v1.1.10

Add

• Add a manual for the compiler configuration file (#665)
• Add a built-in subroutine
  • C$SLEEP (#669)

Miscellaneous

• Improve the test coverage slightly (#667, #664)

v1.1.9

Add

• Add a sub command create for cobj-idx. (#648)
  • cobj-idx create creates a new indexed file.

Fix

• Handle non-ASCII characters in EXTERNAL items and EXTERNAL file names correctly. (#624)
• Show error messages if unimplemented subroutines are used. (#628)
• Fix non-ASCII characters in external items and external file descriptors. (#629)
  • The older version does not convert － in EXTERNAL items and EXTERNAL file names into another valid character in Java.
  • The older version converts ＢＣ in EXTERNAL file names into invalid characters.

v1.1.8-hotfix1

Fix

• Fix code generations for some WRITE statements.
  • Older versions of the compiler converts some WRITE statements into invalid Java code.
• Fix the compiler to handle source code with UTF-8 encoded half-width Katakana characters correctly.

v1.1.8

Fix

• Fix a bug of duplicate record keys of indexed files

v1.1.7-hotfix1

Fix

• Fix bugs related to UTF-8
  • With older versions, DISPLAY statements output invalid UTF-8 data when the arguments are sub items of group fields.
  • With older versions, initilizing data with VALUE clauses fails in some cases.

v1.1.7

Add

• Implement the environment variable COB_TERMINAL_ENCODING.
  • If the value of COB_TERMINAL_ENCODING is UTF-8, the DISPLAY statement outputs data after converting it to UTF-8. Otherwise, it outputs raw data (typically encoded in Shift JIS).
  • If the value of COB_TERMINAL_ENCODING is UTF-8, the ACCEPT statement processes input data as UTF-8 encoded. Otherwise, it processes input data as Shift JIS encoded.
• Fix the UTF-8 compatible compiler to recognize that the size of multibyte characters in COBOL source code as 2 bytes instead of 3 bytes.
  • This change allows users to write more multibyte characters on each line of the COBOL source code.

v1.1.6

Add

• Compile UTF-8 Source Code
  • The latest version of cobj is now able to accept UTF-8 encoded COBOL and generate UTF-8 encoded Java.
  • See documentation to install the UTF-8 compatible compiler.

v1.1.5

Add

• String literals containing SJIS characters are remained as they are, not being converted into byte arrays.
  • The older versions convert string literals containing SJIS characters into byte arrays with hexdecimal integers.
    • Older versions convert "日本語" into CobolUtil.toBytes((byte)0x93, (byte)0xfa, (byte)0x96, (byte)0x7b, (byte)0x8c, (byte)0xea).
    • Latest version converts "日本語" into CobolUtil.stringToBytes("日本語").

Fix

• Fix comparisons of COMP data.
  • With older versions, comparisons of COMP data are invalid in some cases.

v1.1.4

Fixed

• Process Japanese identifiers in COBOL correctly.
  • The older versions convert some Japanese characters in COBOL identifiers to another characters in Java.
• Fix the conditions PERFORM UNTIL.
  • If COBOL source code contains divisions in conditions of PERFORM UNTIL, the older versions emit compile errors.
• Fix SEARCH statements with data specified OCCURS and DEPENDING.
  • The older versions emit compile errors when SEARCH statements with data specified OCCURS and DEPENDING
• Fix ADD statements and SUBTRACT statements.
  • In some case, the result of SUBTRACT was previously incorrect when the operands contais PIC S9(n) negative values.
  • In some case, the result of ADD was -0 instead of +0 previously.
• Fix the process of checking signs of PIC 9(n) containing spaces.
• Fix INSPECT statements.
  • The older versions change signs of some values accidentally because of the bug of INSPECT statements.

v1.1.3

Added

• Add a new option -variable
  • With -variable, cobj allows 73 or more characters for each lines
• Add documentations that describe the structure of libcobj and generated Java files
• Add Visual Studio Code Dev Container of opensource COBOL 4J

Fixed

• Fix a build error of opensource COBOL 4J on Windows 10
• Fix a build error of opensource COBOL 4J on Docker
• Fix cobj-api
  • cobj-api previously generated invalid constructors in record classses
• Fix typos of cobj-idx command
• Fix a command line option -fserial-variable

Miscellaneous

• Improve the readability of variable names containing multi-byte characters
• Convert some EVALUATE statements to switch statements