updated grammar, wrapping_visitor and tests for the changes

[FitashUlHaq]

updated grammar, wrapping_visitor and tests for the changes

[ArmenSl]

Hey Fitash can you do a quick review of the PR #586

The OCL grammar changes collide. Both PRs edit BOCL.g4 plus all the regenerated ANTLR artifacts, and both independently invent the same operators: asSet, intersection, and a dot-form allInstances() — but with different rule names (#arrowAsSet vs #arrowasSet, #dotAllInstances vs #allInstancesExp). Whichever merges second gets unresolvable conflicts in generated parser files and a semantically diverged grammar. Worse, #586 removes the existing Type::allInstances() syntax entirely (breaking saved models that use it), while #585 keeps it. If you can arbitrate one unified grammar; the cleanest path is to extract the grammar/parser changes from #586 into a joint OCL PR (or land #585 first and have the other branch rebase on it).

[jcabot]

For sure, accepting a PR that breaks existing code already merged long time ago makes no sense unless there is a very strong reason, which doesn't seem to be the case (talking about the allInstances case here)

[ArmenSl]

LGTM. Went through the grammar and the wrapping visitor - keeping Type::allInstances(). Good test coverage on the new constructs too. Two small notes, neither blocking: the new rule labels (#arrowasSet, #isunique) don't follow the casing of the existing ones. Let's merge this one first so #586 can rebase its Alloy work on top of this grammar.