4.2.1 Unexepcted JIT Menu error

[Enjay]

I'm sorry that this may be too vague to be useful, however, I thought I'd record it JIC (just in case).

I was playing a game with a map loaded but, as far as I know, nothing in it has any ZScript (or even DECORATE) modifications.

I pressed Esc and the menu did come up, but there was also an error printed to the screen/console. I thought that I had a logfile running, so I didn't pay too much attntion to the exact working but, it turns out that I forgot to set up a log file.

The error was something like:

Unexpected JIT error cannot make native menu call.

It was something like that anyway. Certainly "unexpected JIT" was in there, as was "menu" and "native".

Like I said, sorry if it's too vague.

[Marisa the Magician]

I think I've seen that happen a couple times, but it's almost impossible to reproduce intentionally.

[Nash]

Yeah I've had this randomly happen to me this year, probably less than 10 times. Absolutely zero clue what causes it or how to even reproduce it.

[dpJudas]

I'm afraid we need the exact error message. Unexpected JIT error means that something threw an exception while it was compiling a script function to native code. The rest of the error is the most important part of the message. Filtering by "menu" or "native" lists way too many source lines.

[Enjay]

It hasn't happened since. From what others have said it looks like it is a recurring thing, but a rare one. But I really can't remember more than what I posted. If it happens again, I'll be sure to make a proper record of the message.

[phantombeta]

Moving this to the On Hold Bugs forum so it doesn't get lost or accidentally closed.

[phantombeta]

This should be fixed now. If it somehow still happens (it shouldn't), this bug report can be reopened.

[Player701]

Is there a chance this may also fix the invalid instruction error? If the answer is yes, I can run some automated tests tonight to see if I get any more of those.

[phantombeta]

Is there a chance this may also fix the invalid instruction error? If the answer is yes, I can run some automated tests tonight to see if I get any more of those.

Very unlikely, but you can test it anyway. This requires the specific case of a parameterless function that doesn't trigger the VTBL opcode ending up as the first opcode of a function, and the only case where it could cause that other bug is with calls to non-native functions, as native ones just spit out the error reported in this thread. (AFAIK the function called needs to be a parameterless static function to trigger this)

[Player701]

OK, I'll leave it running and report tomorrow in the other thread.