Let me draw your attention to
https://github.com/w3c/qt3tests/issues/63
if you haven't seen it.
I believe that under the rewritten 4.0 tokenization rules, the expression `map{a?b:true()}` is clearly a syntax error, but it is in the QT3 and QT4 test suites as a passing test. I think the test was written to establish how the "longest token consistent with the EBNF" rule in the 3.1 spec should be interpreted. That rule has now gone.
I propose that we make this test case an error and accept that for some implementations it may be an incompatibility. I suggest if you have an opinion, put it in the qt3tests bug tracker.

