V1.2.4

[mayanje]

V1.2.4
This release 1.2.4 mainly fixes various issues and Code cleanup for better error diagnostics, better performances and better code generation.

Code generation

• A WHEN OTHER was missing in our generated template code (Codegen : when other are missing #1089)
• Better emitted code for IF instruction alignment. (Codegen Merge condition and align IF #1018)
• Mixed code generation can emit and read source using another metacharacter than £, this in case of another charset encoding. (Transform : decode fails if PART2 of a file is modified outside RDZ #1115).

Code Cleanup, Coding guide & Grammar simplification

• Port of the Cobol directives preprocessor from ANTLR to CUP (Migrate from Antlr to CUPS for preprocessor phase #870)
• Adding Coding guide rules (Misc code cleanup #1114)
• Various Grammar simplifications. (CobolWords, CobolExpressions and CodeElement Grammar refactoring #586).

Performance improvements and error diagnostics

• Emit a warning diagnostic when a WHEN OTHER clause is missing inside an EVALUATE statement (Detect missing "when other" in evaluate #1098)
• Add a warning diagnostic when a IF statement is missing a END IF clause (Detect missing end-if #1102)
• New error diagnostic for PICTURE declarations that cannot have children. (Data declaration that contains picture, type or certain usage cannot have children #729)
• Error Diagnostics set on instructions are now more precise by specifying the target symbol reference or definition. (Diagnostics set on Node should be more precise on specify which Token, SymbolReference, SymbolDefinition, ... is in error #1091)
• Symbol Table lookup improvement for better performances. About 20% gain in incremental mode.(SymbolTable.GetCustomTypesSubordinatesNamed is very slow #1033)
• Make TypeCobol Node Linker visitor only visit nodes that contains data declaration (TypeCobolLinker : don't visit Node that cannot contains type usage #1039).

Bug fix

• Temporary fix to display TALLY by using the source text. (TALLY is changed to ALL in error message #1120)
• Add an error diagnostic when a pointer name used in instruction set xxx up by / down by exceed 22 characters (Incremented pointer must be limited to 22 chars and not inside copy #1011)

Language server issues and completion

• Avoid a Goto to Symbol definition for a symbol located in a COPY. (LS : GoToDefinition can locate a Token inside a Copy instead of the main document #1116).

Build configuration

• Remove configuration Net40 on AppVeyor. (Deactivate builds _Net40 in AppVeyor #1105).

CSCup parser generator

• Make the CSCup parser generator support the #use directive for additional grammar inclusion. (Add an include directive in CSCup to allow sub grammars inclusion. #1000)

Other

• Ensure with LSR unit tests that the LSR process is killed after each test. (Incremental : Create unit tests for each step of parser #849).