I am pleased to submit this pull request to resolve the issue mentioned in the comments of #294. While #294 addressed the problem of "logical and" being mistakenly parsed as "cast + dereference", it inadvertently caused invocation expressions with generics to be parsed as binary expressions. So #294 was reverted and the test was removed. A new test about generic invocations was added in the commit fc6be5ae61c8788e0c028e757f132d67bfc538f5
To address this issue, I have added precedence to the generate type argument list <XXX> to make it bind more tightly. With this update, the parser now passes both the test added in #294 and the test added in the commit fc6be5ae61c8788e0c028e757f132d67bfc538f5
I am pleased to submit this pull request to resolve the issue mentioned in the comments of #294. While #294 addressed the problem of "logical and" being mistakenly parsed as "cast + dereference", it inadvertently caused invocation expressions with generics to be parsed as binary expressions. So #294 was reverted and the test was removed. A new test about generic invocations was added in the commit fc6be5ae61c8788e0c028e757f132d67bfc538f5
To address this issue, I have added precedence to the generate type argument list
<XXX>
to make it bind more tightly. With this update, the parser now passes both the test added in #294 and the test added in the commit fc6be5ae61c8788e0c028e757f132d67bfc538f5