Open mtdudek opened 1 year ago
Related is #481 -- I think a pragma would be non-ideal is we can avoid it vs like, a builtin that traps failure in the bytecode interpreter. Should still be synthesizable in the limit if we think about fail!
as a fatal error signal.
Thanks for reporting! As discussed in our April enhancement proposal review meeting: we'll have to think through the syntax we want, so we'll consider this at a lower priority for now.
Right now
#[test_proc]
and#[test]
always has to succeed.Adding new language testing feature will allow for negative testing. I think it would only require adding
#[test_fail]
and#[test_proc_fail]
to the language parser. Test procs or fn should keep their current type format, and the bulk of the changes would be applied to the DSLX interpreter.