group field.

Test Plan

1. APPEND Unicode/National to Unicode/National
2. APPEND Raw to Unicode and National
3. APPEND Unicode/National to Raw
4. APPEND literal to Unicode/National
5. APPEND Group to Unicode/National
6. APPEND Unicode/National to Group

Bugs

1. APPEND statement with Unicode Destination and Raw Source appends extra junk to the end of the field * FIXED *
2. APPEND statement with Raw Destination and Unicode Source fails with ILF runtime error, ie, APPEND --- WIDGET LABEL 0 --- TEMP 80 * FIXED ***
3. APPEND Truncates incorrectly. APPEND 15 byte RAW alpha (contains string FIELD) to 15 character Unicode or National (contains string NORMAL) results in NORMAL FIEL in the Unicode/National field. * FIXED* March 31 engine
4. APPEND Group to Unicode/National does not produce compiler error
5. APPEND National/Unicode to Group field does not produce compiler error
6. When appending literals to National field, it accepts and displays any Unicode character, not just ones in the National set
7. When using invalid literals, ie, \U with only 4 hex digits or \u with non hex characters, process does not run (no errors, just nothing happens). Second attempt at running process crashes Appx.
8. Transcode Error thrown when it shouldn't be. Decimal 164 (Euro Symbol) is put into TEMP 1 via CNV BIN, then TEMP 1 is moved to a National field, then that field is APPENDED to Raw. That throws the error, but it shouldn't since National set is US-ASCII, which is the same as Raw (isn't it?).

-- SteveFrizzell - 2011-03-08