• [ ] Recheck this ticket and clean it up.
• [ ] Check grammar for all repos.
  • [ ] Create equivalent to repo process, but for projects.
  • [ ] Create net.splitcells.project.process project, in order to process such projects.
  • [ ] Move net.splitcells.project.files.standard content into net.splitcells.project.process and delete original. This way a generic project exists for such things and improves the existing project name, which is too long.
• [x] Create alternative implementations for all Domable#toDom instances.
• [x] Fix grammar error, that causes ANTLR4 to not show, what the grammar error. -> Found a workaround: The rule (.)*? Keysymbol_at Keyword_JavaLegacyArtifact (.)*? EOF breaks every error message. Disabling this rule for a file in question creates normal error messages.
  • [x] Print out lexer's tokens result on error. Maybe that is enough? https://stackoverflow.com/questions/70701595/is-there-a-way-to-extract-tokens-in-order-with-antlr
  • [x] Maybe try out ANTLR4 UI first.
• [ ] Remove usage of `org.w3c, because of the problems associated with this lib.
  • [ ] Remove usage of Domable#toDom except inside Domable#toDom implementations.
    • [ ] toDom is used in order to implement toDom like functions like Constraint#graph.
    • [ ] Remove other toDom usage.
      • [ ] Update Rater#argumentation(GroupId, Table).
        • [x] Test whether Rater#argumentation(GroupId, Table) is used at all. -> No it is not really used. So We can replace the method in one go or commit. We could even delete it with no issue, but it may have a need, that is forgotten, so we just convert it instead. Rater#argumentation(GroupId, Table) does not have anything to do with Constraint#naturalArgumentation() and co.
        • [x] Create test for argumentation. -> Is not needed
        • [ ] Remove Rater#argumentation(GroupId, Table) from Rater interface.
        • [ ] Convert all Rater#argumentation(GroupId, Table) overrides to normal methods, that use Perspective instead.
      • [ ] Update Solution#createAnalysis.
        • [ ] Create a test for this.
      • [ ] Update SolutionView#createAnalysis.
        • [ ] Create a test for this.
      • [ ] Update SolutionView#toLinesFodsAnalysis.
        • [ ] Create a test for this.
      • [ ] Update History#toAnalysisFods.
        • [ ] Create a test for this.
      • [ ] Update ProjectsRendererI#createLayout.
        • [ ] Create a test for this.
  • [ ] Remove Domable#toDom instances.

Sub Tasks

For this to work, the source code also have to comply with the grammar, when this issue is done.

• [x] Ensure that Dem complies with new custom Java grammar.
• [ ] Make grammar clean, so it's straight forward to process for translations and to understand.
• [x] Detailed Javadoc grammar. -> task is migrated into repo and will be done later.
• [x] Review grammar.
• [x] Disable checks for bootstrap code. -> In the grammar it is called legacy code.
• [x] Split up Dem into core and merger.
• [x] Add grammar check to CI build. -> Source code check was added to capabilities check.
• [x] Fix usage fragments. Currently, they often do not work and require Tokens for matches and the respective Regex do not work.
• [x] Use mixed grammar, which makes it possible to used fragment tokens more effectively (currently these are broken). -> Split up grammars are better, because lexer work in ANTLR more like validator if one ignores string patterns.
• [x] Remove explicit whitespace matches.
• [x] Document how to do non combined grammar.
• [x] Document how to do combined grammar. -> Not needed.
• [x] Use upper case for Keywords in Lexer.
• [x] Use constants in parser instead of tokens, in order to make these 2 more independent of each other. -> This does not have a real advantage.
• [x] Reformat grammar source code.
• [x] Create core project for Dem in order to test interface and implementation splitting.
• [x] Test speed up via using faster alternatives for data structures: Maps, Sets, Lists -> What is this?
• [x] Only allow foreign code import like the Java standard library in Java legacy code with some exceptions to the Java standard library like java.util.List. Currently, only Java standard library imports are blocked.
• [x] Let all source code comply with this Java subset grammar.

Tasks For Future Tickets

• [ ] Translate code to dlang with working tests in order to be sure, that everything is handled by the grammar: https://tomassetti.me/how-to-write-a-transpiler/

• [ ] Remove direct usage of Path (because resolve method allows resolving absolute paths), Files (because changing the backend does not seem to be viable) and similar in the future.

• [ ] Make main part GPL compliant and only extensions non GPL compliant: https://github.com/www-splitcells-net/net.splitcells.network/issues/88

  • [ ] Note on planned future license change.

• [ ] Create basis for Translation. Use JavaParser for translation only instead of the Java grammer, in order to support some standard Java things/classes/method via name resolution without creating own wrapper interfaces and objects.